Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

RiO-DETR: DETR for Real-time Oriented Object Detection

Zhangchi Hu, Yifan Zhao, Yansong Peng, Wenzhang Sun, Xiangchen Yin, Jie Chen, Peixi Wu, Hebei Li, Xinghao Wang, Dongsheng Jiang, Xiaoyan Sun

TL;DR

RiO-DETR establishes a new speed--accuracy trade-off for real-time oriented detection, and proposes Content-Driven Angle Estimation by decoupling angle from positional queries, together with Rotation-Rectified Orthogonal Attention to capture complementary cues for reliable orientation.

Abstract

We present RiO-DETR: DETR for Real-time Oriented Object Detection, the first real-time oriented detection transformer to the best of our knowledge. Adapting DETR to oriented bounding boxes (OBBs) poses three challenges: semantics-dependent orientation, angle periodicity that breaks standard Euclidean refinement, and an enlarged search space that slows convergence. RiO-DETR resolves these issues with task-native designs while preserving real-time efficiency. First, we propose Content-Driven Angle Estimation by decoupling angle from positional queries, together with Rotation-Rectified Orthogonal Attention to capture complementary cues for reliable orientation. Second, Decoupled Periodic Refinement combines bounded coarse-to-fine updates with a Shortest-Path Periodic Loss for stable learning across angular seams. Third, Oriented Dense O2O injects angular diversity into dense supervision to speed up angle convergence at no extra cost. Extensive experiments on DOTA-1.0, DIOR-R, and FAIR-1M-2.0 demonstrate RiO-DETR establishes a new speed--accuracy trade-off for real-time oriented detection. Code will be made publicly available.

RiO-DETR: DETR for Real-time Oriented Object Detection

TL;DR

RiO-DETR establishes a new speed--accuracy trade-off for real-time oriented detection, and proposes Content-Driven Angle Estimation by decoupling angle from positional queries, together with Rotation-Rectified Orthogonal Attention to capture complementary cues for reliable orientation.

Abstract

We present RiO-DETR: DETR for Real-time Oriented Object Detection, the first real-time oriented detection transformer to the best of our knowledge. Adapting DETR to oriented bounding boxes (OBBs) poses three challenges: semantics-dependent orientation, angle periodicity that breaks standard Euclidean refinement, and an enlarged search space that slows convergence. RiO-DETR resolves these issues with task-native designs while preserving real-time efficiency. First, we propose Content-Driven Angle Estimation by decoupling angle from positional queries, together with Rotation-Rectified Orthogonal Attention to capture complementary cues for reliable orientation. Second, Decoupled Periodic Refinement combines bounded coarse-to-fine updates with a Shortest-Path Periodic Loss for stable learning across angular seams. Third, Oriented Dense O2O injects angular diversity into dense supervision to speed up angle convergence at no extra cost. Extensive experiments on DOTA-1.0, DIOR-R, and FAIR-1M-2.0 demonstrate RiO-DETR establishes a new speed--accuracy trade-off for real-time oriented detection. Code will be made publicly available.
Paper Structure (33 sections, 10 equations, 6 figures, 9 tables)

This paper contains 33 sections, 10 equations, 6 figures, 9 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. CNN-based Oriented Object Detection.
  4. Real-time Object Detection.
  5. DETR-based Oriented Object Detection.
  6. Methods
  7. Best of Both Worlds: Building a Strong Real-time Baseline
  8. Content-Driven Angle Estimation
  9. Geometry-Decoupled Query Encoding
  10. Rotation-Rectified Orthogonal Attention
  11. Decoupled Periodic Refinement
  12. Oriented Dense O2O
  13. Experiments
  14. Evaluation Datasets
  15. DOTA-1.0.
  16. ...and 18 more sections

Figures (6)

  • Figure 1: Comparisons with other detectors in terms of model size (left), latency (mid), and computational cost (right) on DOTA-1.0 under single-scale training and testing protocol. * denotes a community implemented version.
  • Figure 2: The main architecture of our proposed RiO-DETR. The framework highlights three key components: (A) Content-Driven Angle Estimation, (B) Rotation-Rectified Orthogonal Attention, and (C) Decoupled Periodic Refinement.
  • Figure 3: An intuitive illustration of (D) Decoupled Periodic Refinement and (E) Oriented Dense O2O.
  • Figure 4: Visualization of deformable attention sampling points on DOTA.
  • Figure 5: Visual illustration of layer-wise angular refinement for a specific instance.
  • ...and 1 more figures