Our solutions of Konyang Health Datathon 2019 (Breast Cancer Classification challenge)

• 13st place (as team Paten, F1 Score 0.473)
• No pretrain models allowed(Only scratch based training allowed), No contour annotations are provided.

We prepare large codes for competition, but can't test them due to time limit(24hr). Alternatively we train our data with XCeptionNet and baseline CNN(typical CNN). (see dummyTestCodes/team059_KHD2019_MAMMO_274/main_pytorch.py)

We will reaserch more with national breast cancer dataset later. More infos for code will be coming soon. (we need to take a break..)