PyPuzzle provides python bindings for libpuzzle. Libpuzzle is designed to quickly find visually similar images (gif, png, jpg), even if they have been resized, recompressed, recolored or slightly modified. Since libpuzzle is a C library, PyPuzzle connects it to python.
For more information about libpuzzle, consult its home page.
At present, PyPuzzle has only been tested under Linux and OS X 10.9.
Before compiling, make sure you have header files of libpuzzle and libgd2 installed. See 'COMPILATION' section in the doc of libpuzzle.
For Ubuntu users, you can just install them from source:
apt-get install libgd-dev libpuzzle-dev
Note: In case of missing some image formats support, make sure you have corresponding image libraries installed. For example, libjpeg/libpng, etc.
Then you can install PyPuzzle with pip:
$ pip install pypuzzle
or easy_install:
$ esay_install pypuzzle
In addition, you can compile it from source:
$ python setup.py install
import pypuzzle
puzzle = pypuzzle.Puzzle()
# Get distance between two image files
distance = puzzle.get_distance_from_file("file_1.jpg", "file_2.jpg")
# Get vector from image file
vec_1 = puzzle.get_cvec_from_file("file_1.jpg")
vec_2 = puzzle.get_cvec_from_file("file_2.jpg")
# Get distance between two vectors
distance = puzzle.get_distance_from_cvec(vec_1, vec_2)
# Compress vector
cmp_vec_1 = puzzle.compress_cvec(vec_1)
# Uncompress vector
uncmp_vec_1 = puzzle.uncompress_cvec(cmp_vec_1)
PyPuzzle allows you to change related parameters of the algorithm with set_*() methods:
Puzzle.set_max_width(max_width)
Puzzle.set_max_height(max_height)
Puzzle.set_lambdas(lambdas)
Puzzle.set_p_ratio(p_ratio)
Puzzle.set_noise_cutoff(noise_cutoff)
Puzzle.set_contrast_barrier_for_cropping(barrier)
Puzzle.set_max_cropping_ratio(ratio)
Puzzle.set_autocrop(enable)
Consult original page of libpuzzle for more details.
- Add Python 3 support
- Raise IOError in
get_cvec_from_file
andget_distance_from_file
if error occurred when reading from file.