Skip to content

Latest commit

 

History

History
10 lines (7 loc) · 407 Bytes

README.md

File metadata and controls

10 lines (7 loc) · 407 Bytes

Attention2majority

This is the official implementation of the Attention2majority code on Camelyon16 dataset.

Each WSI need to be patchified and saved in .h5 files. Each .h5 file has two datasets: "bag" and "label", where "bag" contains a matrix in Nx224x224x3, and "label" contains the label value of the WSI.

intelligent sampling

$ python discriminator.py

slide train

$ python MIL.py