Skip to content

Latest commit

 

History

History
33 lines (24 loc) · 4.15 KB

README.md

File metadata and controls

33 lines (24 loc) · 4.15 KB
Raw

Focal loss for dense object detection

Abstract

illustration

The highest accuracy object detectors to date are based on a two-stage approach popularized by R-CNN, where a classifier is applied to a sparse set of candidate object locations. In contrast, one-stage detectors that are applied over a regular, dense sampling of possible object locations have the potential to be faster and simpler, but have trailed the accuracy of two-stage detectors thus far. In this paper, we investigate why this is the case. We discover that the extreme foreground-background class imbalance encountered during training of dense detectors is the central cause. We propose to address this class imbalance by reshaping the standard cross entropy loss such that it down-weights the loss assigned to well-classified examples. Our novel Focal Loss focuses training on a sparse set of hard examples and prevents the vast number of easy negatives from overwhelming the detector during training. To evaluate the effectiveness of our loss, we design and train a simple dense detector we call RotatedRetinaNet. Our results show that when trained with the focal loss, RetinaNet is able to match the speed of previous one-stage detectors while surpassing the accuracy of all existing state-of-the-art two-stage detectors.

Results and Models

DOTA1.0

Backbone mAP Angle lr schd Mem (GB) Inf Time (fps) Aug Batch Size Configs Download
ResNet50 (1024,1024,200) 64.55 oc 1x 3.38 14.8 - 2 rotated_retinanet_hbb_r50_fpn_1x_dota_oc model | log
ResNet50 (1024,1024,200) 68.42 le90 1x 3.38 16.2 - 2 rotated_retinanet_obb_r50_fpn_1x_dota_le90 model | log
ResNet50 (1024,1024,200) 69.79 le135 1x 3.38 16.6 - 2 rotated_retinanet_obb_r50_fpn_1x_dota_le135 model | log
ResNet50 (1024,1024,200) 76.50 le90 1x 16.7 MS+RR 2 rotated_retinanet_obb_r50_fpn_1x_dota_ms_rr_le90 model | log
  • MS means multiple scale image split.
  • RR means random rotation.

Citation

@inproceedings{lin2017focal,
  title={Focal loss for dense object detection},
  author={Lin, Tsung-Yi and Goyal, Priya and Girshick, Ross and He, Kaiming and Doll{\'a}r, Piotr},
  booktitle={Proceedings of the IEEE international conference on computer vision},
  year={2017}
}