Summary of Scribformer: Transformer Makes Cnn Work Better For Scribble-based Medical Image Segmentation, by Zihan Li et al.
ScribFormer: Transformer Makes CNN Work Better for Scribble-based Medical Image Segmentation
by Zihan Li, Yuan Zheng, Dandan Shan, Shuzhou Yang, Qingde Li, Beizhan Wang, Yuanting Zhang, Qingqi Hong, Dinggang Shen
First submitted to arxiv on: 3 Feb 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This research proposes a novel approach to scribble-supervised medical image segmentation, addressing limitations of existing methods that rely on convolutional neural networks (CNNs) with encoder-decoder architectures. The proposed ScribFormer model combines the strengths of CNNs and transformers to capture both local and global features from scribble annotations. The triple-branch structure includes a CNN branch, a Transformer branch, and an attention-guided class activation map (ACAM) branch. Experimental results on two public datasets and one private dataset demonstrate superior performance compared to state-of-the-art scribble-supervised segmentation methods, even outperforming fully-supervised methods in some cases. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Scribble-supervised medical image segmentation is a challenging task that requires learning global shape information from limited annotations. A new CNN-Transformer hybrid solution called ScribFormer addresses this issue by combining the strengths of CNNs and transformers. The model has a triple-branch structure, including a CNN branch, a Transformer branch, and an ACAM branch. This approach allows for capturing both local and global features from scribble annotations. Results show that ScribFormer outperforms state-of-the-art methods on public and private datasets. |
Keywords
* Artificial intelligence * Attention * Cnn * Encoder decoder * Image segmentation * Supervised * Transformer
