English

Boosting Long-tailed Object Detection via Step-wise Learning on Smooth-tail Data

Computer Vision and Pattern Recognition 2023-05-23 v1

Abstract

Real-world data tends to follow a long-tailed distribution, where the class imbalance results in dominance of the head classes during training. In this paper, we propose a frustratingly simple but effective step-wise learning framework to gradually enhance the capability of the model in detecting all categories of long-tailed datasets. Specifically, we build smooth-tail data where the long-tailed distribution of categories decays smoothly to correct the bias towards head classes. We pre-train a model on the whole long-tailed data to preserve discriminability between all categories. We then fine-tune the class-agnostic modules of the pre-trained model on the head class dominant replay data to get a head class expert model with improved decision boundaries from all categories. Finally, we train a unified model on the tail class dominant replay data while transferring knowledge from the head class expert model to ensure accurate detection of all categories. Extensive experiments on long-tailed datasets LVIS v0.5 and LVIS v1.0 demonstrate the superior performance of our method, where we can improve the AP with ResNet-50 backbone from 27.0% to 30.3% AP, and especially for the rare categories from 15.5% to 24.9% AP. Our best model using ResNet-101 backbone can achieve 30.7% AP, which suppresses all existing detectors using the same backbone.

Keywords

Cite

@article{arxiv.2305.12833,
  title  = {Boosting Long-tailed Object Detection via Step-wise Learning on Smooth-tail Data},
  author = {Na Dong and Yongqiang Zhang and Mingli Ding and Gim Hee Lee},
  journal= {arXiv preprint arXiv:2305.12833},
  year   = {2023}
}

Comments

10 pages, 5 figures

R2 v1 2026-06-28T10:41:06.605Z