Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox


Equalization Loss v2: A New Gradient Balance Approach for Long-tailed Object Detection

Dec 15, 2020
Jingru Tan, Xin Lu, Gang Zhang, Changqing Yin, Quanquan Li



Recently proposed decoupled training methods emerge as a dominant paradigm for long-tailed object detection. But they require an extra fine-tuning stage, and the disjointed optimization of representation and classifier might lead to suboptimal results. However, end-to-end training methods, like equalization loss (EQL), still perform worse than decoupled training methods. In this paper, we reveal the main issue in long-tailed object detection is the imbalanced gradients between positives and negatives, and find that EQL does not solve it well. To address the problem of imbalanced gradients, we introduce a new version of equalization loss, called equalization loss v2 (EQL v2), a novel gradient guided reweighing mechanism that re-balances the training process for each category independently and equally. Extensive experiments are performed on the challenging LVIS benchmark. EQL v2 outperforms origin EQL by about 4 points overall AP with 14-18 points improvements on the rare categories. More importantly, It also surpasses decoupled training methods. Without further tuning for the Open Images dataset, EQL v2 improves EQL by 6.3 points AP, showing strong generalization ability. Codes will be released at https://github.com/tztztztztz/eqlv2



Share this with someone who'll enjoy it:

   Access Paper Source



Share this with someone who'll enjoy it: