A Study on Transformers and CNNs in Medical Image Segmentation

Research on artificial intelligence (AI) has developed considerably in recent years. In particular, the emergence of ChatGPT in the field of natural language processing has had a greater influence on human life than AlphaGo several years ago. Attention-based Transformers, which play a significant role in the field of deep learning, are gradually spreading not only to natural images but also to medical image analysis. In the new field of medical image segmentation, models that implement a similar appearance to U-Net through the merging of Transformers and convolutional neural networks (CNN) have emerged; these models show significantly better results than the existing U-Net. In this study, we discuss the applications of Transformers and CNN-based networks in the field of medical image segmentation