Summary of Self-supervised Learning Featuring Small-scale Image Dataset For Treatable Retinal Diseases Classification, by Luffina C. Huang et al.
Self-Supervised Learning Featuring Small-Scale Image Dataset for Treatable Retinal Diseases Classification
by Luffina C. Huang, Darren J. Chiu, Manish Mehta
First submitted to arxiv on: 15 Apr 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 The paper proposes Self-Supervised Learning (SSL) as an alternative to Transfer Learning (TL) for medical image classification. It evaluates six models – four SSL and two TL – on a small-scale Optical Coherence Tomography (OCT) dataset, focusing on treatable retinal diseases. The results show that the proposed SSL model achieves state-of-the-art accuracy of 98.84% using only 4,000 training images. The study highlights the superiority of SSL models in both balanced and imbalanced training scenarios, with MoCo-v2 scheme performing consistently well under imbalanced conditions. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Automated medical diagnosis is getting better at recognizing images, but it’s limited by having too few images and needing a lot of human work to prepare them. Researchers found a way to train machines using just existing images without extra help from humans. They tested four special training methods – called Self-Supervised Learning (SSL) – and two more traditional methods on small pictures of eyes taken with Optical Coherence Tomography (OCT). The results show that the new method works really well, especially when there aren’t many examples to train with. |
Keywords
» Artificial intelligence » Image classification » Self supervised » Transfer learning
