Document the texture feature evaluation and the dataset that can be used for further case-control validation.

/cbica/comp_space/hsiehm/for_sarthak/compare_images.py which measure the following:

perc_nan: percentage of NaN in the masked area.
MeanSquaredError: between the image of question and the reference image
StructuralSimilarityIndex: See definition in http://scikit-image.org/docs/dev/api/skimage.measure.html#skimage.measure.compare_ssim
ExactMatch: Whether the two masked images exactly matched down to 2 decimal points.
Correlation: as it sounds.
perc_match: % of matching down to 2 decimal points.

An example of running the comparison for two list of images in python is in: /cbica/comp_space/hsiehm/for_sarthak/compare_images_example.py

This code would create a csv like /cbica/comp_space/hsiehm/for_sarthak/7D7F656B7F7A6565787F_LCC_20090122/feature/texture_feature_comparison_result_20180905.csv

PPG: /cbica/home/hsiehm/Projects/Breasts/Data/PPG_Case_Control_Final BreastAnatomy: /cbica/projects/Prospr/Data/BreastAnatomy/dicoms/all_symlinks

hsieh42 / exit_process

Document the texture feature evaluation and the dataset that can be used for further case-control validation. #28