Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Rethinking Rotated Object Detection with Gaussian Wasserstein Distance Loss

About

Boundary discontinuity and its inconsistency to the final detection metric have been the bottleneck for rotating detection regression loss design. In this paper, we propose a novel regression loss based on Gaussian Wasserstein distance as a fundamental approach to solve the problem. Specifically, the rotated bounding box is converted to a 2-D Gaussian distribution, which enables to approximate the indifferentiable rotational IoU induced loss by the Gaussian Wasserstein distance (GWD) which can be learned efficiently by gradient back-propagation. GWD can still be informative for learning even there is no overlapping between two rotating bounding boxes which is often the case for small object detection. Thanks to its three unique properties, GWD can also elegantly solve the boundary discontinuity and square-like problem regardless how the bounding box is defined. Experiments on five datasets using different detectors show the effectiveness of our approach. Codes are available at https://github.com/yangxue0827/RotationDetection and https://github.com/open-mmlab/mmrotate.

Xue Yang, Junchi Yan, Qi Ming, Wentao Wang, Xiaopeng Zhang, Qi Tian• 2021

Related benchmarks

TaskDatasetResultRank
Oriented Object DetectionDOTA v1.0 (test)
SV79.76
378
Object DetectionDOTA 1.0 (test)
Plane AP89.66
256
Object DetectionHRSC 2016 (test)
mAP@0.0789.85
72
Oriented Object DetectionDOTA v1.5 (test)--
58
Oriented Object DetectionHRSC 2016 (test)--
55
Rotated Object DetectionDOTA 1.0 (test)
mAP80.23
46
Oriented Object DetectionHRSC2016--
35
Oriented Object DetectionDIOR-R (test)
APL69.68
28
Object DetectionDOTA v1.0
Overall mAP76.34
24
Oriented Object DetectionDOTA v1.0
AP5071.66
16
Showing 10 of 25 rows

Other info

Code

Follow for update