A²Seg: Adversarial learning with channel-wise Attention for skin lesion Segmentation Dependencies python 3.11 (below should be fine) Pytorch 2.0 (below should be fine) CUDA Dataset ISBI International Skin Imaging Collaboration (ISIC) 2017 challenge, Part I Lesion Segmentation Project Structure File Description dataLoader.py (as it says) transform.py Functions for data augmentation segmentor.py Model that predict segmentation masks, based on Attention U-net critic.py Model that takes masked images to produce multi-scale feature map train.py (as it says) Model Structure Results References The code of this project is greatly inspired by SegAN (Yuan Xue, et al., 2018), sincere respects to their work and briliant ideas.
