WeaklySegNet

📋 Requirements

Create the conda environment with all dependencies:

conda env create --name weaklySeg --environment.yml

Activate the environment:

conda activate weaklySeg

📊 Datasets

The following datasets are supported:

• MoNuSeg: Download from Grand Challenge
• CPM17: Download from TCGA

🛠️ Data Preparation

Before training, prepare the data with the following steps:

1. Generate Voronoi and K-means labels

python prepare_data.py

This script generates Voronoi label and k-means clustering label for the dataset.

2. Generate color-aware pseudo labels

python labelbycolor.py
python post_processing.py

These scripts generate and refine color-aware pseudo labels.

🚀 WeaklySegNet Pipeline

Run the following scripts in order:

Step 1: Coarse training

python train_coarse.py

Step 2: Fine training

python train_fine.py

Step 3: Testing

python test_fine.py

📧 Contact

If you have any questions, please feel free to contact:

Email: Jiaooo111@126.com