Dense Label Encoding for Boundary Discontinuity Free Rotation Detection

Abstract

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.

arXiv comment12 pages, 6 figures, 9 tables, accepted by CVPR21

References in corpus (15)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• R-FCN: Object Detection via Region-based Fully Convolutional Networks
• Gliding vertex on the horizontal bounding box for multi-oriented object detection
• TextBoxes++: A Single-Shot Oriented Scene Text Detector
• FCOS: Fully Convolutional One-Stage Object Detection
• Automatic Ship Detection of Remote Sensing Images from Google Earth in Complex Scenes Based on Multi-Scale Rotation Dense Feature Pyramid Networks
• R2CNN: Rotational Region CNN for Orientation Robust Scene Text Detection
• Position Detection and Direction Prediction for Arbitrary-Oriented Ships via Multitask Rotation Region Convolutional Neural Network
• R3Det: Refined Single-Stage Detector with Feature Refinement for Rotating Object
• SCRDet: Towards More Robust Detection for Small, Cluttered and Rotated Objects
• PolarDet: A Fast, More Precise Detector for Rotated Target in Aerial Images
• IENet: Interacting Embranchment One Stage Anchor Free Detector for Orientation Aerial Object Detection
• Learning Modulated Loss for Rotated Object Detection
• Objects detection for remote sensing images based on polar coordinates
• Dual Path Networks

Cited by in corpus (4)

• Rethinking Rotated Object Detection with Gaussian Wasserstein Distance Loss
• Arbitrary-Oriented Ship Detection through Center-Head Point Extraction
• Sampling Equivariant Self-attention Networks for Object Detection in Aerial Images
• Learning Polar Encodings for Arbitrary-Oriented Ship Detection in SAR Images