Summary of Cross-patient Pseudo Bags Generation and Curriculum Contrastive Learning For Imbalanced Multiclassification Of Whole Slide Image, by Yonghuang Wu et al.
Cross-Patient Pseudo Bags Generation and Curriculum Contrastive Learning for Imbalanced Multiclassification of Whole Slide Image
by Yonghuang Wu, Xuan Xie, Xinyuan Niu, Chengqian Zhao, Jinhua Yu
First submitted to arxiv on: 18 Nov 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
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 The proposed learning framework addresses the challenge of multi-classification under sample imbalance in whole slide image (WSI) analysis for pathology computing. The authors generate sub-bags with feature distributions similar to original WSIs and utilize a pseudo-bag generation algorithm to leverage abundant information, enabling efficient training in unbalanced-sample tasks. They also introduce an affinity-based sample selection and curriculum contrastive learning strategy to enhance representation learning stability. Experimental results on three datasets demonstrate significant performance improvements in tumor classification and lymph node metastasis, with an average 4.39-point F1 score improvement compared to the second-best method. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Pathologists can use computers to analyze images of slides under a microscope, which helps them make better decisions about what they see. However, there is still a big problem: when we try to teach the computer to recognize many different types of things in these slides, it gets confused because some types are much more common than others. To fix this, scientists came up with a new way to train computers using tiny groups of slide images that have similar features. This helps the computer learn how to recognize patterns even when there aren’t enough examples of rare things. The new approach also helps the computer learn more efficiently and accurately identify what’s in these slides. |
Keywords
» Artificial intelligence » Classification » F1 score » Representation learning
