Skip to content

Latest commit

 

History

History
37 lines (26 loc) · 4.02 KB

README.md

File metadata and controls

37 lines (26 loc) · 4.02 KB

FreeAnchor

FreeAnchor: Learning to Match Anchors for Visual Object Detection

Abstract

Modern CNN-based object detectors assign anchors for ground-truth objects under the restriction of object-anchor Intersection-over-Unit (IoU). In this study, we propose a learning-to-match approach to break IoU restriction, allowing objects to match anchors in a flexible manner. Our approach, referred to as FreeAnchor, updates hand-crafted anchor assignment to "free" anchor matching by formulating detector training as a maximum likelihood estimation (MLE) procedure. FreeAnchor targets at learning features which best explain a class of objects in terms of both classification and localization. FreeAnchor is implemented by optimizing detection customized likelihood and can be fused with CNN-based detectors in a plug-and-play manner. Experiments on COCO demonstrate that FreeAnchor consistently outperforms their counterparts with significant margins.

Results and Models

Backbone Style Lr schd Mem (GB) Inf time (fps) box AP Config Download
R-50 pytorch 1x 4.9 18.4 38.7 config model | log
R-101 pytorch 1x 6.8 14.9 40.3 config model | log
X-101-32x4d pytorch 1x 8.1 11.1 41.9 config model | log

Notes:

  • We use 8 GPUs with 2 images/GPU.
  • For more settings and models, please refer to the official repo.

Citation

@inproceedings{zhang2019freeanchor,
  title   =  {{FreeAnchor}: Learning to Match Anchors for Visual Object Detection},
  author  =  {Zhang, Xiaosong and Wan, Fang and Liu, Chang and Ji, Rongrong and Ye, Qixiang},
  booktitle =  {Neural Information Processing Systems},
  year    =  {2019}
}