This file can help with selecting images for model-building, validation and test sets. For data splitting we should likely stratify on the basis of Orientation, View and Pathology.
Pathology serves as the classification objective (labelled as NORMAL, MALIGNANT or BENIGN).
Some current issues with images in the data folder:
Image B_3159_1.LEFT_CC.LJPEG.png from DDSM was not supplied with the other 3 remaining files for this patient. Need to decide what to do with this.
There is a missing file in the DDSM data - the file should be A_1160_1.RIGHT_MLO.LJPEG.pgm which should be in subfolder 1
Number of files in the data/png/subfolders amount to 4005, while the number of files available in the data/pgm/subfolders amount to 4291 .. some files may have been missed in the conversion process.
Will update the meta_data file when the issues have been resolved.
This file can help with selecting images for model-building, validation and test sets. For data splitting we should likely stratify on the basis of Orientation, View and Pathology.
Pathology serves as the classification objective (labelled as NORMAL, MALIGNANT or BENIGN).
Some current issues with images in the data folder:
Image B_3159_1.LEFT_CC.LJPEG.png from DDSM was not supplied with the other 3 remaining files for this patient. Need to decide what to do with this.
There is a missing file in the DDSM data - the file should be A_1160_1.RIGHT_MLO.LJPEG.pgm which should be in subfolder 1
Number of files in the data/png/subfolders amount to 4005, while the number of files available in the data/pgm/subfolders amount to 4291 .. some files may have been missed in the conversion process.
Will update the meta_data file when the issues have been resolved.