FreeAnchor: Learning to Match Anchors for Visual Object Detection

5 Sep 2019Xiaosong ZhangFang WanChang LiuRongrong JiQixiang Ye

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... (read more)

PDF Abstract

Evaluation results from the paper


Task Dataset Model Metric name Metric value Global rank Compare
Object Detection COCO test-dev FreeAnchor (ResNeXt-101) box AP 44.8 # 11
Object Detection COCO test-dev FreeAnchor (ResNeXt-101) AP50 64.3 # 17
Object Detection COCO test-dev FreeAnchor (ResNeXt-101) AP75 48.4 # 14
Object Detection COCO test-dev FreeAnchor (ResNeXt-101) APS 27 # 15
Object Detection COCO test-dev FreeAnchor (ResNeXt-101) APM 47.9 # 11
Object Detection COCO test-dev FreeAnchor (ResNeXt-101) APL 56 # 16