The Equalization Losses: Gradient-Driven Training for Long-tailed Object Recognition

October 11, 2022 ยท Entered Twilight ยท ๐Ÿ› IEEE Transactions on Pattern Analysis and Machine Intelligence

๐Ÿ’ค TWILIGHT: Eternal Rest
Repo abandoned since publication

Repo contents: .flake8, .gitignore, LICENSE, README.md, benchmark, configs, docs, easy_setup.sh, requirements.txt, scripts, setup.py, up-logo.png, up

Authors Jingru Tan, Bo Li, Xin Lu, Yongqiang Yao, Fengwei Yu, Tong He, Wanli Ouyang arXiv ID 2210.05566 Category cs.CV: Computer Vision Citations 24 Venue IEEE Transactions on Pattern Analysis and Machine Intelligence Repository https://github.com/ModelTC/United-Perception โญ 436 Last Checked 1 month ago
Abstract
Long-tail distribution is widely spread in real-world applications. Due to the extremely small ratio of instances, tail categories often show inferior accuracy. In this paper, we find such performance bottleneck is mainly caused by the imbalanced gradients, which can be categorized into two parts: (1) positive part, deriving from the samples of the same category, and (2) negative part, contributed by other categories. Based on comprehensive experiments, it is also observed that the gradient ratio of accumulated positives to negatives is a good indicator to measure how balanced a category is trained. Inspired by this, we come up with a gradient-driven training mechanism to tackle the long-tail problem: re-balancing the positive/negative gradients dynamically according to current accumulative gradients, with a unified goal of achieving balance gradient ratios. Taking advantage of the simple and flexible gradient mechanism, we introduce a new family of gradient-driven loss functions, namely equalization losses. We conduct extensive experiments on a wide spectrum of visual tasks, including two-stage/single-stage long-tailed object detection (LVIS), long-tailed image classification (ImageNet-LT, Places-LT, iNaturalist), and long-tailed semantic segmentation (ADE20K). Our method consistently outperforms the baseline models, demonstrating the effectiveness and generalization ability of the proposed equalization losses. Codes will be released at https://github.com/ModelTC/United-Perception.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

๐Ÿ“œ Similar Papers

In the same crypt โ€” Computer Vision