Summary of Procns: Progressive Prototype Calibration and Noise Suppression For Weakly-supervised Medical Image Segmentation, by Y. Liu et al.
ProCNS: Progressive Prototype Calibration and Noise Suppression for Weakly-Supervised Medical Image Segmentation
by Y. Liu, L. Lin, K. K. Y. Wong, X. Tang
First submitted to arxiv on: 25 Jan 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: 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 novel approach to weakly-supervised segmentation (WSS) aims to bridge the gap between annotation cost and model performance by leveraging sparse annotations. The proposed ProCNS method combines two synergistic modules: progressive prototype calibration and noise suppression. Specifically, the Prototype-based Regional Spatial Affinity (PRSA) loss maximizes pair-wise affinities between spatial and semantic elements, providing reliable guidance for the model. Meanwhile, the Adaptive Noise Perception and Masking (ANPM) module identifies and masks noisy regions, reducing potential errors. The framework generates specialized soft pseudo-labels for noisy regions, providing supplementary supervision. Extensive experiments on six medical image segmentation tasks demonstrate significant performance gains compared to state-of-the-art methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper solves a big problem in computer vision by making it easier to label images without needing to draw perfect borders. They develop a new way to use small labels to train models, which is more accurate and efficient than previous methods. The approach combines two clever ideas: using the image itself to understand what’s going on, and identifying noisy areas that can mess up the results. By doing this, they’re able to make better predictions and outperform other state-of-the-art techniques. |
Keywords
* Artificial intelligence * Image segmentation * Supervised
