A while back we worked with someone from UCSF and implemented a new mask indexing scheme for rectangular images. This seemed compatible with UCSF images, but they don't work for MIBI data. We need to update the coordinate multiplier
How did you implement your changes
Change coordinates = x_coords * img_data.shape[0] + y_coords to coordinates = x_coords * img_data.shape[1] + y_coords. I confirmed this on Patricia's MALDI dataset.
What is the purpose of this PR?
A while back we worked with someone from UCSF and implemented a new mask indexing scheme for rectangular images. This seemed compatible with UCSF images, but they don't work for MIBI data. We need to update the coordinate multiplier
How did you implement your changes
Change
coordinates = x_coords * img_data.shape[0] + y_coords
tocoordinates = x_coords * img_data.shape[1] + y_coords
. I confirmed this on Patricia's MALDI dataset.Remaining issues
More rectangular image testing needed.