Dense Label Encoding for Boundary Discontinuity Free Rotation Detection
About
Rotation detection serves as a fundamental building block in many visual applications involving aerial image, scene text, and face etc. Differing from the dominant regression-based approaches for orientation estimation, this paper explores a relatively less-studied methodology based on classification. The hope is to inherently dismiss the boundary discontinuity issue as encountered by the regression-based detectors. We propose new techniques to push its frontier in two aspects: i) new encoding mechanism: the design of two Densely Coded Labels (DCL) for angle classification, to replace the Sparsely Coded Label (SCL) in existing classification-based detectors, leading to three times training speed increase as empirically observed across benchmarks, further with notable improvement in detection accuracy; ii) loss re-weighting: we propose Angle Distance and Aspect Ratio Sensitive Weighting (ADARSW), which improves the detection accuracy especially for square-like objects, by making DCL-based detectors sensitive to angular distance and object's aspect ratio. Extensive experiments and visual analysis on large-scale public datasets for aerial images i.e. DOTA, UCAS-AOD, HRSC2016, as well as scene text dataset ICDAR2015 and MLT, show the effectiveness of our approach. The source code is available at https://github.com/Thinklab-SJTU/DCL_RetinaNet_Tensorflow and is also integrated in our open source rotation detection benchmark: https://github.com/yangxue0827/RotationDetection.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Oriented Object Detection | DOTA v1.0 (test) | SV79.04 | 378 | |
| Object Detection | DOTA 1.0 (test) | Plane AP89.26 | 256 | |
| Oriented Object Detection | DOTA (test) | AP (Plane)89.26 | 92 | |
| Object Detection | HRSC 2016 (test) | mAP@0.0789.46 | 72 | |
| Oriented Object Detection | DOTA v1.5 (test) | -- | 58 | |
| Oriented Object Detection | HRSC 2016 (test) | mAP89.46 | 55 | |
| Object Detection | UCAS-AOD (test) | -- | 27 | |
| Oriented Object Detection | DOTA single-scale 1.0 (test) | AP (PL)89.78 | 24 | |
| Rotation Object Detection | DOTA v2.0 (test) | AP5048.43 | 14 | |
| Rotation Object Detection | DOTA v1.0 (val) | AP5070.66 | 14 |