Multi-Scale Heterogeneity-Aware Hypergraph Representation for Histopathology Whole Slide Images（ICME 2024）

Pytorch implementation for the Heterogeneous Hypergraph Representation learning in the paper Multi-Scale Heterogeneity-Aware Hypergraph Representation for Histopathology Whole Slide Images.

Installation

a. Create a conda virtual environment and activate it.

conda create -n H2GT python=3.9 -y
conda activate H2GT

b. Install PyTorch and torchvision following the official instructions, e.g.,

conda install pytorch torchvision -c pytorch

c. Install other libraries.

• Install OpenSlide and openslide-python.
  Tutorial 1 and Tutorial 2 (Windows).

• Install dgl

  pip install  dgl -f https://data.dgl.ai/wheels/cu118/repo.html
  pip install  dglgo -f https://data.dgl.ai/wheels-test/repo.html

Stage 1: Data pre-processing

Please refer to CLAM for data pre-processing.

Data pre-processing: Download the raw WSI data and Prepare the patches.

Stage 2: Construct heterogeneous hypergraph

The aggregator is firstly trained with bag-level labels end to end.

python construct_hypergraph.py --config /path/to/the/config

Stage 3: Training

For different methods, we pre-set their config files in folder configs.

python main.py --config /path/to/the/config