This repository has been archived. The new source code is hosted at https://gitlab.com/flywheel-io/scientific-solutions/gears/wsi-to-dicom

fwgear-wsi-to-dicom-converter

Flywheel Gear implementation of GCP's conversion tool for converting whole slide images to DICOM

This gear packages the google WSI to DICOM tool. This tool relies on OpenSlide, which supports a variety of file formats.

For more detailed information on this tool, and access to sample data, see the wsiToDcm github repository

The information you need to know to run this as a gear in flywheel is duplicated here:

Supported file formats:

Currently, this gear supports all file types readable by OpenSlide:

• Aperio (.svs, .tif)
• Hamamatsu (.vms, .vmu, .ndpi)
• Leica (.scn)
• MIRAX (.mrxs)
• Philips (.tiff)
• Sakura (.svslide)
• Trestle (.tif)
• Ventana (.bif, .tif)
• Generic tiled TIFF (.tif)

Preprocessing

No preprocessing is required for this gear

Outputs

This gear will generate a number of output dicom files dependent on the input settings. The wsiToDcm script frequently returns non-zero codes, despite running successfully with no error messages. Because of this, the gear cannot rely on the return code to indicate success or failure of the code. Therefor, the creation of any number of output files is considered a successful run. Please carefully examine the output of this gear to ensure that it did run successfully.

Complete set of parameters

Inputs

• Input_file (required): Input wsi file, supported by openslide.
• jsonFile (optional): Dicom json file with additional tags

Config

• tileHeight (required): Tile height px. (Default 500).

• tileWidth (required): Tile width px. (Default 500).

• levels (required): Number of levels to generate, levels == 0 means number of levels will be read from wsi file. (Default 0).

• downsamples (optional): Size factor for each level for each level, downsample is size factor for each level.

  eg: if base level size is 100x100 and downsamples is (1, 2, 10) then

  • level0 100x100
  • level1 50x50
  • level2 10x10

• startOn (required): Level to start generation. (Default 0).

• stopOn (required): Level to stop generation. ("Max Level" is -1, Default -1).

• sparse (required): Use TILED_SPARSE frame organization, by default it's TILED_FULL (Default False).

• compression (required): Compression, supported compressions: jpeg, jpeg2000, raw. (Default jpeg).

• seriesDescription (optional): (0008,103E) [LO] SeriesDescription Dicom tag.

• studyId (optional): (0020,000D) [UI] StudyInstanceUID Dicom tag.

• seriesId (optional): (0020,000E) [UI] SeriesInstanceUID Dicom tag.

• batch (required): Maximum frames in one file, as limit is exceeded new files is started. ("No Maximum" is -1, Default -1).

  eg: 3 files will be generated if batch is 10 and 30 frames in level

• threads (required): Threads to consume during execution ("Maximum Threads" is -1, Default -1)

• debug (required): Print debug messages: dimensions of levels, size of frames. (Default False).

• dropFirstRowAndColumn (required): Drop first row and column of the source image in order to workaround bug (Default False).