AP-DPM: A Dual-Path Merging Network via Adversarial Anatomical Prior Guidance for Wrist Bone Segmentation

This is the official implementation of AP-DPM: A Dual-Path Merging Network via Adversarial Anatomical Prior Guidance for Wrist Bone Segmentation.

Setup

• Install the conda environment

conda create -n ramw600 python=3.10
conda activate ramw600

• Install Pytroch

# CUDA 12.6
pip3 install torch torchvision torchaudio

• Install other requirements, such as albumentations.

pip install -r requirements.txt

Dataset

Please download the dataset from https://huggingface.co/datasets/TokyoTechMagicYang/RAM-W600.

Run

• Training. The training configurations are in ./main_seg_dual_path.py and ./main_seg_dual_path_gan.py. We also provide scripts in ./train.sh. Before running, you should refer to ./main_seg_dual_path.py and ./main_seg_dual_path_gan.py and add your paths to the bash files. After running, the checkpoints will be saved in ./ckpts/.
• The training of AP-DPM contains 2 stages. You need to use ./main_seg_dual_path.py for the 1st stage and ./main_seg_dual_path_gan.py for the 2nd stage.

bash train.sh

• Testing. The testing configurations are in ./main_seg_dual_path.py and ./main_seg_dual_path_gan.py. We also provide scripts in ./test.sh. After running, the results of the visualization will be saved in the folder you chose for testing.

bash test.sh