TransUNet — No More CNNs for Image Segmentation

TransUNet — Revolutionize Traditional Image Segmentation

Renovating U-Net by combining CNNs and Transformer to achieve SOTA results on image segmentation tasks.

Table of Contents

· 🔥 Intuition
· 🔥 TransUNet
∘ Down-Sampling (Encoding)
∘ Up-Sampling (Decoding)
· 🔥 Results
· 🔥 Implementations
· Reference

🔥 Intuition

Nowadays, U-Net has dominated image segmentation tasks, especially in the field of medical imaging. Among most of the hitherto proposed U-Nets, Convolution Neural Networks (CNN) are widely used …

artificial intelligence cnns computer vision data science deep learning image machine learning segmentation