Two machine-learning models for the detection of breast cancers from medical imaging of breast cancer biopsy tissue. Big focus on making models accessible to cl

Two machine-learning models for the detection of breast cancers from medical imaging of breast cancer biopsy tissue. Big focus on making models accessible to clinicians. The models are: (1) PPReCOGG, one of the models based on Gabor filters and k-NN (https://github.com/jszym/pprecogg https://github.com/jszym/pprecogg) (2) DeepDuct, the second model, based on a pre-trained VGG16 network and the Grad-CAM algorithm, localises lesions _and_ informs clinicians about why the model has chosen the lesion type it did. You can find more details in my master's thesis, for which the models were written: http://cs.mcgill.ca/~jszymb/thesis/260528685_Szymborski_Joseph_Experimental_Medicine_thesis.pdf http://cs.mcgill.ca/~jszymb/thesis/260528685_Szymborski_Jose... (Edit: Also, if you're hiring machine learning people, medical or otherwise, please get in touch at hn at jszym point com)