Add nvimgcodec GPU accelerated decoder to Pydicom for decoding compressed pixel data (#570)

* Add custom pydicom decoder using nvimgcodec

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Module name change

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* fix complaints

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Correct module name

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Spacing change

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* comments change

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Fix linting complaints

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Fix Copilot nits

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* More improvements

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* version parser

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Reorg the code

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Add more logging

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Refine register and unregister functions

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* upate decoder

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Add decoder instance caching as it gets used for every frame

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Added unit tests and updated dependencies for compressed DICOM support

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Fix complaints on unit test

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Check for GDCM but it should have been taken care of by the reqs

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Catch exception on checking cuda device on Github build machine

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Varaible name changes

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Fix nvimgcodec decoder (#571)

* Fix handling of JPEG color space
* Added JPEG2000 precision support
* Updates PhotometricInterpretation to RGB when color space conversion is applied

Passes all JPEG, JPEG2000 and HTJ2K tests

Signed-off-by: Joaquin Anton Guirao <janton@nvidia.com>
Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Updated the unit test code and the decoder itself

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Added note on using nvimgcodec and enhancing perf test to support user supplied dicom files

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Update due to depcreation of dependency

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Fix complaints

Signed-off-by: M Q <mingmelvinq@nvidia.com>

* Proactively removed the use deprecated pydicom class properties

Signed-off-by: M Q <mingmelvinq@nvidia.com>

---------

Signed-off-by: M Q <mingmelvinq@nvidia.com>
Signed-off-by: Joaquin Anton Guirao <janton@nvidia.com>
Co-authored-by: Joaquin Anton <j.anton.guirao@gmail.com>

Author: MMelQin

monai/deploy/operators/__init__.py

@@ -37,6 +37,7 @@
 # from holoscan.operators import *
 from holoscan.operators import PingRxOp, PingTxOp, VideoStreamRecorderOp, VideoStreamReplayerOp
 
+from . import decoder_nvimgcodec
 from .clara_viz_operator import ClaraVizOperator
 from .dicom_data_loader_operator import DICOMDataLoaderOperator
 from .dicom_encapsulated_pdf_writer_operator import DICOMEncapsulatedPDFWriterOperator
@@ -52,3 +53,29 @@
 from .png_converter_operator import PNGConverterOperator
 from .publisher_operator import PublisherOperator
 from .stl_conversion_operator import STLConversionOperator, STLConverter
+
+__all__ = [
+    "decoder_nvimgcodec",
+    "BundleConfigNames",
+    "ClaraVizOperator",
+    "DICOMDataLoaderOperator",
+    "DICOMEncapsulatedPDFWriterOperator",
+    "DICOMSegmentationWriterOperator",
+    "DICOMSeriesSelectorOperator",
+    "DICOMSeriesToVolumeOperator",
+    "DICOMTextSRWriterOperator",
+    "EquipmentInfo",
+    "InferenceOperator",
+    "IOMapping",
+    "ModelInfo",
+    "MonaiBundleInferenceOperator",
+    "MonaiSegInferenceOperator",
+    "NiftiDataLoader",
+    "PNGConverterOperator",
+    "PublisherOperator",
+    "random_with_n_digits",
+    "save_dcm_file",
+    "write_common_modules",
+    "STLConversionOperator",
+    "STLConverter",
+]