TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation

TLDR

Medical image segmentation is essential for diagnosis and treatment planning, yet the U‑Net architecture, while dominant, struggles to model long‑range dependencies due to convolutional locality, whereas Transformers provide global self‑attention but lack fine‑grained localization. This work introduces TransUNet, a hybrid model that leverages Transformers as powerful encoders while integrating U‑Net to recover detailed spatial information for medical image segmentation. TransUNet encodes tokenized image patches from a CNN feature map with a Transformer to capture global context, then upsamples and fuses these features with high‑resolution CNN maps in a U‑Net decoder to achieve precise localization. The resulting model surpasses state‑of‑the‑art methods on multi‑organ and cardiac segmentation benchmarks, and its code and pretrained models are publicly available.

Abstract

Medical image segmentation is an essential prerequisite for developing healthcare systems, especially for disease diagnosis and treatment planning. On various medical image segmentation tasks, the u-shaped architecture, also known as U-Net, has become the de-facto standard and achieved tremendous success. However, due to the intrinsic locality of convolution operations, U-Net generally demonstrates limitations in explicitly modeling long-range dependency. Transformers, designed for sequence-to-sequence prediction, have emerged as alternative architectures with innate global self-attention mechanisms, but can result in limited localization abilities due to insufficient low-level details. In this paper, we propose TransUNet, which merits both Transformers and U-Net, as a strong alternative for medical image segmentation. On one hand, the Transformer encodes tokenized image patches from a convolution neural network (CNN) feature map as the input sequence for extracting global contexts. On the other hand, the decoder upsamples the encoded features which are then combined with the high-resolution CNN feature maps to enable precise localization. We argue that Transformers can serve as strong encoders for medical image segmentation tasks, with the combination of U-Net to enhance finer details by recovering localized spatial information. TransUNet achieves superior performances to various competing methods on different medical applications including multi-organ segmentation and cardiac segmentation. Code and models are available at https://github.com/Beckschen/TransUNet.