higt's Introduction

HIGT: Hierarchical Interaction Graph-Transformer for Whole Slide Image Analysis

Tutorial | Paper

Authors

Ziyu Guo, Weiqin Zhao, Shujun Wang, Lequan Yu

Overview:

Update:

We apologize for the delay in updates. Below is a tutorial that encompasses the usage of HIGT and MulGT , covering aspects from data preprocessing to the training and testing phases. This tutorial is designed with the goal of standardizing the entire process and is implemented on the foundation of CLAM for ease of use and deployment.

To ensure a seamless setup, please follow the instructions below for Ubuntu systems. Note that core_utils_re.py, create_splits_seq_re.py, and extract_features_fp_re.py are revised versions of their respective files in CLAM, changed a little bit for more convenience with HIGT.

1.

!git clone https://github.com/mahmoodlab/CLAM.git .

2.

!mv /CLAM/* /

We hope this tutorial aids you in your projects. Should you encounter any issues, please don't hesitate to raise an issue on our GitHub page. Our team is committed to providing prompt and helpful feedback.

higt's People

Contributors

Stargazers

Watchers

higt's Issues

models.resnet_baseline_re not found

Hello,

I guess the resnet_baseline from CLAM was re-written but unfortunately this function is nowhere in the github repo. Could you provide it please?

Best!

Data organization

Hello, I wonder if you can upload the data processing code or how you organize your data, thx.

Input Data Preparation

Hi, I am trying to implement your method with my method.
May I know what exactly is in the 'all_data_path' and 'patient_and_label_path'? Is it that inside the 'all_data_path', I should provide the directories of the extracted features for each WSI? or can u provide me example?

Thanks in advance!!

hku-medai / higt Goto Github PK