diff --git a/.mypy.ini b/.mypy.ini
index 6284e54b..be6968bc 100644
--- a/.mypy.ini
+++ b/.mypy.ini
@@ -21,9 +21,30 @@ disallow_untyped_defs = True
 
 ###############################################################################
 # TPL-module options
+# Should only include exclusions for non-optional dependencies. Modules that
+# utilize optional dependencies should `type: ignore` within their module
+# scope.
+
+[mypy-cv2.*]
+ignore_missing_imports = True
+
+[mypy-matplotlib.*]
+ignore_missing_imports = True
+
+[mypy-PIL.*]
+ignore_missing_imports = True
 
 [mypy-pkg_resources.extern]
 ignore_missing_imports = True
 
 [mypy-pytest]
 ignore_missing_imports = True
+
+[mypy-skimage.*]
+ignore_missing_imports = True
+
+[mypy-sklearn.*]
+ignore_missing_imports = True
+
+[mypy-tqdm.*]
+ignore_missing_imports = True
diff --git a/poetry.lock b/poetry.lock
index 291c00e4..55f91a40 100644
--- a/poetry.lock
+++ b/poetry.lock
@@ -90,11 +90,22 @@ python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*, <4"
 [package.extras]
 toml = ["toml"]
 
+[[package]]
+name = "cycler"
+version = "0.10.0"
+description = "Composable style cycles"
+category = "main"
+optional = false
+python-versions = "*"
+
+[package.dependencies]
+six = "*"
+
 [[package]]
 name = "decorator"
 version = "5.0.7"
 description = "Decorators for Humans"
-category = "dev"
+category = "main"
 optional = false
 python-versions = ">=3.5"
 
@@ -139,6 +150,25 @@ category = "main"
 optional = false
 python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*"
 
+[[package]]
+name = "imageio"
+version = "2.9.0"
+description = "Library for reading and writing a wide range of image, video, scientific, and volumetric data formats."
+category = "main"
+optional = false
+python-versions = ">=3.5"
+
+[package.dependencies]
+numpy = "*"
+pillow = "*"
+
+[package.extras]
+ffmpeg = ["imageio-ffmpeg"]
+fits = ["astropy"]
+full = ["astropy", "gdal", "imageio-ffmpeg", "itk"]
+gdal = ["gdal"]
+itk = ["itk"]
+
 [[package]]
 name = "imagesize"
 version = "1.2.0"
@@ -239,6 +269,22 @@ MarkupSafe = ">=0.23"
 [package.extras]
 i18n = ["Babel (>=0.8)"]
 
+[[package]]
+name = "joblib"
+version = "1.0.1"
+description = "Lightweight pipelining with Python functions"
+category = "main"
+optional = false
+python-versions = ">=3.6"
+
+[[package]]
+name = "kiwisolver"
+version = "1.3.1"
+description = "A fast implementation of the Cassowary constraint solver"
+category = "main"
+optional = false
+python-versions = ">=3.6"
+
 [[package]]
 name = "livereload"
 version = "2.6.3"
@@ -259,6 +305,22 @@ category = "dev"
 optional = false
 python-versions = ">=2.7,!=3.0.*,!=3.1.*,!=3.2.*,!=3.3.*"
 
+[[package]]
+name = "matplotlib"
+version = "3.4.1"
+description = "Python plotting package"
+category = "main"
+optional = false
+python-versions = ">=3.7"
+
+[package.dependencies]
+cycler = ">=0.10"
+kiwisolver = ">=1.0.1"
+numpy = ">=1.16"
+pillow = ">=6.2.0"
+pyparsing = ">=2.2.1"
+python-dateutil = ">=2.7"
+
 [[package]]
 name = "mccabe"
 version = "0.6.1"
@@ -291,6 +353,30 @@ category = "dev"
 optional = false
 python-versions = "*"
 
+[[package]]
+name = "networkx"
+version = "2.5"
+description = "Python package for creating and manipulating graphs and networks"
+category = "main"
+optional = false
+python-versions = ">=3.6"
+
+[package.dependencies]
+decorator = ">=4.3.0"
+
+[package.extras]
+all = ["numpy", "scipy", "pandas", "matplotlib", "pygraphviz", "pydot", "pyyaml", "lxml", "pytest"]
+gdal = ["gdal"]
+lxml = ["lxml"]
+matplotlib = ["matplotlib"]
+numpy = ["numpy"]
+pandas = ["pandas"]
+pydot = ["pydot"]
+pygraphviz = ["pygraphviz"]
+pytest = ["pytest"]
+pyyaml = ["pyyaml"]
+scipy = ["scipy"]
+
 [[package]]
 name = "numpy"
 version = "1.20.2"
@@ -299,6 +385,17 @@ category = "main"
 optional = false
 python-versions = ">=3.7"
 
+[[package]]
+name = "opencv-python-headless"
+version = "4.5.1.48"
+description = "Wrapper package for OpenCV python bindings."
+category = "main"
+optional = false
+python-versions = ">=3.6"
+
+[package.dependencies]
+numpy = ">=1.13.3"
+
 [[package]]
 name = "packaging"
 version = "20.9"
@@ -418,7 +515,7 @@ python-versions = ">=3.5"
 name = "pyparsing"
 version = "2.4.7"
 description = "Python parsing module"
-category = "dev"
+category = "main"
 optional = false
 python-versions = ">=2.6, !=3.0.*, !=3.1.*, !=3.2.*"
 
@@ -459,6 +556,17 @@ pytest = ">=4.6"
 [package.extras]
 testing = ["fields", "hunter", "process-tests (==2.0.2)", "six", "pytest-xdist", "virtualenv"]
 
+[[package]]
+name = "python-dateutil"
+version = "2.8.1"
+description = "Extensions to the standard Python datetime module"
+category = "main"
+optional = false
+python-versions = "!=3.0.*,!=3.1.*,!=3.2.*,>=2.7"
+
+[package.dependencies]
+six = ">=1.5"
+
 [[package]]
 name = "pytz"
 version = "2021.1"
@@ -467,6 +575,17 @@ category = "dev"
 optional = false
 python-versions = "*"
 
+[[package]]
+name = "pywavelets"
+version = "1.1.1"
+description = "PyWavelets, wavelet transform module"
+category = "main"
+optional = false
+python-versions = ">=3.5"
+
+[package.dependencies]
+numpy = ">=1.13.3"
+
 [[package]]
 name = "requests"
 version = "2.25.1"
@@ -485,11 +604,66 @@ urllib3 = ">=1.21.1,<1.27"
 security = ["pyOpenSSL (>=0.14)", "cryptography (>=1.3.4)"]
 socks = ["PySocks (>=1.5.6,!=1.5.7)", "win-inet-pton"]
 
+[[package]]
+name = "scikit-image"
+version = "0.18.1"
+description = "Image processing in Python"
+category = "main"
+optional = false
+python-versions = ">=3.7"
+
+[package.dependencies]
+imageio = ">=2.3.0"
+matplotlib = ">=2.0.0,<3.0.0 || >3.0.0"
+networkx = ">=2.0"
+numpy = ">=1.16.5"
+pillow = ">=4.3.0,<7.1.0 || >7.1.0,<7.1.1 || >7.1.1"
+PyWavelets = ">=1.1.1"
+scipy = ">=1.0.1"
+tifffile = ">=2019.7.26"
+
+[package.extras]
+data = ["pooch (>=1.3.0)"]
+docs = ["sphinx (>=1.8,<=2.4.4)", "sphinx-gallery (>=0.7.0,!=0.8.0)", "numpydoc (>=1.0)", "sphinx-copybutton", "pytest-runner", "scikit-learn", "matplotlib (>=3.0.1)", "dask[array] (>=0.15.0,!=2.17.0)", "cloudpickle (>=0.2.1)", "pandas (>=0.23.0)", "seaborn (>=0.7.1)", "pooch (>=1.3.0)", "tifffile (>=2020.5.30)", "myst-parser", "ipywidgets", "plotly (>=4.10.0)"]
+optional = ["simpleitk", "astropy (>=3.1.2)", "qtpy", "pyamg", "dask[array] (>=1.0.0,!=2.17.0)", "cloudpickle (>=0.2.1)", "pooch (>=1.3.0)"]
+test = ["pytest (>=5.2.0)", "pytest-cov (>=2.7.0)", "pytest-localserver", "pytest-faulthandler", "flake8", "codecov", "pooch (>=1.3.0)"]
+
+[[package]]
+name = "scikit-learn"
+version = "0.24.2"
+description = "A set of python modules for machine learning and data mining"
+category = "main"
+optional = false
+python-versions = ">=3.6"
+
+[package.dependencies]
+joblib = ">=0.11"
+numpy = ">=1.13.3"
+scipy = ">=0.19.1"
+threadpoolctl = ">=2.0.0"
+
+[package.extras]
+benchmark = ["matplotlib (>=2.1.1)", "pandas (>=0.25.0)", "memory-profiler (>=0.57.0)"]
+docs = ["matplotlib (>=2.1.1)", "scikit-image (>=0.13)", "pandas (>=0.25.0)", "seaborn (>=0.9.0)", "memory-profiler (>=0.57.0)", "sphinx (>=3.2.0)", "sphinx-gallery (>=0.7.0)", "numpydoc (>=1.0.0)", "Pillow (>=7.1.2)", "sphinx-prompt (>=1.3.0)"]
+examples = ["matplotlib (>=2.1.1)", "scikit-image (>=0.13)", "pandas (>=0.25.0)", "seaborn (>=0.9.0)"]
+tests = ["matplotlib (>=2.1.1)", "scikit-image (>=0.13)", "pandas (>=0.25.0)", "pytest (>=5.0.1)", "pytest-cov (>=2.9.0)", "flake8 (>=3.8.2)", "mypy (>=0.770)", "pyamg (>=4.0.0)"]
+
+[[package]]
+name = "scipy"
+version = "1.6.1"
+description = "SciPy: Scientific Library for Python"
+category = "main"
+optional = false
+python-versions = ">=3.7"
+
+[package.dependencies]
+numpy = ">=1.16.5"
+
 [[package]]
 name = "six"
 version = "1.15.0"
 description = "Python 2 and 3 compatibility utilities"
-category = "dev"
+category = "main"
 optional = false
 python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*"
 
@@ -675,6 +849,28 @@ python-versions = ">=3.5"
 lint = ["flake8", "mypy", "docutils-stubs"]
 test = ["pytest"]
 
+[[package]]
+name = "threadpoolctl"
+version = "2.1.0"
+description = "threadpoolctl"
+category = "main"
+optional = false
+python-versions = ">=3.5"
+
+[[package]]
+name = "tifffile"
+version = "2021.4.8"
+description = "Read and write TIFF files"
+category = "main"
+optional = false
+python-versions = ">=3.7"
+
+[package.dependencies]
+numpy = ">=1.15.1"
+
+[package.extras]
+all = ["imagecodecs (>=2021.3.31)", "matplotlib (>=3.2)", "lxml"]
+
 [[package]]
 name = "toml"
 version = "0.10.2"
@@ -691,6 +887,19 @@ category = "dev"
 optional = false
 python-versions = ">= 3.5"
 
+[[package]]
+name = "tqdm"
+version = "4.60.0"
+description = "Fast, Extensible Progress Meter"
+category = "main"
+optional = false
+python-versions = "!=3.0.*,!=3.1.*,!=3.2.*,!=3.3.*,>=2.7"
+
+[package.extras]
+dev = ["py-make (>=0.1.0)", "twine", "wheel"]
+notebook = ["ipywidgets (>=6)"]
+telegram = ["requests"]
+
 [[package]]
 name = "traitlets"
 version = "5.0.5"
@@ -757,7 +966,7 @@ testing = ["pytest (>=4.6)", "pytest-checkdocs (>=1.2.3)", "pytest-flake8", "pyt
 [metadata]
 lock-version = "1.1"
 python-versions = "^3.7"
-content-hash = "a3b7e78407487c0af3dcefebd49e5bccc548b19cb9107ccad0b8d174724a35a1"
+content-hash = "accc27b19d356ac8f0f7636b0a1d83de37441679b6ea11524de539fa8007ef5d"
 
 [metadata.files]
 alabaster = [
@@ -850,6 +1059,10 @@ coverage = [
     {file = "coverage-5.5-pp37-none-any.whl", hash = "sha256:2a3859cb82dcbda1cfd3e6f71c27081d18aa251d20a17d87d26d4cd216fb0af4"},
     {file = "coverage-5.5.tar.gz", hash = "sha256:ebe78fe9a0e874362175b02371bdfbee64d8edc42a044253ddf4ee7d3c15212c"},
 ]
+cycler = [
+    {file = "cycler-0.10.0-py2.py3-none-any.whl", hash = "sha256:1d8a5ae1ff6c5cf9b93e8811e581232ad8920aeec647c37316ceac982b08cb2d"},
+    {file = "cycler-0.10.0.tar.gz", hash = "sha256:cd7b2d1018258d7247a71425e9f26463dfb444d411c39569972f4ce586b0c9d8"},
+]
 decorator = [
     {file = "decorator-5.0.7-py3-none-any.whl", hash = "sha256:945d84890bb20cc4a2f4a31fc4311c0c473af65ea318617f13a7257c9a58bc98"},
     {file = "decorator-5.0.7.tar.gz", hash = "sha256:6f201a6c4dac3d187352661f508b9364ec8091217442c9478f1f83c003a0f060"},
@@ -870,6 +1083,10 @@ idna = [
     {file = "idna-2.10-py2.py3-none-any.whl", hash = "sha256:b97d804b1e9b523befed77c48dacec60e6dcb0b5391d57af6a65a312a90648c0"},
     {file = "idna-2.10.tar.gz", hash = "sha256:b307872f855b18632ce0c21c5e45be78c0ea7ae4c15c828c20788b26921eb3f6"},
 ]
+imageio = [
+    {file = "imageio-2.9.0-py3-none-any.whl", hash = "sha256:3604d751f03002e8e0e7650aa71d8d9148144a87daf17cb1f3228e80747f2e6b"},
+    {file = "imageio-2.9.0.tar.gz", hash = "sha256:52ddbaeca2dccf53ba2d6dec5676ca7bc3b2403ef8b37f7da78b7654bb3e10f0"},
+]
 imagesize = [
     {file = "imagesize-1.2.0-py2.py3-none-any.whl", hash = "sha256:6965f19a6a2039c7d48bca7dba2473069ff854c36ae6f19d2cde309d998228a1"},
     {file = "imagesize-1.2.0.tar.gz", hash = "sha256:b1f6b5a4eab1f73479a50fb79fcf729514a900c341d8503d62a62dbc4127a2b1"},
@@ -898,6 +1115,44 @@ jinja2 = [
     {file = "Jinja2-2.11.3-py2.py3-none-any.whl", hash = "sha256:03e47ad063331dd6a3f04a43eddca8a966a26ba0c5b7207a9a9e4e08f1b29419"},
     {file = "Jinja2-2.11.3.tar.gz", hash = "sha256:a6d58433de0ae800347cab1fa3043cebbabe8baa9d29e668f1c768cb87a333c6"},
 ]
+joblib = [
+    {file = "joblib-1.0.1-py3-none-any.whl", hash = "sha256:feeb1ec69c4d45129954f1b7034954241eedfd6ba39b5e9e4b6883be3332d5e5"},
+    {file = "joblib-1.0.1.tar.gz", hash = "sha256:9c17567692206d2f3fb9ecf5e991084254fe631665c450b443761c4186a613f7"},
+]
+kiwisolver = [
+    {file = "kiwisolver-1.3.1-cp36-cp36m-macosx_10_9_x86_64.whl", hash = "sha256:fd34fbbfbc40628200730bc1febe30631347103fc8d3d4fa012c21ab9c11eca9"},
+    {file = "kiwisolver-1.3.1-cp36-cp36m-manylinux1_i686.whl", hash = "sha256:d3155d828dec1d43283bd24d3d3e0d9c7c350cdfcc0bd06c0ad1209c1bbc36d0"},
+    {file = "kiwisolver-1.3.1-cp36-cp36m-manylinux1_x86_64.whl", hash = "sha256:5a7a7dbff17e66fac9142ae2ecafb719393aaee6a3768c9de2fd425c63b53e21"},
+    {file = "kiwisolver-1.3.1-cp36-cp36m-manylinux2014_aarch64.whl", hash = "sha256:f8d6f8db88049a699817fd9178782867bf22283e3813064302ac59f61d95be05"},
+    {file = "kiwisolver-1.3.1-cp36-cp36m-manylinux2014_ppc64le.whl", hash = "sha256:5f6ccd3dd0b9739edcf407514016108e2280769c73a85b9e59aa390046dbf08b"},
+    {file = "kiwisolver-1.3.1-cp36-cp36m-win32.whl", hash = "sha256:225e2e18f271e0ed8157d7f4518ffbf99b9450fca398d561eb5c4a87d0986dd9"},
+    {file = "kiwisolver-1.3.1-cp36-cp36m-win_amd64.whl", hash = "sha256:cf8b574c7b9aa060c62116d4181f3a1a4e821b2ec5cbfe3775809474113748d4"},
+    {file = "kiwisolver-1.3.1-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:232c9e11fd7ac3a470d65cd67e4359eee155ec57e822e5220322d7b2ac84fbf0"},
+    {file = "kiwisolver-1.3.1-cp37-cp37m-manylinux1_i686.whl", hash = "sha256:b38694dcdac990a743aa654037ff1188c7a9801ac3ccc548d3341014bc5ca278"},
+    {file = "kiwisolver-1.3.1-cp37-cp37m-manylinux1_x86_64.whl", hash = "sha256:ca3820eb7f7faf7f0aa88de0e54681bddcb46e485beb844fcecbcd1c8bd01689"},
+    {file = "kiwisolver-1.3.1-cp37-cp37m-manylinux2014_aarch64.whl", hash = "sha256:c8fd0f1ae9d92b42854b2979024d7597685ce4ada367172ed7c09edf2cef9cb8"},
+    {file = "kiwisolver-1.3.1-cp37-cp37m-manylinux2014_ppc64le.whl", hash = "sha256:1e1bc12fb773a7b2ffdeb8380609f4f8064777877b2225dec3da711b421fda31"},
+    {file = "kiwisolver-1.3.1-cp37-cp37m-win32.whl", hash = "sha256:72c99e39d005b793fb7d3d4e660aed6b6281b502e8c1eaf8ee8346023c8e03bc"},
+    {file = "kiwisolver-1.3.1-cp37-cp37m-win_amd64.whl", hash = "sha256:8be8d84b7d4f2ba4ffff3665bcd0211318aa632395a1a41553250484a871d454"},
+    {file = "kiwisolver-1.3.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:31dfd2ac56edc0ff9ac295193eeaea1c0c923c0355bf948fbd99ed6018010b72"},
+    {file = "kiwisolver-1.3.1-cp38-cp38-manylinux1_i686.whl", hash = "sha256:563c649cfdef27d081c84e72a03b48ea9408c16657500c312575ae9d9f7bc1c3"},
+    {file = "kiwisolver-1.3.1-cp38-cp38-manylinux1_x86_64.whl", hash = "sha256:78751b33595f7f9511952e7e60ce858c6d64db2e062afb325985ddbd34b5c131"},
+    {file = "kiwisolver-1.3.1-cp38-cp38-manylinux2014_aarch64.whl", hash = "sha256:a357fd4f15ee49b4a98b44ec23a34a95f1e00292a139d6015c11f55774ef10de"},
+    {file = "kiwisolver-1.3.1-cp38-cp38-manylinux2014_ppc64le.whl", hash = "sha256:5989db3b3b34b76c09253deeaf7fbc2707616f130e166996606c284395da3f18"},
+    {file = "kiwisolver-1.3.1-cp38-cp38-win32.whl", hash = "sha256:c08e95114951dc2090c4a630c2385bef681cacf12636fb0241accdc6b303fd81"},
+    {file = "kiwisolver-1.3.1-cp38-cp38-win_amd64.whl", hash = "sha256:44a62e24d9b01ba94ae7a4a6c3fb215dc4af1dde817e7498d901e229aaf50e4e"},
+    {file = "kiwisolver-1.3.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:50af681a36b2a1dee1d3c169ade9fdc59207d3c31e522519181e12f1b3ba7000"},
+    {file = "kiwisolver-1.3.1-cp39-cp39-manylinux1_i686.whl", hash = "sha256:a53d27d0c2a0ebd07e395e56a1fbdf75ffedc4a05943daf472af163413ce9598"},
+    {file = "kiwisolver-1.3.1-cp39-cp39-manylinux1_x86_64.whl", hash = "sha256:834ee27348c4aefc20b479335fd422a2c69db55f7d9ab61721ac8cd83eb78882"},
+    {file = "kiwisolver-1.3.1-cp39-cp39-manylinux2014_aarch64.whl", hash = "sha256:5c3e6455341008a054cccee8c5d24481bcfe1acdbc9add30aa95798e95c65621"},
+    {file = "kiwisolver-1.3.1-cp39-cp39-manylinux2014_ppc64le.whl", hash = "sha256:acef3d59d47dd85ecf909c359d0fd2c81ed33bdff70216d3956b463e12c38a54"},
+    {file = "kiwisolver-1.3.1-cp39-cp39-win32.whl", hash = "sha256:c5518d51a0735b1e6cee1fdce66359f8d2b59c3ca85dc2b0813a8aa86818a030"},
+    {file = "kiwisolver-1.3.1-cp39-cp39-win_amd64.whl", hash = "sha256:b9edd0110a77fc321ab090aaa1cfcaba1d8499850a12848b81be2222eab648f6"},
+    {file = "kiwisolver-1.3.1-pp36-pypy36_pp73-macosx_10_9_x86_64.whl", hash = "sha256:0cd53f403202159b44528498de18f9285b04482bab2a6fc3f5dd8dbb9352e30d"},
+    {file = "kiwisolver-1.3.1-pp36-pypy36_pp73-manylinux2010_x86_64.whl", hash = "sha256:33449715e0101e4d34f64990352bce4095c8bf13bed1b390773fc0a7295967b3"},
+    {file = "kiwisolver-1.3.1-pp36-pypy36_pp73-win32.whl", hash = "sha256:401a2e9afa8588589775fe34fc22d918ae839aaaf0c0e96441c0fdbce6d8ebe6"},
+    {file = "kiwisolver-1.3.1.tar.gz", hash = "sha256:950a199911a8d94683a6b10321f9345d5a3a8433ec58b217ace979e18f16e248"},
+]
 livereload = [
     {file = "livereload-2.6.3.tar.gz", hash = "sha256:776f2f865e59fde56490a56bcc6773b6917366bce0c267c60ee8aaf1a0959869"},
 ]
@@ -955,6 +1210,27 @@ markupsafe = [
     {file = "MarkupSafe-1.1.1-cp39-cp39-win_amd64.whl", hash = "sha256:b7d644ddb4dbd407d31ffb699f1d140bc35478da613b441c582aeb7c43838dd8"},
     {file = "MarkupSafe-1.1.1.tar.gz", hash = "sha256:29872e92839765e546828bb7754a68c418d927cd064fd4708fab9fe9c8bb116b"},
 ]
+matplotlib = [
+    {file = "matplotlib-3.4.1-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:7a54efd6fcad9cb3cd5ef2064b5a3eeb0b63c99f26c346bdcf66e7c98294d7cc"},
+    {file = "matplotlib-3.4.1-cp37-cp37m-manylinux1_i686.whl", hash = "sha256:86dc94e44403fa0f2b1dd76c9794d66a34e821361962fe7c4e078746362e3b14"},
+    {file = "matplotlib-3.4.1-cp37-cp37m-manylinux1_x86_64.whl", hash = "sha256:574306171b84cd6854c83dc87bc353cacc0f60184149fb00c9ea871eca8c1ecb"},
+    {file = "matplotlib-3.4.1-cp37-cp37m-manylinux2014_aarch64.whl", hash = "sha256:84a10e462120aa7d9eb6186b50917ed5a6286ee61157bfc17c5b47987d1a9068"},
+    {file = "matplotlib-3.4.1-cp37-cp37m-win32.whl", hash = "sha256:81e6fe8b18ef5be67f40a1d4f07d5a4ed21d3878530193898449ddef7793952f"},
+    {file = "matplotlib-3.4.1-cp37-cp37m-win_amd64.whl", hash = "sha256:c45e7bf89ea33a2adaef34774df4e692c7436a18a48bcb0e47a53e698a39fa39"},
+    {file = "matplotlib-3.4.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:1f83a32e4b6045191f9d34e4dc68c0a17c870b57ef9cca518e516da591246e79"},
+    {file = "matplotlib-3.4.1-cp38-cp38-manylinux1_i686.whl", hash = "sha256:a18cc1ab4a35b845cf33b7880c979f5c609fd26c2d6e74ddfacb73dcc60dd956"},
+    {file = "matplotlib-3.4.1-cp38-cp38-manylinux1_x86_64.whl", hash = "sha256:ac2a30a09984c2719f112a574b6543ccb82d020fd1b23b4d55bf4759ba8dd8f5"},
+    {file = "matplotlib-3.4.1-cp38-cp38-manylinux2014_aarch64.whl", hash = "sha256:a97781453ac79409ddf455fccf344860719d95142f9c334f2a8f3fff049ffec3"},
+    {file = "matplotlib-3.4.1-cp38-cp38-win32.whl", hash = "sha256:2eee37340ca1b353e0a43a33da79d0cd4bcb087064a0c3c3d1329cdea8fbc6f3"},
+    {file = "matplotlib-3.4.1-cp38-cp38-win_amd64.whl", hash = "sha256:90dbc007f6389bcfd9ef4fe5d4c78c8d2efe4e0ebefd48b4f221cdfed5672be2"},
+    {file = "matplotlib-3.4.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:7f16660edf9a8bcc0f766f51c9e1b9d2dc6ceff6bf636d2dbd8eb925d5832dfd"},
+    {file = "matplotlib-3.4.1-cp39-cp39-manylinux1_i686.whl", hash = "sha256:a989022f89cda417f82dbf65e0a830832afd8af743d05d1414fb49549287ff04"},
+    {file = "matplotlib-3.4.1-cp39-cp39-manylinux1_x86_64.whl", hash = "sha256:be4430b33b25e127fc4ea239cc386389de420be4d63e71d5359c20b562951ce1"},
+    {file = "matplotlib-3.4.1-cp39-cp39-manylinux2014_aarch64.whl", hash = "sha256:7561fd541477d41f3aa09457c434dd1f7604f3bd26d7858d52018f5dfe1c06d1"},
+    {file = "matplotlib-3.4.1-cp39-cp39-win32.whl", hash = "sha256:9f374961a3996c2d1b41ba3145462c3708a89759e604112073ed6c8bdf9f622f"},
+    {file = "matplotlib-3.4.1-cp39-cp39-win_amd64.whl", hash = "sha256:53ceb12ef44f8982b45adc7a0889a7e2df1d758e8b360f460e435abe8a8cd658"},
+    {file = "matplotlib-3.4.1.tar.gz", hash = "sha256:84d4c4f650f356678a5d658a43ca21a41fca13f9b8b00169c0b76e6a6a948908"},
+]
 mccabe = [
     {file = "mccabe-0.6.1-py2.py3-none-any.whl", hash = "sha256:ab8a6258860da4b6677da4bd2fe5dc2c659cff31b3ee4f7f5d64e79735b80d42"},
     {file = "mccabe-0.6.1.tar.gz", hash = "sha256:dd8d182285a0fe56bace7f45b5e7d1a6ebcbf524e8f3bd87eb0f125271b8831f"},
@@ -987,6 +1263,10 @@ mypy-extensions = [
     {file = "mypy_extensions-0.4.3-py2.py3-none-any.whl", hash = "sha256:090fedd75945a69ae91ce1303b5824f428daf5a028d2f6ab8a299250a846f15d"},
     {file = "mypy_extensions-0.4.3.tar.gz", hash = "sha256:2d82818f5bb3e369420cb3c4060a7970edba416647068eb4c5343488a6c604a8"},
 ]
+networkx = [
+    {file = "networkx-2.5-py3-none-any.whl", hash = "sha256:8c5812e9f798d37c50570d15c4a69d5710a18d77bafc903ee9c5fba7454c616c"},
+    {file = "networkx-2.5.tar.gz", hash = "sha256:7978955423fbc9639c10498878be59caf99b44dc304c2286162fd24b458c1602"},
+]
 numpy = [
     {file = "numpy-1.20.2-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:e9459f40244bb02b2f14f6af0cd0732791d72232bbb0dc4bab57ef88e75f6935"},
     {file = "numpy-1.20.2-cp37-cp37m-manylinux1_i686.whl", hash = "sha256:a8e6859913ec8eeef3dbe9aed3bf475347642d1cdd6217c30f28dee8903528e6"},
@@ -1013,6 +1293,33 @@ numpy = [
     {file = "numpy-1.20.2-pp37-pypy37_pp73-manylinux2010_x86_64.whl", hash = "sha256:97ce8b8ace7d3b9288d88177e66ee75480fb79b9cf745e91ecfe65d91a856042"},
     {file = "numpy-1.20.2.zip", hash = "sha256:878922bf5ad7550aa044aa9301d417e2d3ae50f0f577de92051d739ac6096cee"},
 ]
+opencv-python-headless = [
+    {file = "opencv-python-headless-4.5.1.48.tar.gz", hash = "sha256:d16825755e7b5a6d8737f93e116670229e1510199e0af9213004e187ae0dbcc5"},
+    {file = "opencv_python_headless-4.5.1.48-cp36-cp36m-macosx_10_13_x86_64.whl", hash = "sha256:f2011ecb3980bbed283d17d43e0f1221bac88c0cac1a6fb59a056544de2df2f7"},
+    {file = "opencv_python_headless-4.5.1.48-cp36-cp36m-manylinux2014_aarch64.whl", hash = "sha256:777fb596e04331f73ef5b0c1faa4d33348f29ca58216d9286355c16f5489c939"},
+    {file = "opencv_python_headless-4.5.1.48-cp36-cp36m-manylinux2014_i686.whl", hash = "sha256:924aa4c34ead0b817309f42291dc526b2a7755476afe3009d1e275fc3090d92d"},
+    {file = "opencv_python_headless-4.5.1.48-cp36-cp36m-manylinux2014_x86_64.whl", hash = "sha256:0e02809db2968e54f3c23340be6ff9a1428b3f03f5dca7cd5aceda66e319ce86"},
+    {file = "opencv_python_headless-4.5.1.48-cp36-cp36m-win32.whl", hash = "sha256:522f12dd994e064a30562adfd63b9439099bd7c80819f5261c37ebe593283c9e"},
+    {file = "opencv_python_headless-4.5.1.48-cp36-cp36m-win_amd64.whl", hash = "sha256:e3027e0d1b71b68b5cfe1ea9c627e323dff71112c854ba19805258d8fc6c630e"},
+    {file = "opencv_python_headless-4.5.1.48-cp37-cp37m-macosx_10_13_x86_64.whl", hash = "sha256:f5e40a06116460ef2fd2d1c24be3b65f8bfb5fcdfe433f3fc01bcb4c2eb485bf"},
+    {file = "opencv_python_headless-4.5.1.48-cp37-cp37m-manylinux2014_aarch64.whl", hash = "sha256:ba2f0bd46e9534f29969e39f7895cbea9764173102b0c04b7818b8a9910d66e4"},
+    {file = "opencv_python_headless-4.5.1.48-cp37-cp37m-manylinux2014_i686.whl", hash = "sha256:f7f8e4f7c63c8e95eb210f4cba88d0069a0a964d6335d7a35b07f0d0baa13558"},
+    {file = "opencv_python_headless-4.5.1.48-cp37-cp37m-manylinux2014_x86_64.whl", hash = "sha256:aa562c520f46283423ba8fac29099458e42deab697a9abd0491622e421c5c454"},
+    {file = "opencv_python_headless-4.5.1.48-cp37-cp37m-win32.whl", hash = "sha256:657cdd9fbbbbb7e898ae3d9b0649b367d0e440d429c714104d069c5612d578bb"},
+    {file = "opencv_python_headless-4.5.1.48-cp37-cp37m-win_amd64.whl", hash = "sha256:fe02a943b1a28b505e954fbce24e867119a3eb4351f93adad55c6cfe81a70484"},
+    {file = "opencv_python_headless-4.5.1.48-cp38-cp38-macosx_10_13_x86_64.whl", hash = "sha256:526b9e19cf6300f0891d8f427eac1048091912332bb781eb4957a4994bfbe608"},
+    {file = "opencv_python_headless-4.5.1.48-cp38-cp38-manylinux2014_aarch64.whl", hash = "sha256:5f0c7d34fa9953706c2a1a6d2760a91dc5a68cab3df16af61609894c6c8586f1"},
+    {file = "opencv_python_headless-4.5.1.48-cp38-cp38-manylinux2014_i686.whl", hash = "sha256:1f7c14f5d4e5af4dc4669fc6b4a983b36072a934c439ac11266b930496da8255"},
+    {file = "opencv_python_headless-4.5.1.48-cp38-cp38-manylinux2014_x86_64.whl", hash = "sha256:7c75680ffdf32d7044415a215d1fc60dbec14a7f2f0b59a85f0f74ef5efcb6ad"},
+    {file = "opencv_python_headless-4.5.1.48-cp38-cp38-win32.whl", hash = "sha256:2560dcf3c1158226b066302f777bfe0f65282410b8d90871dd872306c967d1f1"},
+    {file = "opencv_python_headless-4.5.1.48-cp38-cp38-win_amd64.whl", hash = "sha256:9aa04a491c534531029fbac61da961fae0bf4abb1786eed9c91befde4ca7bd81"},
+    {file = "opencv_python_headless-4.5.1.48-cp39-cp39-macosx_10_13_x86_64.whl", hash = "sha256:96d1da6ad061d8f3509668d398d14dd8265e529d6407f87eb26e7f4ddf043cc0"},
+    {file = "opencv_python_headless-4.5.1.48-cp39-cp39-manylinux2014_aarch64.whl", hash = "sha256:243aea91cc1e36a47c46da4cc408071af39444a48df1fe1539ea8d7990500fd2"},
+    {file = "opencv_python_headless-4.5.1.48-cp39-cp39-manylinux2014_i686.whl", hash = "sha256:2e6a9a88617a0ef7219cff24ba78a58416670a77e6ac63975f9009af3319ab63"},
+    {file = "opencv_python_headless-4.5.1.48-cp39-cp39-manylinux2014_x86_64.whl", hash = "sha256:372149d007e20bf556b7687591d22b58b56b3c225f492da051d81587e5dc7411"},
+    {file = "opencv_python_headless-4.5.1.48-cp39-cp39-win32.whl", hash = "sha256:65c9ea57be5ebcaed009b3fee14fe59f3b6aa1e573fe08c5039ff057ad593c53"},
+    {file = "opencv_python_headless-4.5.1.48-cp39-cp39-win_amd64.whl", hash = "sha256:a322d4df14c3a5f19701a023b3da91e9b8af8653b0d2ee0c50b0b341212f7343"},
+]
 packaging = [
     {file = "packaging-20.9-py2.py3-none-any.whl", hash = "sha256:67714da7f7bc052e064859c05c595155bd1ee9f69f76557e21f051443c20947a"},
     {file = "packaging-20.9.tar.gz", hash = "sha256:5b327ac1320dc863dca72f4514ecc086f31186744b84a230374cc1fd776feae5"},
@@ -1104,14 +1411,124 @@ pytest-cov = [
     {file = "pytest-cov-2.11.1.tar.gz", hash = "sha256:359952d9d39b9f822d9d29324483e7ba04a3a17dd7d05aa6beb7ea01e359e5f7"},
     {file = "pytest_cov-2.11.1-py2.py3-none-any.whl", hash = "sha256:bdb9fdb0b85a7cc825269a4c56b48ccaa5c7e365054b6038772c32ddcdc969da"},
 ]
+python-dateutil = [
+    {file = "python-dateutil-2.8.1.tar.gz", hash = "sha256:73ebfe9dbf22e832286dafa60473e4cd239f8592f699aa5adaf10050e6e1823c"},
+    {file = "python_dateutil-2.8.1-py2.py3-none-any.whl", hash = "sha256:75bb3f31ea686f1197762692a9ee6a7550b59fc6ca3a1f4b5d7e32fb98e2da2a"},
+]
 pytz = [
     {file = "pytz-2021.1-py2.py3-none-any.whl", hash = "sha256:eb10ce3e7736052ed3623d49975ce333bcd712c7bb19a58b9e2089d4057d0798"},
     {file = "pytz-2021.1.tar.gz", hash = "sha256:83a4a90894bf38e243cf052c8b58f381bfe9a7a483f6a9cab140bc7f702ac4da"},
 ]
+pywavelets = [
+    {file = "PyWavelets-1.1.1-cp35-cp35m-macosx_10_6_intel.whl", hash = "sha256:35959c041ec014648575085a97b498eafbbaa824f86f6e4a59bfdef8a3fe6308"},
+    {file = "PyWavelets-1.1.1-cp35-cp35m-manylinux1_i686.whl", hash = "sha256:55e39ec848ceec13c9fa1598253ae9dd5c31d09dfd48059462860d2b908fb224"},
+    {file = "PyWavelets-1.1.1-cp35-cp35m-manylinux1_x86_64.whl", hash = "sha256:c06d2e340c7bf8b9ec71da2284beab8519a3908eab031f4ea126e8ccfc3fd567"},
+    {file = "PyWavelets-1.1.1-cp35-cp35m-win32.whl", hash = "sha256:be105382961745f88d8196bba5a69ee2c4455d87ad2a2e5d1eed6bd7fda4d3fd"},
+    {file = "PyWavelets-1.1.1-cp35-cp35m-win_amd64.whl", hash = "sha256:076ca8907001fdfe4205484f719d12b4a0262dfe6652fa1cfc3c5c362d14dc84"},
+    {file = "PyWavelets-1.1.1-cp36-cp36m-macosx_10_9_x86_64.whl", hash = "sha256:7947e51ca05489b85928af52a34fe67022ab5b81d4ae32a4109a99e883a0635e"},
+    {file = "PyWavelets-1.1.1-cp36-cp36m-manylinux1_i686.whl", hash = "sha256:9e2528823ccf5a0a1d23262dfefe5034dce89cd84e4e124dc553dfcdf63ebb92"},
+    {file = "PyWavelets-1.1.1-cp36-cp36m-manylinux1_x86_64.whl", hash = "sha256:80b924edbc012ded8aa8b91cb2fd6207fb1a9a3a377beb4049b8a07445cec6f0"},
+    {file = "PyWavelets-1.1.1-cp36-cp36m-manylinux2014_aarch64.whl", hash = "sha256:c2a799e79cee81a862216c47e5623c97b95f1abee8dd1f9eed736df23fb653fb"},
+    {file = "PyWavelets-1.1.1-cp36-cp36m-win32.whl", hash = "sha256:d510aef84d9852653d079c84f2f81a82d5d09815e625f35c95714e7364570ad4"},
+    {file = "PyWavelets-1.1.1-cp36-cp36m-win_amd64.whl", hash = "sha256:889d4c5c5205a9c90118c1980df526857929841df33e4cd1ff1eff77c6817a65"},
+    {file = "PyWavelets-1.1.1-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:68b5c33741d26c827074b3d8f0251de1c3019bb9567b8d303eb093c822ce28f1"},
+    {file = "PyWavelets-1.1.1-cp37-cp37m-manylinux1_i686.whl", hash = "sha256:18a51b3f9416a2ae6e9a35c4af32cf520dd7895f2b69714f4aa2f4342fca47f9"},
+    {file = "PyWavelets-1.1.1-cp37-cp37m-manylinux1_x86_64.whl", hash = "sha256:cfe79844526dd92e3ecc9490b5031fca5f8ab607e1e858feba232b1b788ff0ea"},
+    {file = "PyWavelets-1.1.1-cp37-cp37m-manylinux2014_aarch64.whl", hash = "sha256:2f7429eeb5bf9c7068002d0d7f094ed654c77a70ce5e6198737fd68ab85f8311"},
+    {file = "PyWavelets-1.1.1-cp37-cp37m-win32.whl", hash = "sha256:720dbcdd3d91c6dfead79c80bf8b00a1d8aa4e5d551dc528c6d5151e4efc3403"},
+    {file = "PyWavelets-1.1.1-cp37-cp37m-win_amd64.whl", hash = "sha256:bc5e87b72371da87c9bebc68e54882aada9c3114e640de180f62d5da95749cd3"},
+    {file = "PyWavelets-1.1.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:98b2669c5af842a70cfab33a7043fcb5e7535a690a00cd251b44c9be0be418e5"},
+    {file = "PyWavelets-1.1.1-cp38-cp38-manylinux1_i686.whl", hash = "sha256:e02a0558e0c2ac8b8bbe6a6ac18c136767ec56b96a321e0dfde2173adfa5a504"},
+    {file = "PyWavelets-1.1.1-cp38-cp38-manylinux1_x86_64.whl", hash = "sha256:6162dc0ae04669ea04b4b51420777b9ea2d30b0a9d02901b2a3b4d61d159c2e9"},
+    {file = "PyWavelets-1.1.1-cp38-cp38-manylinux2014_aarch64.whl", hash = "sha256:39c74740718e420d38c78ca4498568fa57976d78d5096277358e0fa9629a7aea"},
+    {file = "PyWavelets-1.1.1-cp38-cp38-win32.whl", hash = "sha256:79f5b54f9dc353e5ee47f0c3f02bebd2c899d49780633aa771fed43fa20b3149"},
+    {file = "PyWavelets-1.1.1-cp38-cp38-win_amd64.whl", hash = "sha256:935ff247b8b78bdf77647fee962b1cc208c51a7b229db30b9ba5f6da3e675178"},
+    {file = "PyWavelets-1.1.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:6ebfefebb5c6494a3af41ad8c60248a95da267a24b79ed143723d4502b1fe4d7"},
+    {file = "PyWavelets-1.1.1-cp39-cp39-manylinux1_i686.whl", hash = "sha256:6bc78fb9c42a716309b4ace56f51965d8b5662c3ba19d4591749f31773db1125"},
+    {file = "PyWavelets-1.1.1-cp39-cp39-manylinux1_x86_64.whl", hash = "sha256:411e17ca6ed8cf5e18a7ca5ee06a91c25800cc6c58c77986202abf98d749273a"},
+    {file = "PyWavelets-1.1.1-cp39-cp39-manylinux2014_aarch64.whl", hash = "sha256:83c5e3eb78ce111c2f0b45f46106cc697c3cb6c4e5f51308e1f81b512c70c8fb"},
+    {file = "PyWavelets-1.1.1-cp39-cp39-win32.whl", hash = "sha256:2b634a54241c190ee989a4af87669d377b37c91bcc9cf0efe33c10ff847f7841"},
+    {file = "PyWavelets-1.1.1-cp39-cp39-win_amd64.whl", hash = "sha256:732bab78435c48be5d6bc75486ef629d7c8f112e07b313bf1f1a2220ab437277"},
+    {file = "PyWavelets-1.1.1.tar.gz", hash = "sha256:1a64b40f6acb4ffbaccce0545d7fc641744f95351f62e4c6aaa40549326008c9"},
+]
 requests = [
     {file = "requests-2.25.1-py2.py3-none-any.whl", hash = "sha256:c210084e36a42ae6b9219e00e48287def368a26d03a048ddad7bfee44f75871e"},
     {file = "requests-2.25.1.tar.gz", hash = "sha256:27973dd4a904a4f13b263a19c866c13b92a39ed1c964655f025f3f8d3d75b804"},
 ]
+scikit-image = [
+    {file = "scikit-image-0.18.1.tar.gz", hash = "sha256:fbb618ca911867bce45574c1639618cdfb5d94e207432b19bc19563d80d2f171"},
+    {file = "scikit_image-0.18.1-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:1cd05c882ffb2a271a1f20b4afe937d63d55b8753c3d652f11495883a7800ebe"},
+    {file = "scikit_image-0.18.1-cp37-cp37m-manylinux1_i686.whl", hash = "sha256:e972c628ad9ba52c298b032368e29af9bd5eeb81ce33bc2d9b039a81661c99c5"},
+    {file = "scikit_image-0.18.1-cp37-cp37m-manylinux1_x86_64.whl", hash = "sha256:1256017c513e8e1b8b9da73e5fd1e605d0077bbbc8e5c8d6c2cab36400131c6c"},
+    {file = "scikit_image-0.18.1-cp37-cp37m-win32.whl", hash = "sha256:ec25e4110951d3a280421bb10dd510a082ba83d86e20d706294faf7899cdb3d5"},
+    {file = "scikit_image-0.18.1-cp37-cp37m-win_amd64.whl", hash = "sha256:2eea42706a25ae6e0cebaf1914e2ab1c04061b1f3c9966d76025d58a2e9188fc"},
+    {file = "scikit_image-0.18.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:76446e2402e64d7dba78eeae8aa86e92a0cafe5b1c9e6235bd8d067471ed2788"},
+    {file = "scikit_image-0.18.1-cp38-cp38-manylinux1_i686.whl", hash = "sha256:d5ad4a9b4c9797d4c4c48f45fa224c5ebff22b9b0af636c3ecb8addbb66c21e6"},
+    {file = "scikit_image-0.18.1-cp38-cp38-manylinux1_x86_64.whl", hash = "sha256:23f9178b21c752bfb4e4ea3a3fa0ff79bc5a401bc75ddb4661f2cebd1c2b0e24"},
+    {file = "scikit_image-0.18.1-cp38-cp38-win32.whl", hash = "sha256:d746540cafe7776c6d05a0b40ec744bb8d33d1ddc51faba601d26c02593d8bcc"},
+    {file = "scikit_image-0.18.1-cp38-cp38-win_amd64.whl", hash = "sha256:30447af3f5b7c9491f2d3db5bc275493d1b91bf1dd16b67e2fd79a6bb95d8ee9"},
+    {file = "scikit_image-0.18.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:ae6659b3a8bd4bba7e9dcbfd0064e443b32c7054bf09174749db896730fcf42e"},
+    {file = "scikit_image-0.18.1-cp39-cp39-manylinux1_i686.whl", hash = "sha256:2c058770c6ad6e0fe6c30f59970c9c65fa740ff014d121d8c341664cd792cf49"},
+    {file = "scikit_image-0.18.1-cp39-cp39-manylinux1_x86_64.whl", hash = "sha256:c700336a7f96109c74154090c5e693693a8e3fa09ed6156a5996cdc9a3bb1534"},
+    {file = "scikit_image-0.18.1-cp39-cp39-win32.whl", hash = "sha256:3515b890e771f99bbe1051a0dcfe0fc477da961da933c34f89808a0f1eeb7dc2"},
+    {file = "scikit_image-0.18.1-cp39-cp39-win_amd64.whl", hash = "sha256:5f602779258807d03e72c0a439cfb221f647e628be166fb3594397435f13c76b"},
+]
+scikit-learn = [
+    {file = "scikit-learn-0.24.2.tar.gz", hash = "sha256:d14701a12417930392cd3898e9646cf5670c190b933625ebe7511b1f7d7b8736"},
+    {file = "scikit_learn-0.24.2-cp36-cp36m-macosx_10_13_x86_64.whl", hash = "sha256:d5bf9c863ba4717b3917b5227463ee06860fc43931dc9026747de416c0a10fee"},
+    {file = "scikit_learn-0.24.2-cp36-cp36m-manylinux1_i686.whl", hash = "sha256:5beaeb091071625e83f5905192d8aecde65ba2f26f8b6719845bbf586f7a04a1"},
+    {file = "scikit_learn-0.24.2-cp36-cp36m-manylinux1_x86_64.whl", hash = "sha256:06ffdcaaf81e2a3b1b50c3ac6842cfb13df2d8b737d61f64643ed61da7389cde"},
+    {file = "scikit_learn-0.24.2-cp36-cp36m-manylinux2010_i686.whl", hash = "sha256:fec42690a2eb646b384eafb021c425fab48991587edb412d4db77acc358b27ce"},
+    {file = "scikit_learn-0.24.2-cp36-cp36m-manylinux2010_x86_64.whl", hash = "sha256:5ff3e4e4cf7592d36541edec434e09fb8ab9ba6b47608c4ffe30c9038d301897"},
+    {file = "scikit_learn-0.24.2-cp36-cp36m-manylinux2014_aarch64.whl", hash = "sha256:3cbd734e1aefc7c5080e6b6973fe062f97c26a1cdf1a991037ca196ce1c8f427"},
+    {file = "scikit_learn-0.24.2-cp36-cp36m-win32.whl", hash = "sha256:f74429a07fedb36a03c159332b914e6de757176064f9fed94b5f79ebac07d913"},
+    {file = "scikit_learn-0.24.2-cp36-cp36m-win_amd64.whl", hash = "sha256:dd968a174aa82f3341a615a033fa6a8169e9320cbb46130686562db132d7f1f0"},
+    {file = "scikit_learn-0.24.2-cp37-cp37m-macosx_10_13_x86_64.whl", hash = "sha256:49ec0b1361da328da9bb7f1a162836028e72556356adeb53342f8fae6b450d47"},
+    {file = "scikit_learn-0.24.2-cp37-cp37m-manylinux1_i686.whl", hash = "sha256:f18c3ed484eeeaa43a0d45dc2efb4d00fc6542ccdcfa2c45d7b635096a2ae534"},
+    {file = "scikit_learn-0.24.2-cp37-cp37m-manylinux1_x86_64.whl", hash = "sha256:cdf24c1b9bbeb4936456b42ac5bd32c60bb194a344951acb6bfb0cddee5439a4"},
+    {file = "scikit_learn-0.24.2-cp37-cp37m-manylinux2010_i686.whl", hash = "sha256:d177fe1ff47cc235942d628d41ee5b1c6930d8f009f1a451c39b5411e8d0d4cf"},
+    {file = "scikit_learn-0.24.2-cp37-cp37m-manylinux2010_x86_64.whl", hash = "sha256:f3ec00f023d84526381ad0c0f2cff982852d035c921bbf8ceb994f4886c00c64"},
+    {file = "scikit_learn-0.24.2-cp37-cp37m-manylinux2014_aarch64.whl", hash = "sha256:ae19ac105cf7ce8c205a46166992fdec88081d6e783ab6e38ecfbe45729f3c39"},
+    {file = "scikit_learn-0.24.2-cp37-cp37m-win32.whl", hash = "sha256:f0ed4483c258fb23150e31b91ea7d25ff8495dba108aea0b0d4206a777705350"},
+    {file = "scikit_learn-0.24.2-cp37-cp37m-win_amd64.whl", hash = "sha256:39b7e3b71bcb1fe46397185d6c1a5db1c441e71c23c91a31e7ad8cc3f7305f9a"},
+    {file = "scikit_learn-0.24.2-cp38-cp38-macosx_10_13_x86_64.whl", hash = "sha256:90a297330f608adeb4d2e9786c6fda395d3150739deb3d42a86d9a4c2d15bc1d"},
+    {file = "scikit_learn-0.24.2-cp38-cp38-manylinux1_i686.whl", hash = "sha256:f1d2108e770907540b5248977e4cff9ffaf0f73d0d13445ee938df06ca7579c6"},
+    {file = "scikit_learn-0.24.2-cp38-cp38-manylinux1_x86_64.whl", hash = "sha256:1eec963fe9ffc827442c2e9333227c4d49749a44e592f305398c1db5c1563393"},
+    {file = "scikit_learn-0.24.2-cp38-cp38-manylinux2010_i686.whl", hash = "sha256:2db429090b98045d71218a9ba913cc9b3fe78e0ba0b6b647d8748bc6d5a44080"},
+    {file = "scikit_learn-0.24.2-cp38-cp38-manylinux2010_x86_64.whl", hash = "sha256:62214d2954377fcf3f31ec867dd4e436df80121e7a32947a0b3244f58f45e455"},
+    {file = "scikit_learn-0.24.2-cp38-cp38-manylinux2014_aarch64.whl", hash = "sha256:8fac72b9688176922f9f54fda1ba5f7ffd28cbeb9aad282760186e8ceba9139a"},
+    {file = "scikit_learn-0.24.2-cp38-cp38-win32.whl", hash = "sha256:ae426e3a52842c6b6d77d00f906b6031c8c2cfdfabd6af7511bb4bc9a68d720e"},
+    {file = "scikit_learn-0.24.2-cp38-cp38-win_amd64.whl", hash = "sha256:038f4e9d6ef10e1f3fe82addc3a14735c299866eb10f2c77c090410904828312"},
+    {file = "scikit_learn-0.24.2-cp39-cp39-macosx_10_13_x86_64.whl", hash = "sha256:48f273836e19901ba2beecd919f7b352f09310ce67c762f6e53bc6b81cacf1f0"},
+    {file = "scikit_learn-0.24.2-cp39-cp39-manylinux1_i686.whl", hash = "sha256:a2a47449093dcf70babc930beba2ca0423cb7df2fa5fd76be5260703d67fa574"},
+    {file = "scikit_learn-0.24.2-cp39-cp39-manylinux1_x86_64.whl", hash = "sha256:0e71ce9c7cbc20f6f8b860107ce15114da26e8675238b4b82b7e7cd37ca0c087"},
+    {file = "scikit_learn-0.24.2-cp39-cp39-manylinux2010_i686.whl", hash = "sha256:2754c85b2287333f9719db7f23fb7e357f436deed512db3417a02bf6f2830aa5"},
+    {file = "scikit_learn-0.24.2-cp39-cp39-manylinux2010_x86_64.whl", hash = "sha256:7be1b88c23cfac46e06404582215a917017cd2edaa2e4d40abe6aaff5458f24b"},
+    {file = "scikit_learn-0.24.2-cp39-cp39-manylinux2014_aarch64.whl", hash = "sha256:4e6198675a6f9d333774671bd536668680eea78e2e81c0b19e57224f58d17f37"},
+    {file = "scikit_learn-0.24.2-cp39-cp39-win32.whl", hash = "sha256:cbdb0b3db99dd1d5f69d31b4234367d55475add31df4d84a3bd690ef017b55e2"},
+    {file = "scikit_learn-0.24.2-cp39-cp39-win_amd64.whl", hash = "sha256:40556bea1ef26ef54bc678d00cf138a63069144a0b5f3a436eecd8f3468b903e"},
+]
+scipy = [
+    {file = "scipy-1.6.1-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:a15a1f3fc0abff33e792d6049161b7795909b40b97c6cc2934ed54384017ab76"},
+    {file = "scipy-1.6.1-cp37-cp37m-manylinux1_i686.whl", hash = "sha256:e79570979ccdc3d165456dd62041d9556fb9733b86b4b6d818af7a0afc15f092"},
+    {file = "scipy-1.6.1-cp37-cp37m-manylinux1_x86_64.whl", hash = "sha256:a423533c55fec61456dedee7b6ee7dce0bb6bfa395424ea374d25afa262be261"},
+    {file = "scipy-1.6.1-cp37-cp37m-manylinux2014_aarch64.whl", hash = "sha256:33d6b7df40d197bdd3049d64e8e680227151673465e5d85723b3b8f6b15a6ced"},
+    {file = "scipy-1.6.1-cp37-cp37m-win32.whl", hash = "sha256:6725e3fbb47da428794f243864f2297462e9ee448297c93ed1dcbc44335feb78"},
+    {file = "scipy-1.6.1-cp37-cp37m-win_amd64.whl", hash = "sha256:5fa9c6530b1661f1370bcd332a1e62ca7881785cc0f80c0d559b636567fab63c"},
+    {file = "scipy-1.6.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:bd50daf727f7c195e26f27467c85ce653d41df4358a25b32434a50d8870fc519"},
+    {file = "scipy-1.6.1-cp38-cp38-manylinux1_i686.whl", hash = "sha256:f46dd15335e8a320b0fb4685f58b7471702234cba8bb3442b69a3e1dc329c345"},
+    {file = "scipy-1.6.1-cp38-cp38-manylinux1_x86_64.whl", hash = "sha256:0e5b0ccf63155d90da576edd2768b66fb276446c371b73841e3503be1d63fb5d"},
+    {file = "scipy-1.6.1-cp38-cp38-manylinux2014_aarch64.whl", hash = "sha256:2481efbb3740977e3c831edfd0bd9867be26387cacf24eb5e366a6a374d3d00d"},
+    {file = "scipy-1.6.1-cp38-cp38-win32.whl", hash = "sha256:68cb4c424112cd4be886b4d979c5497fba190714085f46b8ae67a5e4416c32b4"},
+    {file = "scipy-1.6.1-cp38-cp38-win_amd64.whl", hash = "sha256:5f331eeed0297232d2e6eea51b54e8278ed8bb10b099f69c44e2558c090d06bf"},
+    {file = "scipy-1.6.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:0c8a51d33556bf70367452d4d601d1742c0e806cd0194785914daf19775f0e67"},
+    {file = "scipy-1.6.1-cp39-cp39-manylinux1_i686.whl", hash = "sha256:83bf7c16245c15bc58ee76c5418e46ea1811edcc2e2b03041b804e46084ab627"},
+    {file = "scipy-1.6.1-cp39-cp39-manylinux1_x86_64.whl", hash = "sha256:794e768cc5f779736593046c9714e0f3a5940bc6dcc1dba885ad64cbfb28e9f0"},
+    {file = "scipy-1.6.1-cp39-cp39-manylinux2014_aarch64.whl", hash = "sha256:5da5471aed911fe7e52b86bf9ea32fb55ae93e2f0fac66c32e58897cfb02fa07"},
+    {file = "scipy-1.6.1-cp39-cp39-win32.whl", hash = "sha256:8e403a337749ed40af60e537cc4d4c03febddcc56cd26e774c9b1b600a70d3e4"},
+    {file = "scipy-1.6.1-cp39-cp39-win_amd64.whl", hash = "sha256:a5193a098ae9f29af283dcf0041f762601faf2e595c0db1da929875b7570353f"},
+    {file = "scipy-1.6.1.tar.gz", hash = "sha256:c4fceb864890b6168e79b0e714c585dbe2fd4222768ee90bc1aa0f8218691b11"},
+]
 six = [
     {file = "six-1.15.0-py2.py3-none-any.whl", hash = "sha256:8b74bedcbbbaca38ff6d7491d76f2b06b3592611af620f8426e82dddb04a5ced"},
     {file = "six-1.15.0.tar.gz", hash = "sha256:30639c035cdb23534cd4aa2dd52c3bf48f06e5f4a941509c8bafd8ce11080259"},
@@ -1166,6 +1583,14 @@ sphinxcontrib-serializinghtml = [
     {file = "sphinxcontrib-serializinghtml-1.1.4.tar.gz", hash = "sha256:eaa0eccc86e982a9b939b2b82d12cc5d013385ba5eadcc7e4fed23f4405f77bc"},
     {file = "sphinxcontrib_serializinghtml-1.1.4-py2.py3-none-any.whl", hash = "sha256:f242a81d423f59617a8e5cf16f5d4d74e28ee9a66f9e5b637a18082991db5a9a"},
 ]
+threadpoolctl = [
+    {file = "threadpoolctl-2.1.0-py3-none-any.whl", hash = "sha256:38b74ca20ff3bb42caca8b00055111d74159ee95c4370882bbff2b93d24da725"},
+    {file = "threadpoolctl-2.1.0.tar.gz", hash = "sha256:ddc57c96a38beb63db45d6c159b5ab07b6bced12c45a1f07b2b92f272aebfa6b"},
+]
+tifffile = [
+    {file = "tifffile-2021.4.8-py3-none-any.whl", hash = "sha256:1cfc55f5b728e200142580a7bf108b72775c4097d007b4111876559fa1fb7432"},
+    {file = "tifffile-2021.4.8.tar.gz", hash = "sha256:55aa8baad38e1567c9fe450fff52160e4a21294a612f241c5e414da80f87209b"},
+]
 toml = [
     {file = "toml-0.10.2-py2.py3-none-any.whl", hash = "sha256:806143ae5bfb6a3c6e736a764057db0e6a0e05e338b5630894a5f779cabb4f9b"},
     {file = "toml-0.10.2.tar.gz", hash = "sha256:b3bda1d108d5dd99f4a20d24d9c348e91c4db7ab1b749200bded2f839ccbe68f"},
@@ -1213,6 +1638,10 @@ tornado = [
     {file = "tornado-6.1-cp39-cp39-win_amd64.whl", hash = "sha256:548430be2740e327b3fe0201abe471f314741efcb0067ec4f2d7dcfb4825f3e4"},
     {file = "tornado-6.1.tar.gz", hash = "sha256:33c6e81d7bd55b468d2e793517c909b139960b6c790a60b7991b9b6b76fb9791"},
 ]
+tqdm = [
+    {file = "tqdm-4.60.0-py2.py3-none-any.whl", hash = "sha256:daec693491c52e9498632dfbe9ccfc4882a557f5fa08982db1b4d3adbe0887c3"},
+    {file = "tqdm-4.60.0.tar.gz", hash = "sha256:ebdebdb95e3477ceea267decfc0784859aa3df3e27e22d23b83e9b272bf157ae"},
+]
 traitlets = [
     {file = "traitlets-5.0.5-py3-none-any.whl", hash = "sha256:69ff3f9d5351f31a7ad80443c2674b7099df13cc41fc5fa6e2f6d3b0330b0426"},
     {file = "traitlets-5.0.5.tar.gz", hash = "sha256:178f4ce988f69189f7e523337a3e11d91c786ded9360174a3d9ca83e79bc5396"},
diff --git a/pyproject.toml b/pyproject.toml
index 4e6e49ec..f5a1abca 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -32,9 +32,14 @@ classifiers = [
 
 [tool.poetry.dependencies]
 python = "^3.7"
+matplotlib = "^3.4.1"
+opencv-python-headless = "^4.5.1"
+scikit-image = "^0.18.1"
+scikit-learn = "^0.24.2"
 smqtk-core = "^0.17.0"
 smqtk-dataprovider = "^0.15.0"
 smqtk-descriptors = "^0.15.0"
+tqdm = "^4.60.0"
 
 [tool.poetry.dev-dependencies]
 # CI
diff --git a/xaitk_saliency/impls/saliency/logit_sal.py b/xaitk_saliency/impls/saliency/logit_sal.py
index 41173eda..673547e3 100644
--- a/xaitk_saliency/impls/saliency/logit_sal.py
+++ b/xaitk_saliency/impls/saliency/logit_sal.py
@@ -1,100 +1,82 @@
-from PIL import Image
+import io
+import logging
+from typing import Any, Dict, Iterable, Sequence, Tuple, Type
+
+import PIL.Image
 import numpy as np
-import os
-from sklearn.metrics.pairwise import euclidean_distances
-import six
-from smqtk.algorithms.relevancy_index import get_relevancy_index_impls
-from smqtk.algorithms import get_descriptor_generator_impls
-from smqtk.algorithms import SmqtkAlgorithm
-from smqtk.algorithms.saliency import SaliencyBlackbox,ImageSaliencyAugmenter
-from smqtk.representation.data_element.memory_element import DataMemoryElement
-from smqtk.utils import plugin
+from smqtk_core.configuration import (
+    from_config_dict,
+    to_config_dict,
+)
+from smqtk_dataprovider.impls.data_element.memory import DataMemoryElement
+from smqtk_descriptors import DescriptorElement, DescriptorGenerator
+
+from xaitk_saliency.interfaces.saliency import (
+    ImageSaliencyAugmenter,
+    SaliencyBlackbox,
+)
+from xaitk_saliency.utils.masking import generate_masked_images
+
+
+LOG = logging.getLogger(__name__)
+
 
-    
-class Logit_SaliencyBlackbox (SaliencyBlackbox):
+class LogitSaliencyBlackbox(SaliencyBlackbox):
     """
-    Logit_SaliencyBlackbox function yields some floating point scalar value
-    for a given masked base image descriptor element that signifies the 
+    This implementation yields some floating point scalar value
+    for a given masked base image descriptor element that signifies the
     difference between the confidence value of the query image and masked
     image descriptors, used by class implementations of
     'ImageSaliencyMapGenerator'.
     """
 
-    def __init__(self, pos_descriptors, neg_descriptors, rel_index, T_descr):
+    def __init__(
+        self,
+        pos_descriptors: Sequence[DescriptorElement],
+        neg_descriptors: Sequence[DescriptorElement],
+        rel_index: Any,  # "RelevancyIndex"; TODO: switch to RankRelevancy
+        T_descr: DescriptorElement
+    ):
+        """
+        :param pos_descriptors: is a set of positive descriptors.
+        :param neg_descriptors: is a set of negative descriptors.
+        :param rel_index:
+            Plugin implementation of the algorithms used to generate relevance
+            index used to rank images.
+        :param T_descr:
+            Base image feature descriptor
         """
-        :param list[smqtk.representation.DescriptorMemoryElement] 
-            pos_descriptors: is a set of positive descriptors.
-        :param list[smqtk.representation.DescriptorMemoryElement]
-               neg_descriptors: is a set of negative descriptors.
-        :param rel_index: Plugin implementation of the algorithms used to 
-        generate relevance index used to rank images
-        :type: A new instance of a class implementing the
-            ``RelevancyIndex`` class.
-        :param smqtk.representation.DescriptorElement
-              T_descr: Base image feature descriptor  
-        """
-    
         self.pos_de = pos_descriptors
-        self.neg_de =  neg_descriptors
+        self.neg_de = neg_descriptors
         self.ADJs = (self.pos_de, self.neg_de)
         self.rel_index = rel_index
         self.T_descr = T_descr
-    
-    @classmethod
-    def is_usable(self):
-        """
-        Check whether this implementation is available for use.
-        Required valid presence of Ajudications and a new instance of 
-        a class implementing the ``RelevancyIndex`` class.
-        :return: Boolean determination of whether implementation is usable.
-        :rtype: bool
-        """
 
-        valid = True
-        return valid
-        
     @classmethod
-    def from_iqr_session(cls, iqrs, descr_gen, base_image): 
-        """
-        Create an ``SaliencyBlackbox`` instance from iqrs session, descriptor
-        generator and base_image.
-        :param smqtk.iqr.IqrSession iqrs:`smqtk.iqr.IqrSession` instance.
-        :param smqtk.algorithms.DescriptorGenerator descr_gen:
-            The descriptor generator used by smqtk.
-        :param PIL.Image base_image: The Base image for which we 
-           need to calculate a saliency map.
-        :return: A new instance of a class implementing the
-            ``SaliencyBlackbox`` class.
-        :rtype: SaliencyBlackbox
-        """
-
+    def from_iqr_session(
+        cls: Type["LogitSaliencyBlackbox"],
+        iqrs: Any,  # "IqrSession"
+        descr_generator: DescriptorGenerator,
+        base_image: PIL.Image.Image
+    ) -> "LogitSaliencyBlackbox":
         assert iqrs
-        pos = list(iqrs.positive_descriptors |\
-          iqrs.external_positive_descriptors)
-        neg = list(iqrs.negative_descriptors |\
-          iqrs.external_negative_descriptors)
-        rel_index = plugin.from_plugin_config\
-         (plugin.to_plugin_config(iqrs.rel_index),\
-          get_relevancy_index_impls())
-        buff = six.BytesIO()
+        pos = list(iqrs.positive_descriptors | iqrs.external_positive_descriptors)
+        neg = list(iqrs.negative_descriptors |
+                   iqrs.external_negative_descriptors)
+        # TODO: This requires the relevancy package breakout for the types used
+        #       here.
+        rel_index: Any = from_config_dict(
+            to_config_dict(iqrs.rel_index),
+            []  # RelevancyIndex.get_impls()
+        )
+        buff = io.BytesIO()
         base_image.save(buff, format="bmp")
         de = DataMemoryElement(buff.getvalue(),
                                content_type='image/bmp')
-        T_descr=descr_gen.compute_descriptor(de)
-        return Logit_SaliencyBlackbox(pos, neg, rel_index, T_descr)
+        T_descr = descr_generator.generate_one_element(de)
+        return LogitSaliencyBlackbox(pos, neg, rel_index, T_descr)
 
-    def get_config(self):
-        """
-        Returns a JSON-compliant dictionary that could be passed to 
-        the class's ``from_config`` method to produce an instance 
-        with identical configuration.
-        In the common case, this involves naming the keys of the dictionary
-        based on the initialization argument names as if it were to be passed
-        to the constructor via dictionary expansion.
-        :return: JSON type compliant configuration dictionary.
-        :rtype: dict
-        """
-        
+    def get_config(self) -> Dict[str, Any]:
         return {
             'pos_descriptors': self.pos_de,
             'neg_descriptors': self.neg_de,
@@ -102,102 +84,74 @@ def get_config(self):
             'T_descr': self.T_descr,
         }
 
-    def transform(self, descriptors):
-        """
-        Transform some descriptor element into a saliency scalar.
-        :param Iterable[smqtk.representation.DescriptorElement] 
-           descriptors:Descriptor of augmentations to get their scalar value.
-        :return: The saliency value for the given descriptor.
-        :rtype: numpy.ndarray[float]
-        """
-        
-        descriptors_list=list(descriptors)
-        rel_train_set=[single_descr for single_descr in descriptors_list]
+    def transform(self, descriptors: Iterable[DescriptorElement]) -> np.ndarray:
+        descriptors_list = list(descriptors)
+        rel_train_set = [single_descr for single_descr in descriptors_list]
         rel_train_set.append(self.T_descr)
         self.rel_index.build_index(rel_train_set)
-        RI_scores=self.rel_index.rank(*self.ADJs) 
+        RI_scores = self.rel_index.rank(*self.ADJs)
         diff = np.ones(len(descriptors_list))
         base_RI = RI_scores[rel_train_set[-1]]
         for i in range(len(descriptors_list)):
-            diff[i]= RI_scores[rel_train_set[i]] - base_RI
+            diff[i] = RI_scores[rel_train_set[i]] - base_RI
         return diff
 
 
-class Logit_ImageSaliencyAugmenter(ImageSaliencyAugmenter):
-    """
-    Robust Augmenter that yields a number of augmentations 
-    of an input image, as well as preserved-area masks,
-    used for use in saliency map generation. 
+class LogitImageSaliencyAugmenter(ImageSaliencyAugmenter):
     """
+    Robust Augmenter that yields a number of augmentations
+    of an input image, as well as preserved-area masks
+    used in saliency map generation.
 
-    def __init__(self, window_size=50, stride=20):
-        """
-        :param int window_size: the block window size 
-        (with value 0, other areas with value 1)
-        :param int stride: the sliding step
-        """
+    :param window_size: the block window size (with value 0, other areas
+        with value 1)
+    :param stride: the sliding step
+    """
 
+    def __init__(self, window_size: int = 50, stride: int = 20):
         self.window_size = window_size
         self.stride = stride
-        self.masks = self.generate_block_masks(window_size=window_size,\
-        stride=stride)
-
-    @classmethod
-    def is_usable(cls):
-        """
-        Check whether this implementation is available for use.
-        Required valid to be True
-        :return:
-            Boolean determination of whether this implementation is usable.
-        :rtype: bool
-        """
-        
-        valid  = True
-        return valid
-    
-    @classmethod
-    def get_config(self):
-        """
-        Return a JSON-compliant dictionary that could be passed to 
-        this class's ``from_config`` method to produce an 
-        instance with identical configuration.
-        In the common case, this involves naming the keys of the dictionary
-        based on the initialization argument names as if it were to be passed
-        to the constructor via dictionary expansion.
-        :return: JSON type compliant configuration dictionary.
-        :rtype: dict
-        """
+        self.masks = self.generate_block_masks(
+            window_size=window_size,
+            stride=stride
+        )
 
+    def get_config(self) -> Dict[str, Any]:
         return {
             'window_size': self.window_size,
             'stride': self.stride,
         }
 
-    def generate_block_masks(self, window_size, stride, image_size=(224,224)):
+    def generate_block_masks(
+        self,
+        window_size: int,
+        stride: int,
+        image_size: Tuple[int, int] = (224, 224)
+    ) -> np.ndarray:
         """
-        Generates sliding window type binary masks used in augment() to 
-        mask an image. The Images are resized to 224x224 to 
-        enable re-use of masks Generating the sliding window style masks.
-        :param int window_size: the block window size 
-        (with value 0, other areas with value 1)
-        :param int stride: the sliding step
-        :param tuple image_size: the mask size which should be the 
-        same to the image size
+        Generates sliding window type binary masks used in augment() to
+        mask an image. The Images are resized to 224x224 to
+        enable re-use of masks.
+
+        :param window_size: the block window size (with value 0, other areas
+            with value 1)
+        :param stride: the sliding step
+        :param image_size: the mask size which should be the same to the image
+            size
+
         :return: the sliding window style masks
-        :rtype: numpy.ndarray
         """
-
         rows = np.arange(0 + stride - window_size, image_size[0], stride)
         cols = np.arange(0 + stride - window_size, image_size[1], stride)
 
         mask_num = len(rows) * len(cols)
-        self._log.debug('mask_num: {}'.format(mask_num))
-        masks = np.ones((mask_num, image_size[0], image_size[1])\
-                                              , dtype=np.float64)
+        LOG.debug('mask_num: {}'.format(mask_num))
+        masks = np.ones((mask_num, image_size[0], image_size[1]),
+                        dtype=np.float64)
         i = 0
         for r in rows:
             for c in cols:
-                if r<0:
+                if r < 0:
                     r1 = 0
                 else:
                     r1 = r
@@ -205,7 +159,7 @@ def generate_block_masks(self, window_size, stride, image_size=(224,224)):
                     r2 = image_size[0]
                 else:
                     r2 = r + window_size
-                if c<0:
+                if c < 0:
                     c1 = 0
                 else:
                     c1 = c
@@ -219,40 +173,10 @@ def generate_block_masks(self, window_size, stride, image_size=(224,224)):
         masks = masks.reshape(mask_shape)
         return masks
 
-    def generate_masked_imgs(self, masks, img):
-        """
-        Apply the masks onto one input image
-        :param numpy.ndarray masks: sliding window type masks in 
-        [1, Height, Weight, 1] format.
-        :param numpy.ndarray img: Original base image
-        :return: List masked images
-        :rtype: List of PIL Images 
-        """
-
-        masked_imgs = []
-        for mask in masks:
-            masked_img = np.multiply(mask, img, casting='unsafe')
-            masked_imgs.append(Image.fromarray(np.uint8(masked_img)))
-        return masked_imgs
-    
-    def augment(self, image_mat):        
-        """
-        Takes in an image matrix and returns its augmented version
-        :param numpy.ndarray image_mat:
-            Image matrix to be augmented.
-        :return: A numpy arrays of augmented image matrices as well as masks
-            that indicate the regions in the augmented images that are
-            unmodified with respect to the input image (preserved regions).
-            Returned augmented images should be in the dimension format
-            [index, height, width [,channel]] with the the same data type as
-            the input image matrix.
-            Returned masks should be in the dimension format
-            [index, height, width,channel] with the boolean data type.
-        :rtype: (PIL.Image, numpy.ndarray)
-        """
-         
-        masked_images = self.generate_masked_imgs(self.masks, image_mat)
-        return (masked_images, self.masks)
-
-SALIENCY_BLACKBOX_CLASS = Logit_SaliencyBlackbox
-IMG_SALIENCY_AUGMENTER_CLASS = Logit_ImageSaliencyAugmenter
+    def augment(
+        self,
+        image_mat: np.ndarray
+    ) -> Tuple[Sequence[PIL.Image.Image], np.ndarray]:
+        img = PIL.Image.fromarray(image_mat)
+        masked_images = generate_masked_images(self.masks, img)
+        return masked_images, self.masks
diff --git a/xaitk_saliency/impls/saliency/rise.py b/xaitk_saliency/impls/saliency/rise.py
index 2b0dae3a..4e161155 100644
--- a/xaitk_saliency/impls/saliency/rise.py
+++ b/xaitk_saliency/impls/saliency/rise.py
@@ -1,14 +1,22 @@
-from ._interface import ImageSaliencyAugmenter, ImageSaliencyMapGenerator
+import io
+from typing import Any, Dict, Generator, Sequence, Tuple
 
-import six
 import numpy as np
+import PIL.Image
 from skimage.transform import resize
-from smqtk.representation.data_element.memory_element import DataMemoryElement
+from smqtk_dataprovider.impls.data_element.memory import DataMemoryElement
+from smqtk_descriptors import DescriptorGenerator
+
+from xaitk_saliency.interfaces.saliency import (
+    ImageSaliencyAugmenter,
+    ImageSaliencyMapGenerator,
+    SaliencyBlackbox,
+)
 
 
 class RISEAugmenter (ImageSaliencyAugmenter):
 
-    def __init__(self, N, s, p1, input_size):
+    def __init__(self, N: int, s: int, p1: float, input_size: Tuple[int, int]):
         """
         Generate a set of random masks to apply to the image.
         :param int N:
@@ -22,7 +30,6 @@ def __init__(self, N, s, p1, input_size):
             Size of the model's input. Smaller masks are upsampled to this resolution
             to be applied to (multiplied with) the input later. E.g. (224, 224)
         """
-
         self.N = N
         self.s = s
         self.p1 = p1
@@ -30,8 +37,8 @@ def __init__(self, N, s, p1, input_size):
         cell_size = np.ceil(np.array(input_size) / s)
         # Upscale factor
         up_size = (s + 1) * cell_size
-        
-        # Generate a set of random grids of small resolution 
+
+        # Generate a set of random grids of small resolution
         grid = np.random.rand(N, s, s) < p1
         grid = grid.astype('float32')
 
@@ -44,35 +51,11 @@ def __init__(self, N, s, p1, input_size):
             # Linear upsampling and random cropping
             masks[i, :, :] = resize(grid[i], up_size, order=1, mode='reflect',
                                     anti_aliasing=False)[x:x + input_size[0], y:y + input_size[1]]
+        # Reshape brings this to: (N, input_size[0], input_size[1], 1)
         self.masks = masks.reshape(-1, *input_size, 1)
         self.input_size = input_size
 
-
-    @classmethod
-    def is_usable(cls):
-        """
-        Check whether this implementation is available for use.
-        Required valid presence of the six and numpy libraries
-        :return:
-            Boolean determination of whether this implementation is usable.
-        :rtype: bool
-        """
-
-        return six and np
-
-
-    def get_config(self):
-        """
-        Return a JSON-compliant dictionary that could be passed to this
-        class's ``from_config`` method to produce an instance with identical
-        configuration.
-        In the common case, this involves naming the keys of the dictionary
-        based on the initialization argument names as if it were to be passed
-        to the constructor via dictionary expansion.
-        :return: JSON type compliant configuration dictionary.
-        :rtype: dict
-        """
-
+    def get_config(self) -> Dict[str, Any]:
         return {
             'N': self.N,
             's': self.s,
@@ -80,24 +63,10 @@ def get_config(self):
             'input_size': self.input_size,
         }
 
-    def augment(self, image_mat):
-        """
-        :param numpy.ndarray image_mat:
-            Image matrix to be augmented.
-
-        :return: A numpy arrays of augmented image matrices as well as masks
-            that indicate the regions in the augmented images that are
-            unmodified with respect to the input image (preserved regions).
-
-            Returned augmented images should be in the dimension format
-            [index, height, width [,channel]] with the the same data type as
-            the input image matrix.
-
-            Returned masks should be in the dimension format
-            [index, height, width] with the boolean data type.
-        :rtype: (numpy.ndarray, numpy.ndarray)
-        """
-
+    def augment(
+        self,
+        image_mat: np.ndarray
+    ) -> Tuple[Sequence[PIL.Image.Image], np.ndarray]:
         # If image is grayscale
         if len(image_mat.shape) == 2:
             image_mat = np.expand_dims(image_mat, 2).repeat(3, axis=2)
@@ -106,104 +75,53 @@ def augment(self, image_mat):
 
 class RISEGenerator (ImageSaliencyMapGenerator):
 
-    def __init__(self, input_size):
+    def __init__(self, input_size: Tuple[int, int]):
         """
         Interface for randomized input sampling based explanations for blackbox models
         https://arxiv.org/abs/1806.07421
         """
-
         self.org_hw = input_size
 
-
-    @classmethod
-    def is_usable(cls):
-        """
-        Check whether this implementation is available for use.
-        Required valid presence of query image feature
-        and base image descriptor
-        :return:
-            Boolean determination of whether this implementation is usable.
-        :rtype: bool
-        """
-
-        return resize and np
-
-
-    def get_config(self):
-        """
-        Return a JSON-compliant dictionary that could be passed to this
-        class's ``from_config`` method to produce an instance with identical
-        configuration.
-        In the common case, this involves naming the keys of the dictionary
-        based on the initialization argument names as if it were to be passed
-        to the constructor via dictionary expansion.
-        :return: JSON type compliant configuration dictionary.
-        :rtype: dict
-        """
-
+    def get_config(self) -> Dict[str, Any]:
         return {
             'input_size': self.org_hw
         }
 
-
-    def generate(self, image_mat, augmenter, descriptor_generator,
-                 blackbox):
-        """
-        Generate an image saliency heat-map matrix given a blackbox's behavior
-        over the descriptions of an augmented base image.
-
-        :param numpy.ndarray image_mat:
-            Numpy image matrix of the format [height, width [,channel]] that is
-            to be augmented.
-
-        :param ImageSaliencyAugmenter augmenter:
-            Augmentation algorithm following
-            the :py:class:`ImageSaliencyAugmenter` interface.
-
-        :param smqtk.algorithms.DescriptorGenerator descriptor_generator:
-            A descriptor generation algorithm following
-            the :py:class:`smqtk.algorithms.DescriptorGenerator` interface.
-
-        :param SaliencyBlackbox blackbox:
-            Blackbox algorithm implementation following
-            the :py:class:`SaliencyBlackbox` interface.
-
-        :return: A :py:class:`numpy.ndarray` matrix of the same [height, width]
-            shape as the input image matrix but of floating-point type within
-            the range of [0,1], where areas of higher value represent more
-            salient regions according to the given blackbox algorithm.
-        :rtype: numpy.ndarray[float]
-        """
-        
+    def generate(
+        self,
+        image_mat: np.ndarray,
+        augmenter: ImageSaliencyAugmenter,
+        descriptor_generator: DescriptorGenerator,
+        blackbox: SaliencyBlackbox
+    ) -> PIL.Image.Image:
         resized_img = resize(image_mat, self.org_hw, order=1)
         masked_images, masks = augmenter.augment(resized_img)
-        
-        idx_to_uuid = []
-        def iter_aug_img_data_elements():
+
+        def iter_aug_img_data_elements() -> Generator[DataMemoryElement, None, None]:
             for a in masked_images:
-                buff = six.BytesIO()
-                (a).save(buff, format="bmp")
+                buff = io.BytesIO()
+                a.save(buff, format="bmp")
                 de = DataMemoryElement(buff.getvalue(),
                                        content_type='image/bmp')
-                idx_to_uuid.append(de.uuid())
                 yield de
 
-        uuid_to_desc = descriptor_generator.compute_descriptor_async(iter_aug_img_data_elements())
-        scores = blackbox.transform((uuid_to_desc[uuid] for uuid in idx_to_uuid))
-        
+        scores = blackbox.transform(descriptor_generator.generate_elements(
+            iter_aug_img_data_elements()
+        ))
+
         # Compute a weighted average of masks w.r.t. the scores
-        saliency_map = np.average(masks, axis=0, weights=scores)
+        saliency_map: np.ndarray = np.average(masks, axis=0, weights=scores)
         saliency_map = np.squeeze(saliency_map)
         # Normalize
         saliency_map /= masks.mean(axis=0)
         # Resize back to the original image shape
         saliency_map = resize(saliency_map, image_mat.shape, order=1)
-        
+
         # At this point the saliency map will be in some range [a, b], 0 <= a <= b <= 1.
-        # The absolute values characterize the average score of the masked image and 
+        # The absolute values characterize the average score of the masked image and
         # therefore have some important information. However, for visualization purposes,
         # the saliency map can be rescaled to [0, 1].
         # saliency_map = (saliency_map - saliency_map.min()) / (saliency_map.max() - saliency_map.min())
-        
+
+        # TODO: This is not returning currently a PIL Image.
         return saliency_map
-            
diff --git a/xaitk_saliency/impls/saliency/sal_gen.py b/xaitk_saliency/impls/saliency/sal_gen.py
index 0c1675ec..0fc5a040 100644
--- a/xaitk_saliency/impls/saliency/sal_gen.py
+++ b/xaitk_saliency/impls/saliency/sal_gen.py
@@ -1,58 +1,50 @@
-from tqdm import tqdm
+import copy
+import io
+from typing import Any, Dict, Generator
+
+import cv2
 from matplotlib import pyplot as plt
-from datetime import datetime
-import six
 import numpy as np
-import PIL
-import cv2
-import copy
-from smqtk.algorithms.saliency import ImageSaliencyMapGenerator
-from smqtk.algorithms.descriptor_generator import DescriptorGenerator
-from smqtk.representation.data_element.file_element import DataFileElement
-from smqtk.representation.data_element.memory_element \
-      import DataMemoryElement
-from skimage.transform import resize
+import PIL.Image
+from smqtk_dataprovider.impls.data_element.memory import DataMemoryElement
+from smqtk_descriptors import DescriptorGenerator
 
+from xaitk_saliency.interfaces.saliency import (
+    ImageSaliencyAugmenter,
+    ImageSaliencyMapGenerator,
+    SaliencyBlackbox,
+)
 
-class Logit_ImageSaliencyMapGenerator(ImageSaliencyMapGenerator):
+
+class LogitImageSaliencyMapGenerator(ImageSaliencyMapGenerator):
     """
     Interface for the method of generation of a saliency map given an image
     augmentation and blackbox algorithms.
     """
-    def __init__(self,threshold=0.2):
-        
+    def __init__(self, threshold: float = 0.2):
         self.thresh = threshold
 
-        self.org_hw = None
-   
-    def get_config(self):
-
+    def get_config(self) -> Dict[str, Any]:
         return {
             'threshold': self.thresh,
         }
 
-    @classmethod
-    def is_usable(cls):
+    def overlay_saliency_map(
+        self,
+        sa_map: np.ndarray,
+        org_img: np.ndarray
+    ) -> PIL.Image.Image:
         """
-        Check whether this implementation is available for use.
-        Required valid presence of svm and svmutil modules
-        :return:
-            Boolean determination of whether this implementation is usable.
-        :rtype: bool
-        """
-        return plt and np
+        Overlay the saliency map on top of original image
 
-    def overlay_saliency_map(self,sa_map, org_img):
-        """
-        overlay the saliency map on top of original image
-        :param numpy.array sa_map: saliency map
-        :param numpy.array org_img: Original image
-        :return: Overlayed image
-        :rtype: PIL.Image
+        :param sa_map: saliency map
+        :param org_img: Original image
+
+        :return: Overlaid image
         """
         plt.switch_backend('agg')
-        height = float(self.org_hw[0])
-        width = float(self.org_hw[1])
+        height = float(org_img.shape[0])
+        width = float(org_img.shape[1])
         fig = plt.figure(dpi=int(height))
         fig.set_size_inches((width / height), 1, forward=False)
         ax = plt.Axes(fig, [0., 0., 1., 1.])
@@ -61,72 +53,64 @@ def overlay_saliency_map(self,sa_map, org_img):
         ax.imshow(org_img)
         ax.imshow(sa_map, cmap='jet', alpha=0.5)
         fig.canvas.draw()
-        np_data = np.fromstring(fig.canvas.tostring_rgb(), \
-           dtype=np.uint8, sep='')
+        np_data = np.fromstring(fig.canvas.tostring_rgb(),
+                                dtype=np.uint8, sep='')
         np_data = np_data.reshape(fig.canvas.get_width_height()[::-1] + (3,))
         im = PIL.Image.fromarray(np_data)
         plt.close()
         return im
 
-    def weighted_avg(self,scalar_vec,masks):
-        masks = masks.reshape(-1,224,224,1)
-        cur_filters = copy.deepcopy(masks[:,:,:,0])
+    def weighted_avg(self, scalar_vec: np.ndarray, masks: np.ndarray) -> np.ndarray:
+        """
+        :param scalar_vec: Array of floats.
+        :param masks: Array of image masks congruent in first-dim size with
+            `scalar_vec`.
+        :return: Weighted saliency mask in floating point.
+        """
+        masks = masks.reshape(-1, 224, 224, 1)
+        cur_filters = copy.deepcopy(masks[:, :, :, 0])
         count = masks.shape[0] - np.sum(cur_filters, axis=0)
 
         for i in range(len(cur_filters)):
-            cur_filters[i] = (1.0 - cur_filters[i]) * np.clip(scalar_vec[i], \
-                       a_min=0.0, a_max=None)
+            cur_filters[i] = (1.0 - cur_filters[i]) * np.clip(scalar_vec[i],
+                                                              a_min=0.0, a_max=None)
         res_sa = np.sum(cur_filters, axis=0) / count
 
-        res_sa = np.clip(res_sa, a_min=((np.max(res_sa)) * self.thresh), \
-                    a_max = None)
+        res_sa = np.clip(res_sa,
+                         a_min=((np.max(res_sa)) * self.thresh),
+                         a_max=None)
         return res_sa
 
-    def generate(self, base_image, augmenter, descriptor_generator,
-                 blackbox):
-        """
-        Generate an image saliency heat-map matrix given a blackbox's behavior
-        over the descriptions of an augmented base image.
-        :param numpy.ndarray base_image:
-             Numpy matrix of the format [height, width [,channel]] that is
-            to be augmented.
-        :param ImageSaliencyAugmenter augmenter:
-            Augmentation algorithm following
-            the :py:class:`ImageSaliencyAugmenter` interface.
-        :param smqtk.algorithms.DescriptorGenerator descriptor_generator:
-            A descriptor generation algorithm following
-            the :py:class:`smqtk.algorithms.DescriptorGenerator` interface.
-        :param SaliencyBlackbox blackbox:
-            Blackbox algorithm implementation following
-            the :py:class:`SaliencyBlackbox` interface.
-        :return: A :py:class:`PIL.Image` of the same [height, width]
-            shape as the input image matrix but of floating-point type within
-            the range of [0,1], where areas of higher value represent more
-            salient regions according to the given blackbox algorithm.
-        :rtype: PIL.Image
-        """
+    def generate(
+        self,
+        image_mat: np.ndarray,
+        augmenter: ImageSaliencyAugmenter,
+        descriptor_generator: DescriptorGenerator,
+        blackbox: SaliencyBlackbox
+    ) -> PIL.Image.Image:
+        # [width, height] shape of the input image matrix
+        org_wh = tuple(image_mat.shape[:2][::-1])
+
+        image_mat_resized = cv2.resize(image_mat,
+                                       (224, 224),
+                                       interpolation=cv2.INTER_LINEAR)
+        augs, masks = augmenter.augment(image_mat_resized)
 
-        self.org_hw = np.shape(base_image)[0:2]
-        base_image_resized = cv2.resize(base_image,(224,224) \
-                     ,interpolation=cv2.INTER_LINEAR)
-        augs, masks = augmenter.augment(base_image_resized)
-        idx_to_uuid = []
-        def iter_aug_img_data_elements():
+        def iter_aug_img_data_elements() -> Generator[DataMemoryElement, None, None]:
             for a in augs:
-               buff = six.BytesIO()
-               (a).save(buff, format="bmp")
-               de = DataMemoryElement(buff.getvalue(),
-                                   content_type='image/bmp')
-               idx_to_uuid.append(de.uuid())
-               yield de
-        uuid_to_desc = descriptor_generator.compute_descriptor_async \
-                    (iter_aug_img_data_elements())
-        scalar_vec = blackbox.transform( \
-                     (uuid_to_desc[uuid] for uuid in idx_to_uuid))
-        final_sal_map = self.weighted_avg(scalar_vec,masks)
-        final_sal_map = cv2.resize(final_sal_map, \
-         (self.org_hw[1], self.org_hw[0]),interpolation=cv2.INTER_LINEAR)
-        sal_map_ret = self.overlay_saliency_map(final_sal_map,base_image)
-        return sal_map_ret
+                buff = io.BytesIO()
+                a.save(buff, format="bmp")
+                de = DataMemoryElement(buff.getvalue(),
+                                       content_type='image/bmp')
+                yield de
 
-IMG_SALIENCY_GENERATOR_CLASS=Logit_ImageSaliencyMapGenerator
+        scalar_vec = blackbox.transform(descriptor_generator.generate_elements(
+            iter_aug_img_data_elements()
+        ))
+        final_sal_map = self.weighted_avg(scalar_vec, masks)
+        final_sal_map = cv2.resize(
+            final_sal_map, org_wh,
+            interpolation=cv2.INTER_LINEAR
+        )
+        sal_map_ret = self.overlay_saliency_map(final_sal_map, image_mat)
+        return sal_map_ret
diff --git a/xaitk_saliency/impls/saliency/sbsm.py b/xaitk_saliency/impls/saliency/sbsm.py
index 20532432..2a1c203b 100644
--- a/xaitk_saliency/impls/saliency/sbsm.py
+++ b/xaitk_saliency/impls/saliency/sbsm.py
@@ -1,236 +1,164 @@
-from PIL import Image
+import io
+import logging
 import os
+from typing import Any, Dict, Iterable, Sequence, Tuple, Type
+
+import numpy as np
+import PIL.Image
 from sklearn.metrics.pairwise import euclidean_distances
-import six
-from smqtk.algorithms import SmqtkAlgorithm
-from smqtk.algorithms.saliency import SaliencyBlackbox,ImageSaliencyAugmenter
-from smqtk.representation.data_element.memory_element import DataMemoryElement
-
-try:
-    import numpy as np
-    from tqdm import tqdm
-    import copy
-except ImportError as ex:
-    logging.getLogger(__name__).warning("Failed to import numpy/tqdm \
-         /copy module: %s", str(ex))
-    np = None
-    tqdm = None
-    copy = None
-
-
-class SBSM_SaliencyBlackbox (SaliencyBlackbox):
+from smqtk_dataprovider.impls.data_element.memory import DataMemoryElement
+from smqtk_descriptors import DescriptorElement, DescriptorGenerator
+from tqdm import tqdm
+
+from xaitk_saliency.interfaces.saliency import (
+    ImageSaliencyAugmenter,
+    SaliencyBlackbox,
+)
+from xaitk_saliency.utils.masking import generate_masked_images
+
+
+LOG = logging.getLogger(__name__)
+
+
+class SBSMSaliencyBlackbox(SaliencyBlackbox):
     """
-    SBSM_SaliencyBlackbox function that yields some floating point 
-    scalar value for a given masked base image descriptor element that 
+    SBSM_SaliencyBlackbox function that yields some floating point
+    scalar value for a given masked base image descriptor element that
     signifies the proximity between the query image and masked image
     descriptors, used by class implementations of 'ImageSaliencyMapGenerator'.
-    """
 
-    def __init__(self, query_f, base_descr):
-        """
-        :param smqtk.representation.DescriptorElement query_f:
-            Feature of query image.
-        :param smqtk.representation.DescriptorElement base_descr: 
-            Base image descriptor
-        """    
+    :param query_f: Feature of query image.
+    :param base_descr: Base image descriptor
+    """
 
-        self.query_f = query_f        
+    def __init__(self, query_f: np.ndarray, base_descr: np.ndarray):
+        self.query_f = query_f
         self.base_descr = base_descr
 
     @classmethod
-    def is_usable(cls):
-        """
-        Check whether this implementation is available for use.
-        Required valid presence of query image feature 
-        and base image descriptor
-        :return:
-            Boolean determination of whether this implementation is usable.
-        :rtype: bool
-        """
-
-        return True
-                
-    @classmethod
-    def from_iqr_session(cls, iqrs, descr_gen, base_image): 
-        """
-        Create an ``SaliencyBlackbox`` instance from iqrs session,
-        descriptor generator and base_image.
-        :param smqtk.iqr.IqrSession iqrs:`smqtk.iqr.IqrSession` instance.
-        :param smqtk.algorithms.DescriptorGenerator descr_gen:
-           The descriptor generator used by smqtk.
-        :param PIL.Image base_image: The Base image for which we need to
-        calculate a saliency map.
-        :return: A new instance of a class implementing the
-            ``SaliencyBlackbox`` class.
-        :rtype: SaliencyBlackbox
-        """
-
+    def from_iqr_session(
+        cls: Type["SaliencyBlackbox"],
+        iqrs: Any,  # "IqrSession"
+        descr_generator: DescriptorGenerator,
+        base_image: PIL.Image.Image
+    ) -> "SaliencyBlackbox":
         assert iqrs
         if (len(iqrs.external_positive_descriptors)) != 1:
             raise ValueError("Saliency generation in class``{}`` "
-                                  "only supports one query sample, but"
-                                  "recieved more than 1 external positive."
-                                  .format(cls.__name__))
-        query_f = [ext_pos.vector() for ext_pos in \
-           iqrs.external_positive_descriptors]
-        buff = six.BytesIO()
+                             "only supports one query sample, but"
+                             "recieved more than 1 external positive."
+                             .format(cls.__name__))
+        # Implementation only ever considers the first query descriptor, so
+        # we reflect such a selection here based on the previous incorrect
+        # impl.
+        # query_f = [ext_pos.vector() for ext_pos in \
+        #    iqrs.external_positive_descriptors]
+        query_f = list(iqrs.external_positive_descriptors)[0].vector()
+        buff = io.BytesIO()
         base_image.save(buff, format="bmp")
         de = DataMemoryElement(buff.getvalue(),
                                content_type='image/bmp')
-        base_descr = descr_gen.compute_descriptor(de)
-        return SBSM_SaliencyBlackbox(query_f, base_descr)
-       
-    def get_config(self):
-        """
-        Return a JSON-compliant dictionary that could be passed to this
-        class's ``from_config`` method to produce an instance with identical
-        configuration.
-        In the common case, this involves naming the keys of the dictionary
-        based on the initialization argument names as if it were to be passed
-        to the constructor via dictionary expansion.
-        :return: JSON type compliant configuration dictionary.
-        :rtype: dict
-        """
+        base_descr = descr_generator.generate_one_element(de)
+        base_descr_vec = base_descr.vector()
+        assert base_descr_vec is not None  # we literally just generated it...
+        return SBSMSaliencyBlackbox(query_f, base_descr_vec)
 
+    def get_config(self) -> Dict[str, Any]:
         return {
             'query_f': self.query_f,
             'base_descr': self.base_descr,
         }
 
-    def transform(self, descriptors):
-        """
-        Transform some descriptor element into a saliency scalar.
-        :param collections.Iterable[smqtk.representation.DescriptorElement]
-          descriptors: Descriptor to get the saliency of.
-        :return: The saliency value for the given descriptor.
-        :rtype: numpy.ndarray[float]
-        """
-     
-        org_dis = abs(euclidean_distances(self.query_f[0].reshape(1, -1)
-                                ,(self.base_descr.vector()).reshape(1, -1)))
+    def transform(self, descriptors: Iterable[DescriptorElement]) -> np.ndarray:
+        org_dis = abs(euclidean_distances(self.query_f.reshape(1, -1),
+                                          self.base_descr.reshape(1, -1)))
         descriptors_list = list(descriptors)
+        vector_list = [d.vector() for d in descriptors_list]
+        if None in vector_list:
+            raise RuntimeError("One or more input descriptor elements did not "
+                               "refer to an actual vector (some None-valued).")
         diff = np.ones(len(descriptors_list))
         query_f_reshaped = self.query_f[0].reshape(1, -1)
-        for i in range(len(descriptors_list)):
-            diff[i] = max(abs(euclidean_distances \
-              (descriptors_list[i].vector().reshape(1, -1), \
-               query_f_reshaped))-org_dis,0)
+        for i in range(len(vector_list)):
+            v = vector_list[i]
+            assert v is not None
+            diff[i] = max(abs(euclidean_distances(
+                v.reshape(1, -1),
+                query_f_reshaped
+            )) - org_dis, 0)
         return diff
 
 
-class SBSM_ImageSaliencyAugmenter (ImageSaliencyAugmenter):
+class SBSMImageSaliencyAugmenter(ImageSaliencyAugmenter):
     """
     Algorithm that yields a number of augmentations of an input image,
     as well as preserved-area masks, used for use in saliency map generation.
-    For a more robust augmenter that can work for different image_sizes 
+    For a more robust augmenter that can work for different image_sizes
     use implementation class 'Logit_ImageSaliencyAugmenter'.
-    """
 
-    def __init__(self, window_size=20, stride=4):
-        """
-        :param int window_size: the block window size 
-        (with value 0, other areas with value 1)
-        :param int stride: the sliding step
-        """
+    :param window_size: the block window size (with value 0, other areas
+        with value 1)
+    :param stride: the sliding step
+    """
 
+    def __init__(self, window_size: int = 20, stride: int = 4):
         self.window_size = window_size
         self.stride = stride
-        self.masks = self.generate_block_masks(self.window_size,self.stride)
-
-    @classmethod
-    def is_usable(cls):
-        """
-        Check whether this implementation is available for use.
-        Required valid presence of tqdm and copy modules
-        :return:
-            Boolean determination of whether this implementation is usable.
-        :rtype: bool
-        """
-
-        valid = (tqdm is not None) and (copy is not None)
-        if not valid:
-            cls.get_logger().debug("tqdm or copy python \
-                  module cannot be imported")
-        return valid
-
-    def get_config(self):
-        """
-        Return a JSON-compliant dictionary that could be passed to
-        this class's ``from_config`` method to produce an instance 
-        with identical configuration.
-        In the common case, this involves naming the keys of the dictionary
-        based on the initialization argument names as if it were to be passed
-        to the constructor via dictionary expansion.
-        :return: JSON type compliant configuration dictionary.
-        :rtype: dict
-        """
+        self.masks = self.generate_block_masks(self.window_size, self.stride)
 
+    def get_config(self) -> Dict[str, Any]:
         return {
             'window_size': self.window_size,
             'stride': self.stride,
         }
 
-    def generate_block_masks(self, window_size, stride, image_size=(224, 224)):
+    def generate_block_masks(
+        self,
+        window_size: int,
+        stride: int,
+        image_size: Tuple[int, int] = (224, 224)
+    ) -> np.ndarray:
         """
-        Generates sliding window type binary masks used in augment() to mask
-        an image. The Images are resized to 224x224 to enable re-use of masks
-        Generating the sliding window style masks
-        :param int window_size: the block window size
-          (with value 0, other areas with value 1)
-        :param int stride: the sliding step
-        :param tuple(default: (224, 224)) image_size:The mask size which should
-          be the same to the image size.
+        Generates sliding window type binary masks used in augment() to
+        mask an image. The Images are resized to 224x224 to
+        enable re-use of masks.
+
+        :param window_size: the block window size (with value 0, other areas
+            with value 1)
+        :param stride: the sliding step
+        :param image_size: the mask size which should be the same to the image
+            size
+
         :return: the sliding window style masks
-        :rtype: numpy.ndarray
         """
-
-        
-        #The augmenter only supports certain factors of window_size and stride
-        #for a more robust augmenter use Logit_ImageSaliencyAugmenter
-        if (image_size[0]-window_size)%stride != 0: 
+        # The augmenter only supports certain factors of window_size and stride
+        # for a more robust augmenter use LogitImageSaliencyAugmenter
+        if (image_size[0] - window_size) % stride != 0:
             raise ValueError("Change window size and stride to satisfy "
-                              "condition:(image_height-window_size)%stride=0")
+                             "condition:(image_height-window_size)%stride=0")
         if not os.path.isfile('block_mask_{}_{}.npy'.format(window_size, stride)):
-            grid_num_r = ((image_size[0] - window_size) // stride)+1
-            grid_num_c = ((image_size[1] - window_size) // stride)+1
+            grid_num_r = ((image_size[0] - window_size) // stride) + 1
+            grid_num_c = ((image_size[1] - window_size) // stride) + 1
             mask_num = grid_num_r * grid_num_c
             masks = np.ones((mask_num, image_size[0], image_size[1]), dtype=np.float32)
             i = 0
-            for r in tqdm(np.arange(0, image_size[0] - window_size+1, stride), total=grid_num_r, desc="Generating rows"):
-                for c in np.arange(0, image_size[1] - window_size+1, stride):
+            for r in tqdm(np.arange(0, image_size[0] - window_size + 1, stride), total=grid_num_r,
+                          desc="Generating rows"):
+                for c in np.arange(0, image_size[1] - window_size + 1, stride):
                     masks[i, r:r + window_size, c:c + window_size] = 0.0
                     i += 1
             masks = masks.reshape(-1, *image_size)
             masks.tofile('block_mask_{}_{}.npy'.format(window_size, stride))
         else:
-            self._log.debug("Loading masks from file block_mask_{}_{}.npy".format(window_size, stride))
+            LOG.debug("Loading masks from file block_mask_{}_{}.npy".format(window_size, stride))
             masks = np.fromfile('block_mask_{}_{}.npy'.format(window_size, stride),
-                                dtype=np.float32).reshape(-1,  *image_size)
+                                dtype=np.float32).reshape(-1, *image_size)
         return masks
 
-    def generate_masked_imgs(self, masks, img):
-        """
-        Apply the masks onto one input image
-        :param numpy.ndarray masks: sliding window type 
-           masks in [1, Height, Weight, 1] format.
-        :param numpy.ndarray img: Original base image
-        :return: List masked images
-        :rtype: list[PIL.Image] 
-        """
-
-        if (img.ndim == 2):
-            channels = 1 
-        if (img.ndim == 3):
-            channels = 3
-        masked_imgs = []
-        masked_img = copy.deepcopy(img)
-        for cnt,mask in enumerate(masks):
-            for ind in range(channels):
-                masked_img[:,:,ind] = np.multiply(mask, img[:,:,ind])
-            masked_imgs.append(Image.fromarray(np.uint8(masked_img)))
-        return masked_imgs
-        
-    def augment(self, image_mat):
+    def augment(
+        self,
+        image_mat: np.ndarray
+    ) -> Tuple[Sequence[PIL.Image.Image], np.ndarray]:
         """
         Takes in an image matrix and returns its augmented version
         :param numpy.ndarray image_mat:
@@ -245,9 +173,6 @@ def augment(self, image_mat):
             [index, height, width,channel] with the boolean data type.
         :rtype: (PIL.Image, numpy.ndarray)
         """
-         
-        masked_images = self.generate_masked_imgs(self.masks,image_mat)
-        return (masked_images,self.masks)
-
-SALIENCY_BLACKBOX_CLASS=SBSM_SaliencyBlackbox
-IMG_SALIENCY_AUGMENTER_CLASS=SBSM_ImageSaliencyAugmenter
+        img = PIL.Image.fromarray(image_mat)
+        masked_images = generate_masked_images(self.masks, img)
+        return masked_images, self.masks
diff --git a/xaitk_saliency/interfaces/saliency.py b/xaitk_saliency/interfaces/saliency.py
index 57fe02c0..ebdbcccb 100644
--- a/xaitk_saliency/interfaces/saliency.py
+++ b/xaitk_saliency/interfaces/saliency.py
@@ -1,17 +1,28 @@
 import abc
+from typing import Any, Iterable, Sequence, Tuple, Type, TypeVar
 
-from smqtk.algorithms import SmqtkAlgorithm
-from smqtk.representation.data_element.memory_element import DataMemoryElement
+import numpy as np
+import PIL.Image
+from smqtk_core import Plugfigurable
+from smqtk_descriptors import DescriptorElement, DescriptorGenerator
 
 
-class SaliencyBlackbox (SmqtkAlgorithm):
+T = TypeVar("T", bound="SaliencyBlackbox")
+
+
+class SaliencyBlackbox (Plugfigurable):
     """
     Blackbox function that produces some floating point scalar value for a
     given descriptor element.
     """
 
     @classmethod
-    def from_iqr_session(cls, iqrs, descr_generator, base_elem):
+    def from_iqr_session(
+        cls: Type["SaliencyBlackbox"],
+        iqrs: Any,  # "IqrSession"
+        descr_generator: DescriptorGenerator,
+        base_image: PIL.Image.Image
+    ) -> "SaliencyBlackbox":
         """
         Create an ``SaliencyBlackbox`` instance from an
         :class:`smqtk.iqr.IqrSession` instance.
@@ -28,91 +39,97 @@ def from_iqr_session(cls, iqrs, descr_generator, base_elem):
             if a blackbox instance.
         :param descr_generator:
             Descriptor generator instance.
-        :param base_elem:
-            A :py:class:`smqtk.representation.DataElement` representing the
-            data over which a saliency map may be generated.
+        :param base_image:
+            A :py:class:`PIL.Image.Image` over which a saliency map may be
+            generated.
 
         :return: A new instance of of a class implementing the
             ``SaliencyBlackbox`` class.
-        :rtype: SaliencyBlackbox
-
         """
         raise NotImplementedError("The ``from_iqr_session`` classmethod is "
                                   "not implemented for class ``{}``."
                                   .format(cls.__name__))
 
     @abc.abstractmethod
-    def transform(self, descriptors):
+    def transform(self, descriptors: Iterable[DescriptorElement]) -> np.ndarray:
         """
         Transform some number of descriptor elements into a saliency scalar
         values.
 
-        :param collections.Iterable[smqtk.representation.DescriptorElement] descriptors:
+        :param descriptors:
             Descriptors to get the saliency values of.
 
-        :return: The saliency value for the given descriptor.
-        :rtype: numpy.ndarray[float]
+        :return: The floating-point saliency value for the given descriptor.
         """
 
 
-class ImageSaliencyAugmenter (SmqtkAlgorithm):
+class ImageSaliencyAugmenter (Plugfigurable):
     """
     Algorithm that yields a number of augmentations of an input image, as well
     as preserved-area masks, used for use in saliency map generation.
     """
 
     @abc.abstractmethod
-    def augment(self, image_mat):
+    def augment(
+        self,
+        image_mat: np.ndarray
+    ) -> Tuple[Sequence[PIL.Image.Image], np.ndarray]:
         """
-        :param numpy.ndarray image_mat:
-            Image matrix to be augmented.
+        Takes in an image matrix and returns its augmented version
+
+        :param image_mat:
+            Image matrix to be augmented. This should be in the format
+            [height, width [,channel]].
 
-        :return: A numpy arrays of augmented image matrices as well as masks
-            that indicate the regions in the augmented images that are
-            unmodified with respect to the input image (preserved regions).
+        :return: A sequence of augmented images as well as congruently sized
+            array of masks that indicate the regions in the augmented images
+            that are unmodified with respect to the input image (preserved
+            regions).
 
-            Returned augmented images should be in the dimension format
-            [index, height, width [,channel]] with the the same data type as
-            the input image matrix.
+            Returned augmented images should have the same height, width and
+            channel format as the input image matrix.
 
             Returned masks should be in the dimension format
             [index, height, width] with the boolean data type.
-        :rtype: (PIL.Image, numpy.ndarray)
         """
 
 
-class ImageSaliencyMapGenerator (SmqtkAlgorithm):
+class ImageSaliencyMapGenerator (Plugfigurable):
     """
     Interface for the method of generation of a saliency map given an image
     augmentation and blackbox algorithms.
     """
 
     @abc.abstractmethod
-    def generate(self, image_mat, augmenter, descriptor_generator,
-                 blackbox):
+    def generate(
+        self,
+        image_mat: np.ndarray,
+        augmenter: ImageSaliencyAugmenter,
+        descriptor_generator: DescriptorGenerator,
+        blackbox: SaliencyBlackbox
+    ) -> PIL.Image.Image:
         """
         Generate an image saliency heat-map matrix given a blackbox's behavior
         over the descriptions of an augmented base image.
 
-        :param numpy.ndarray image_mat:
+        :param image_mat:
             Numpy image matrix of the format [height, width [,channel]] that is
             to be augmented.
-
-        :param ImageSaliencyAugmenter augmenter:
+        :param augmenter:
             Augmentation algorithm following
             the :py:class:`ImageSaliencyAugmenter` interface.
-
-        :param smqtk.algorithms.DescriptorGenerator descriptor_generator:
+        :param descriptor_generator:
             A descriptor generation algorithm following
             the :py:class:`smqtk.algorithms.DescriptorGenerator` interface.
-
-        :param SaliencyBlackbox blackbox:
+        :param blackbox:
             Blackbox algorithm implementation following
             the :py:class:`SaliencyBlackbox` interface.
 
-        :return: A :py:class:`numpy.ndarray` matrix of the same [height, width]
+        :return: A PIL Image with the same height and width properties as the
+            input image.
+
+            :py:class:`numpy.ndarray` matrix of the same [height, width]
             shape as the input image matrix but of floating-point type within
             the range of [0,1], where areas of higher value represent more
             salient regions according to the given blackbox algorithm.
-        :rtype: numpy.ndarray[float]
         """
diff --git a/xaitk_saliency/py.typed b/xaitk_saliency/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/xaitk_saliency/utils/__init__.py b/xaitk_saliency/utils/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/xaitk_saliency/utils/masking.py b/xaitk_saliency/utils/masking.py
new file mode 100644
index 00000000..4a1acea7
--- /dev/null
+++ b/xaitk_saliency/utils/masking.py
@@ -0,0 +1,30 @@
+from typing import List
+
+import numpy as np
+import PIL.Image
+
+
+def generate_masked_images(
+    masks: np.ndarray,
+    img: PIL.Image.Image
+) -> List[PIL.Image.Image]:
+    """
+    Apply some binary masks onto one common image, generating a number of new
+    images with the masked regions maintained.
+
+    We expect the "mask" matrices and the image to be the same height and
+    width, and be valued in the [0, 1] floating-point range.
+    In the mask matrix, higher values correspond to regions of the image that
+    preserved. E.g. a 0 in the mask will translate to blacking out the
+    corresponding location in the source image.
+
+    :param masks: Mask images in the [N, Height, Weight, 1] shape format.
+    :param img: Original base image
+
+    :return: List of masked images in PIL Image form.
+    """
+    masked_imgs = []
+    for mask in masks:
+        masked_img = np.multiply(mask, img, casting='unsafe')
+        masked_imgs.append(PIL.Image.fromarray(np.uint8(masked_img)))
+    return masked_imgs