We used KIMIA Path960, a publicly available dataset of 960 histopathology images extracted from 20 different tissue scans
Transfer learning using GoogleNet model for classification of histopathological images
test accuracy = 93.93%