909 Update LoadImage to use Nibabel as default

docs/source/highlights.md

@@ -81,7 +81,7 @@ For example:
 ```py
 # define a transform chain for pre-processing
 train_transforms = monai.transforms.Compose([
-    LoadNiftid(keys=['image', 'label']),
+    LoadImaged(keys=['image', 'label']),
     RandRotate90d(keys=['image', 'label'], prob=0.2, spatial_axes=[0, 2]),
     ... ...
 ])

monai/apps/datasets.py

@@ -23,7 +23,7 @@
     partition_dataset,
     select_cross_validation_folds,
 )
-from monai.transforms import LoadNiftid, LoadPNGd, Randomizable
+from monai.transforms import LoadImaged, Randomizable
 from monai.utils import ensure_tuple
 
 
@@ -92,7 +92,7 @@ def __init__(
             )
         data = self._generate_data_list(dataset_dir)
         if transform == ():
-            transform = LoadPNGd("image")
+            transform = LoadImaged("image")
         super().__init__(data, transform, cache_num=cache_num, cache_rate=cache_rate, num_workers=num_workers)
 
     def randomize(self, data: Optional[Any] = None) -> None:
@@ -268,7 +268,7 @@ def __init__(
         ]
         self._properties = load_decathlon_properties(os.path.join(dataset_dir, "dataset.json"), property_keys)
         if transform == ():
-            transform = LoadNiftid(["image", "label"])
+            transform = LoadImaged(["image", "label"])
         super().__init__(data, transform, cache_num=cache_num, cache_rate=cache_rate, num_workers=num_workers)
 
     def get_indices(self) -> np.ndarray:

monai/data/image_reader.py

@@ -14,6 +14,7 @@
 from typing import TYPE_CHECKING, Any, Callable, Dict, List, Optional, Sequence, Tuple, Union
 
 import numpy as np
+from torch.utils.data._utils.collate import np_str_obj_array_pattern
 
 from monai.config import KeysCollection
 from monai.data.utils import correct_nifti_header_if_necessary
@@ -80,6 +81,29 @@ def get_data(self, img) -> Tuple[np.ndarray, Dict]:
         raise NotImplementedError(f"Subclass {self.__class__.__name__} must implement this method.")
 
 
+def _copy_compatible_dict(from_dict: Dict, to_dict: Dict):
+    if not isinstance(to_dict, dict):
+        raise ValueError(f"to_dict must be a Dict, got {type(to_dict)}.")
+    if not to_dict:
+        for key in from_dict:
+            datum = from_dict[key]
+            if isinstance(datum, np.ndarray) and np_str_obj_array_pattern.search(datum.dtype.str) is not None:
+                continue
+            to_dict[key] = datum
+    else:
+        affine_key, shape_key = "affine", "spatial_shape"
+        if affine_key in from_dict and not np.allclose(from_dict[affine_key], to_dict[affine_key]):
+            raise RuntimeError(
+                "affine matrix of all images should be the same for channel-wise concatenation. "
+                f"Got {from_dict[affine_key]} and {to_dict[affine_key]}."
+            )
+        if shape_key in from_dict and not np.allclose(from_dict[shape_key], to_dict[shape_key]):
+            raise RuntimeError(
+                "spatial_shape of all images should be the same for channel-wise concatenation. "
+                f"Got {from_dict[shape_key]} and {to_dict[shape_key]}."
+            )
+
+
 class ITKReader(ImageReader):
     """
     Load medical images based on ITK library.
@@ -159,22 +183,15 @@ def get_data(self, img):
 
         """
         img_array: List[np.ndarray] = list()
-        compatible_meta: Dict = None
+        compatible_meta: Dict = {}
 
         for i in ensure_tuple(img):
             header = self._get_meta_dict(i)
             header["original_affine"] = self._get_affine(i)
             header["affine"] = header["original_affine"].copy()
             header["spatial_shape"] = self._get_spatial_shape(i)
             img_array.append(self._get_array_data(i))
-
-            if compatible_meta is None:
-                compatible_meta = header
-            else:
-                if not np.allclose(header["affine"], compatible_meta["affine"]):
-                    raise RuntimeError("affine matrix of all images should be same.")
-                if not np.allclose(header["spatial_shape"], compatible_meta["spatial_shape"]):
-                    raise RuntimeError("spatial_shape of all images should be same.")
+            _copy_compatible_dict(header, compatible_meta)
 
         img_array_ = np.stack(img_array, axis=0) if len(img_array) > 1 else img_array[0]
         return img_array_, compatible_meta
@@ -188,7 +205,7 @@ def _get_meta_dict(self, img) -> Dict:
 
         """
         img_meta_dict = img.GetMetaDataDictionary()
-        meta_dict = dict()
+        meta_dict = {}
         for key in img_meta_dict.GetKeys():
             # ignore deprecated, legacy members that cause issues
             if key.startswith("ITK_original_"):
@@ -220,7 +237,7 @@ def _get_affine(self, img) -> np.ndarray:
         affine[(slice(-1), -1)] = origin
         return affine
 
-    def _get_spatial_shape(self, img) -> Sequence:
+    def _get_spatial_shape(self, img) -> np.ndarray:
         """
         Get the spatial shape of image data, it doesn't contain the channel dim.
 
@@ -230,7 +247,7 @@ def _get_spatial_shape(self, img) -> Sequence:
         """
         shape = list(itk.size(img))
         shape.reverse()
-        return shape
+        return np.asarray(shape)
 
     def _get_array_data(self, img) -> np.ndarray:
         """
@@ -247,17 +264,15 @@ def _get_array_data(self, img) -> np.ndarray:
         channels = img.GetNumberOfComponentsPerPixel()
         if channels == 1:
             return itk.array_view_from_image(img, keep_axes=False)
-        else:
-            # The memory layout of itk.Image has all pixel's channels adjacent
-            # in memory, i.e. R1G1B1R2G2B2R3G3B3. For PyTorch/MONAI, we need
-            # channels to be contiguous, i.e. R1R2R3G1G2G3B1B2B3.
-            arr = itk.array_view_from_image(img, keep_axes=False)
-            dest = list(range(img.ndim))
-            source = dest.copy()
-            end = source.pop()
-            source.insert(0, end)
-            arr_contiguous_channels = np.moveaxis(arr, source, dest)
-            return arr_contiguous_channels
+        # The memory layout of itk.Image has all pixel's channels adjacent
+        # in memory, i.e. R1G1B1R2G2B2R3G3B3. For PyTorch/MONAI, we need
+        # channels to be contiguous, i.e. R1R2R3G1G2G3B1B2B3.
+        arr = itk.array_view_from_image(img, keep_axes=False)
+        dest = list(range(img.ndim))
+        source = dest.copy()
+        end = source.pop()
+        source.insert(0, end)
+        return np.moveaxis(arr, source, dest)
 
 
 class NibabelReader(ImageReader):
@@ -271,9 +286,10 @@ class NibabelReader(ImageReader):
 
     """
 
-    def __init__(self, as_closest_canonical: bool = False, **kwargs):
+    def __init__(self, as_closest_canonical: bool = False, dtype: Optional[np.dtype] = np.float32, **kwargs):
         super().__init__()
         self.as_closest_canonical = as_closest_canonical
+        self.dtype = dtype
         self.kwargs = kwargs
 
     def verify_suffix(self, filename: Union[Sequence[str], str]) -> bool:
@@ -324,26 +340,19 @@ def get_data(self, img):
 
         """
         img_array: List[np.ndarray] = list()
-        compatible_meta: Dict = None
+        compatible_meta: Dict = {}
 
         for i in ensure_tuple(img):
             header = self._get_meta_dict(i)
+            header["affine"] = self._get_affine(i)
             header["original_affine"] = self._get_affine(i)
-            header["affine"] = header["original_affine"].copy()
+            header["as_closest_canonical"] = self.as_closest_canonical
             if self.as_closest_canonical:
                 i = nib.as_closest_canonical(i)
                 header["affine"] = self._get_affine(i)
-            header["as_closest_canonical"] = self.as_closest_canonical
             header["spatial_shape"] = self._get_spatial_shape(i)
             img_array.append(self._get_array_data(i))
-
-            if compatible_meta is None:
-                compatible_meta = header
-            else:
-                if not np.allclose(header["affine"], compatible_meta["affine"]):
-                    raise RuntimeError("affine matrix of all images should be same.")
-                if not np.allclose(header["spatial_shape"], compatible_meta["spatial_shape"]):
-                    raise RuntimeError("spatial_shape of all images should be same.")
+            _copy_compatible_dict(header, compatible_meta)
 
         img_array_ = np.stack(img_array, axis=0) if len(img_array) > 1 else img_array[0]
         return img_array_, compatible_meta
@@ -367,9 +376,9 @@ def _get_affine(self, img) -> np.ndarray:
             img: a Nibabel image object loaded from a image file.
 
         """
-        return img.affine
+        return img.affine.copy()
 
-    def _get_spatial_shape(self, img) -> Sequence:
+    def _get_spatial_shape(self, img) -> np.ndarray:
         """
         Get the spatial shape of image data, it doesn't contain the channel dim.
 
@@ -379,7 +388,7 @@ def _get_spatial_shape(self, img) -> Sequence:
         """
         ndim = img.header["dim"][0]
         spatial_rank = min(ndim, 3)
-        return list(img.header["dim"][1 : spatial_rank + 1])
+        return np.asarray(img.header["dim"][1 : spatial_rank + 1])
 
     def _get_array_data(self, img) -> np.ndarray:
         """
@@ -389,7 +398,9 @@ def _get_array_data(self, img) -> np.ndarray:
             img: a Nibabel image object loaded from a image file.
 
         """
-        return np.asarray(img.dataobj)
+        _array = np.array(img.get_fdata(dtype=self.dtype))
+        img.uncache()
+        return _array
 
 
 class NumpyReader(ImageReader):
@@ -466,7 +477,7 @@ def get_data(self, img):
 
         """
         img_array: List[np.ndarray] = list()
-        compatible_meta: Dict = None
+        compatible_meta: Dict = {}
         if isinstance(img, np.ndarray):
             img = (img,)
 
@@ -475,12 +486,7 @@ def get_data(self, img):
             if isinstance(i, np.ndarray):
                 header["spatial_shape"] = i.shape
             img_array.append(i)
-
-            if compatible_meta is None:
-                compatible_meta = header
-            else:
-                if not np.allclose(header["spatial_shape"], compatible_meta["spatial_shape"]):
-                    raise RuntimeError("spatial_shape of all images should be same.")
+            _copy_compatible_dict(header, compatible_meta)
 
         img_array_ = np.stack(img_array, axis=0) if len(img_array) > 1 else img_array[0]
         return img_array_, compatible_meta
@@ -551,18 +557,13 @@ def get_data(self, img):
 
         """
         img_array: List[np.ndarray] = list()
-        compatible_meta: Dict = None
+        compatible_meta: Dict = {}
 
         for i in ensure_tuple(img):
             header = self._get_meta_dict(i)
             header["spatial_shape"] = self._get_spatial_shape(i)
             img_array.append(np.asarray(i))
-
-            if compatible_meta is None:
-                compatible_meta = header
-            else:
-                if not np.allclose(header["spatial_shape"], compatible_meta["spatial_shape"]):
-                    raise RuntimeError("spatial_shape of all images should be same.")
+            _copy_compatible_dict(header, compatible_meta)
 
         img_array_ = np.stack(img_array, axis=0) if len(img_array) > 1 else img_array[0]
         return img_array_, compatible_meta
@@ -574,17 +575,17 @@ def _get_meta_dict(self, img) -> Dict:
             img: a PIL Image object loaded from a image file.
 
         """
-        meta = dict()
-        meta["format"] = img.format
-        meta["mode"] = img.mode
-        meta["width"] = img.width
-        meta["height"] = img.height
-        return meta
+        return {
+            "format": img.format,
+            "mode": img.mode,
+            "width": img.width,
+            "height": img.height,
+        }
 
-    def _get_spatial_shape(self, img) -> Sequence:
+    def _get_spatial_shape(self, img) -> np.ndarray:
         """
         Get the spatial shape of image data, it doesn't contain the channel dim.
         Args:
             img: a PIL Image object loaded from a image file.
         """
-        return [img.width, img.height]
+        return np.asarray((img.width, img.height))

monai/data/iterable_dataset.py

@@ -42,5 +42,4 @@ def __iter__(self):
         for data in self.source:
             if self.transform is not None:
                 data = apply_transform(self.transform, data)
-
             yield data