Changing the hardware setup (the image loader) for a more fine graine…

…d image loader selection

.gitignore

@@ -29,7 +29,7 @@ test_database/3DHistech
 test_database/XiaLab
 test_database/dicom_files
 test_database/dicom_leica
-dpyproject.toml
+test_database/dicom_leica_anonymized
+test_database/dicom_wsidicomizer
 mask_ihc.ipynb
-pyproject.toml
 test_output

README.md

@@ -118,7 +118,7 @@ In our Pre-Processing pipeline, we are able to extract quadratic patches from de
       [--apply_prefilter APPLY_PREFILTER]
       [--log_path LOG_PATH]
       [--log_level {critical,error,warning,info,debug}]
-      [--hardware_selection {cucim,openslide}]
+      [--hardware_selection {cucim,openslide,wsidicom}]
       [--wsi_magnification WSI_MAGNIFICATION]
       [--wsi_mpp WSI_MPP]
 
@@ -256,9 +256,9 @@ In our Pre-Processing pipeline, we are able to extract quadratic patches from de
     --log_level {critical,error,warning,info,debug}
                           Set the logging level. Options are ['critical', 'error',
                           'warning', 'info', 'debug'] (default: None)
-    --hardware_selection {cucim,openslide}
+    --hardware_selection {cucim,openslide,wsidicom}
                           Select hardware device (just if available, otherwise always
-                          cucim). Defaults to cucim. (default: None)
+                          cucim). Defaults to None. (default: None)
     --wsi_magnification WSI_MAGNIFICATION
                           Manual WSI magnification, but just applies if metadata
                           cannot be derived from OpenSlide (e.g., for .tiff files).
@@ -370,6 +370,11 @@ In our Pre-Processing pipeline, we are able to extract quadratic patches from de
   ```
 </details>
 
+## A Note on DICOM
+If you use DICOM files directly converted by vendors, first try to use OpenSlide which is the default setting. If this is slow, you could try to enforce the `wsidicom` image loader by setting the hardware to "wsidicom" (`hardware_selection: wsi_dicom`or `--hardware_selection wsidicom`) and check if this works.
+
+If you converted files to DICOM with the `wsidicomizer` tool as explained below, please always set the image loader to "wsidicom" (`hardware_selection: wsi_dicom`or `--hardware_selection wsidicom`).
+
 ## Examples
 An example notebook is given [here](PathoPatch.ipynb):
 <a href="https://colab.research.google.com/github/TIO-IKIM/PathoPatcher/blob/main/PathoPatch.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>

docs/README_pypi.md

@@ -92,7 +92,7 @@ wsi_extraction [-h]
     [--apply_prefilter APPLY_PREFILTER]
     [--log_path LOG_PATH]
     [--log_level {critical,error,warning,info,debug}]
-    [--hardware_selection {cucim,openslide}]
+    [--hardware_selection {cucim,openslide,wsidicom}]
     [--wsi_magnification WSI_MAGNIFICATION]
     [--wsi_mpp WSI_MPP]
 
@@ -230,9 +230,9 @@ options:
   --log_level {critical,error,warning,info,debug}
                         Set the logging level. Options are ['critical', 'error',
                         'warning', 'info', 'debug'] (default: None)
-  --hardware_selection {cucim,openslide}
-                        Select hardware device (just if available, otherwise always
-                        cucim). Defaults to cucim. (default: None)
+    --hardware_selection {cucim,openslide,wsidicom}
+                          Select hardware device (just if available, otherwise always
+                          cucim). Defaults to None. (default: None)
   --wsi_magnification WSI_MAGNIFICATION
                         Manual WSI magnification, but just applies if metadata
                         cannot be derived from OpenSlide (e.g., for .tiff files).
@@ -306,6 +306,8 @@ WSI_Name
 ├── patch_metadata.json   # Patch metadata of WSI merged in one file
 └── thumbnail.png         # WSI thumbnail
 ```
+
+## Further information
 For more information, check out the git.
 
 ## License

examples/patch_extraction.yaml

@@ -80,4 +80,4 @@ filter_patches:               # Post-extraction patch filtering to sort out arte
 # logging
 log_path:                     # Path where log files should be stored. Otherwise, log files are stored in the output folder. [str][Optional, defaults to None]
 log_level:                    # Set the logging level. [str][Optional, defaults to info]
-hardware_selection:           # Select hardware device (just if available, otherwise always cucim). [str] [Optional, defaults to cucim]
+hardware_selection:           # Select hardware device (just if available, otherwise always cucim). Options are openslide,cucim,wsidicom [str] [Optional, defaults to cucim]

pathopatch/cli.py

@@ -148,7 +148,7 @@ class PreProcessingConfig(BaseModel):
         apply_prefilter (bool, optional): Pre-extraction mask filtering to remove marker from mask before applying otsu. Defaults to False.
         log_path (str, optional): Path where log files should be stored. Otherwise, log files are stored in the output folder. Defaults to None.
         log_level (str, optional): Set the logging level. Defaults to "info".
-        hardware_selection (str, optional): Select hardware device (just if available, otherwise always cucim). Defaults to "cucim".
+        hardware_selection (str, optional): Select hardware device (just if available, otherwise always cucim). Defaults to None.
         wsi_properties (dict, optional): Dictionary with manual WSI metadata, but just applies if metadata cannot be derived from OpenSlide (e.g., for .tiff files). Supported keys are slide_mpp and magnification
 
     Raises:
@@ -209,7 +209,7 @@ class PreProcessingConfig(BaseModel):
     # other
     log_path: Optional[str]
     log_level: Optional[str] = "info"
-    hardware_selection: Optional[str] = "cucim"
+    hardware_selection: Optional[str] = None
     wsi_properties: Optional[dict]
 
     def __init__(__pydantic_self__, **data: Any) -> None:
@@ -577,8 +577,8 @@ def __init__(self) -> None:
         parser.add_argument(
             "--hardware_selection",
             type=str,
-            choices=["cucim", "openslide"],
-            help="Select hardware device (just if available, otherwise always cucim). Defaults to cucim.",
+            choices=["cucim", "openslide", "wsidicom"],
+            help="Select hardware device (just if available, otherwise always cucim). Defaults to None.",
         )
         parser.add_argument(
             "--wsi_magnification",

pathopatch/patch_extraction/dataset.py

@@ -28,6 +28,10 @@
 from PIL import Image
 from pathopatch.utils.exceptions import WrongParameterException
 from pathopatch.wsi_interfaces.openslide_deepzoom import DeepZoomGeneratorOS
+from pathopatch.wsi_interfaces.wsidicomizer_openslide import (
+    DicomSlide,
+    DeepZoomGeneratorDicom,
+)
 from pathopatch.utils.patch_util import (
     calculate_background_ratio,
     compute_interesting_patches,
@@ -280,20 +284,44 @@ def __init__(
 
     def _set_hardware(self) -> None:
         """Either load CuCIM (GPU-accelerated) or OpenSlide"""
-        if module_exists("cucim", error="ignore"):
-            self.logger.debug("Using CuCIM")
-            from cucim import CuImage
+        wsi_file = Path(self.config.wsi_path)
+        if wsi_file.is_dir():
+            if len(list(wsi_file.glob("*.dcm"))) != 0:
+                self.logger.debug("Detected dicom files")
+                try:
+                    dcm_files = list(wsi_file.glob("*.dcm"))
+                    dcm_files = [(f, os.path.getsize(f.resolve())) for f in dcm_files]
+                    dcm_files = sorted(dcm_files, key=lambda x: x[1], reverse=True)
+                    OpenSlide(str(dcm_files[0][0]))
+                    wsi_file = dcm_files[0][0]
+                    self.image_loader = OpenSlide
+                    self.deepzoomgenerator = DeepZoomGeneratorOS
+                    self.slide_metadata_loader = OpenSlide
+                except:
+                    try:
+                        DicomSlide(wsi_file)
+                    except Exception as e:
+                        raise e
+                    self.deepzoomgenerator = DeepZoomGeneratorDicom
+                    self.image_loader = DicomSlide
+                    self.slide_metadata_loader = DicomSlide
+        else:
+            if module_exists("cucim", error="ignore"):
+                self.logger.debug("Using CuCIM")
+                from cucim import CuImage
 
-            from pathopatch.wsi_interfaces.cucim_deepzoom import (
-                DeepZoomGeneratorCucim,
-            )
+                from pathopatch.wsi_interfaces.cucim_deepzoom import (
+                    DeepZoomGeneratorCucim,
+                )
 
-            self.deepzoomgenerator = DeepZoomGeneratorCucim
-            self.image_loader = CuImage
-        else:
-            self.logger.debug("Using OpenSlide")
-            self.deepzoomgenerator = DeepZoomGeneratorOS
-            self.image_loader = OpenSlide
+                self.deepzoomgenerator = DeepZoomGeneratorCucim
+                self.image_loader = CuImage
+                self.slide_metadata_loader = OpenSlide
+            else:
+                self.logger.debug("Using OpenSlide")
+                self.deepzoomgenerator = DeepZoomGeneratorOS
+                self.image_loader = OpenSlide
+                self.slide_metadata_loader = OpenSlide
 
     def _set_tissue_detector(self) -> None:
         """Set up the tissue detection model and transformations.
@@ -364,7 +392,7 @@ def _prepare_slide(
                 * List[Polygon]: List of polygons, downsampled to the target level
                 * List[str]: List of region labels
         """
-        self.slide_openslide = OpenSlide(str(self.config.wsi_path))
+        self.slide_openslide = self.slide_metadata_loader(str(self.config.wsi_path))
         self.slide = self.image_loader(str(self.config.wsi_path))
 
         if (