RiO-DETRv2: Geometry-Aware Uncertainty Modeling Meets Real-Time Oriented Object Detection Transformer

Zhangchi Hu; Zilong He; Yifan Zhao; Xiangyu Yu; Jie Chen; Peixi Wu; Yansong Peng; Hebei Li

cs.CVOct 21, 2026

RiO-DETRv2: Geometry-Aware Uncertainty Modeling Meets Real-Time Oriented Object Detection Transformer

Authors: Zhangchi Hu, Zilong He, Yifan Zhao, Xiangyu Yu, Jie Chen, Peixi Wu, Yansong Peng, Hebei Li,

Organizations: University of Science and Technology of China · Institute of Artificial Intelligence, Hefei Comprehensive National Science Center

Abstract

RiO-DETR is the first real-time end-to-end transformer for oriented object detection, yet further localization gains require a more expressive box representation. Although distributional regression has succeeded in horizontal DETRs, extending it to oriented detection introduces geometry-misaligned boundary modeling and topology-inconsistent orientation refinement. We present RiO-DETRv2, a geometry-aware uncertainty modeling framework for real-time oriented object detection. To resolve the geometry mismatch, Object-Aligned Boundary Distribution Refinement represents box boundaries as probability distributions in an object-aligned coordinate system and progressively refines them across decoder layers. For better orientation topology consistency, we introduce Circular-to-Local Angular Residual Refinement that transitions from global periodic search to local fine-grained estimation. Unified OBB Quality-Aligned Learning further aligns classification, boundary and angular refinement with complete-OBB localization quality. Together, these designs improve localization without compromising efficiency, achieving a better balance between speed and accuracy. Specifically, RiO-DETRv2-l/x achieves 53.7/54.1 mAP50-95 on DOTA, outperforming RiO-DETR-l/x by 1.2/1.5 points with nearly unchanged computational cost. Code is available at https://github.com/RicePasteM/RiO-DETRv2.

Figures & tables

Figure 1: (A) Representative localization errors that are counted as TPs under the IoU threshold of 0.5 but fail the stricter 0.75 matching criterion, revealing the need for more precise OBB localization. (B) Comparisons with other detectors of AP ${}_{50\text{-}95}$ and AP 50 versus inference latency on DOTA-1.0 under single-scale training and testing protocol.

Figure 2: Overview of RiO-DETRv2. The OBB is factorized into object-aligned boundary geometry and periodic orientation. Side-distance distributions are progressively refined in the object frame, while the angular branch transitions from global circular search to local residual refinement. Both are jointly optimized by complete-OBB localization quality.

Figure 3: Global-axis decomposition entangles boundary distributions with orientation, whereas object-aligned modeling supervises side distances in the target-local frame and decodes them with the predicted angle, separating boundary geometry from angular estimation.

Model	Params.	FLOPs	Lat.	AP 50 by category															Overall performance
Model	Params.	FLOPs	Lat.	PL	BD	BR	GTF	SV	LV	SH	TC	BC	ST	SBF	RA	HA	SP	HC	mAP ${}_{50\text{-}95}$	AP 50	AP 75
RTMDet-R-t	4.9M	20G	8.3	89.2	80.0	47.9	69.7	82.1	83.3	88.6	90.9	86.3	86.9	59.9	62.3	74.2	72.0	57.0	47.4	75.4	50.6
YOLO26n-obb	2.5M	14G	2.8	89.7	84.7	52.1	71.1	83.1	78.9	88.7	91.1	87.3	86.9	58.5	69.1	76.6	82.8	65.1	49.1	77.4	53.4
RiO-DETR-n	4.0M	17G	2.7	87.3	86.1	54.9	73.6	80.5	85.1	88.0	90.8	87.4	86.9	60.9	72.0	76.8	73.4	71.8	47.1	78.5	48.2
RiO-DETRv2-n	4.3M	18G	2.8	88.8	86.2	57.6	76.6	82.0	85.2	89.0	90.9	88.0	87.0	63.7	71.2	77.3	72.7	70.5	49.3	79.1	50.9
PPYOLOE-R-s	8.1M	44G	8.5	88.8	78.4	45.9	66.9	80.1	82.8	88.2	90.6	79.7	86.4	56.1	64.0	65.0	79.5	50.4	40.6	73.5	39.2

Table 1: Comparison with representative real-time oriented object detectors on the DOTA-1.0 test set. For each matched model scale, the best and second-best overall results are highlighted in bold and underlined , respectively.

Model	Para.	FLOPs	Lat.	mAP ${}_{50\text{-}95}$	AP 50	AP 75
YOLO26n-obb	2.5M	10G	2.23	45.79	71.88	46.92
RiO-DETR-n	4.0M	11G	2.20	44.06	71.92	44.57
RiO-DETRv2-n	4.3M	11G	2.27	46.47	72.21	47.69
YOLO26s-obb	9.8M	39G	2.84	48.41	74.07	51.30
RiO-DETR-s	8.2M	33G	3.01	47.94	74.44	49.09
RiO-DETRv2-s	8.9M	33G	3.29	49.95	75.13	51.29

Table 2: Performance comparison with state-of-the-art real-time oriented object detectors on DIOR-R.

Model	Params.	FLOPs	Lat.	mAP ${}_{50\text{-}95}$	AP 50	AP 75
Oriented R-CNN †	41.1M	297G	–	11.2	33.2	4.3
$\mathrm{S}^{2}\mathrm{A}$ -Net †	38.6M	271G	–	10.8	33.4	3.3
Oriented RepPoints †	36.6M	269G	–	13.0	40.3	4.2
ARS-DETR †	41.1M	250G	25.32	14.3	41.1	5.8
DCFL ( $\mathrm{S}^{2}\mathrm{A}$ -Net) †	38.6M	271G	–	17.5	49.6	7.9
RHINO-DETR	50.8M	383G	35.56	17.7	53.9	5.7

Table 3: Comparison with state-of-the-art oriented object detectors on the AI-TOD-R benchmark.

\dagger

denotes results reported by the official benchmark ( Xu et al. 2024 ) .

Figure 4: Visualization of geometry-aware uncertainty refinement. The top row compares the initial and final OBBs, while the lower rows show layer-wise boundary and angular distributions. Boundary predictions progressively concentrate in the object-aligned frame, and angular refinement transitions from large periodic corrections to fine local adjustments.

Dataset	OABDR	CTLAR	UniOBB-QAL	Params.	FLOPs	AP ${}_{50\text{-}95}$	AP 50
DOTA-1.0				18.6M	158G	51.6	80.6
	✓			19.5M	158G	52.2	81.0
		✓		18.7M	158G	51.9	81.1
	✓	✓		19.5M	158G	52.7	81.2
	✓	✓	✓	19.5M	158G	53.2	80.8
DIOR-R				18.6M	97G	48.0	75.7

Table 4: Component ablation study of RiO-DETRv2-m.

Boundary Representation (on DIOR-R)	AP ${}_{50\text{-}95}$	AP 50	AP 75
Original RiO-DETR (Base)	51.80	77.43	53.82
w/ Global-axis distribution	52.19	77.22	54.20
w/ Object-aligned distribution (Ours)	52.88	77.74	54.66

Table 5: Ablation study on OABDR of RiO-DETRv2-x.

Angular Refinement (on DIOR-R)	AP ${}_{50\text{-}95}$	AP 50	AP 75
Original RiO-DETR w/ OABDR (Base)	52.88	77.74	54.66
w/ Linear angular distribution	52.66	77.31	53.82
w/ Circular distribution	53.02	78.14	54.78
w/ Circular-to-local refinement (Ours)	53.16	78.39	55.13

Table 6: Ablation study on CTLAR of RiO-DETRv2-x.

Appendix figures & tables7 assets

Supplementary material from the paper’s appendix.

Appendix

Setting	RiO-DETRv2-x	RiO-DETRv2-l	RiO-DETRv2-m	RiO-DETRv2-s	RiO-DETRv2-n
Backbone Name	HGNetv2-B5	HGNetv2-B4	HGNetv2-B2	HGNetv2-B0	HGNetv2-B0
Input Resolution	$1024^{2}$	$1024^{2}$	$1024^{2}$	$1024^{2}$	$1024^{2}$
Embedding Dimension	384	256	256	224	128
Feedforward Dimension	2048	1024	1024	1024	512
Decoder Layers	4	4	3	3	3
Queries	300	300	300	300	300

Appendix

Table 7: Hyperparameter configurations for RiO-DETRv2 models.

Figure 5: Behavior across GT aspect-ratio groups on DIOR-R. (a) The standard angular error develops a pronounced secondary mode near

90^{\circ}

for near-square objects, directly exposing occasional canonical-axis exchange. (b) Despite this parameterization ambiguity, matched-box IoU remains nearly unchanged, while angular distribution entropy increases toward the square limit.

Figure 6: Risk–coverage analysis over all top-300 predictions on DIOR-R. The left panel presents the complete risk–coverage curves, and the right panel shows the risk difference between the angle-aware estimator and the mean of the two boundary-only runs.

Uncertainty signal	Spearman $\rho$	10% retained	All matched
$H_{\mathrm{bnd}}$ vs. $1-\mathrm{IoU}_{\mathrm{OBB}}$	$0.782\,[0.774,0.788]$	$0.049$	$0.180$
$\bar{H}_{\theta}$ vs. angle error	$0.672\,[0.662,0.682]$	$0.240^{\circ}$	$1.877^{\circ}$
$H_{\theta}^{L}$ vs. angle error	$0.685\,[0.674,0.696]$	$0.237^{\circ}$	$1.877^{\circ}$

Appendix

Table 8: Relationship between distributional uncertainty and localization error on matched DIOR-R detections. Spearman correlation is reported with a 95% image-level bootstrap confidence interval. “10% retained” denotes the 10% predictions with the lowest estimated uncertainty.

GT aspect ratio	#GT	R@50 (%)	R@75 (%)	Mean IoU	$E_{180}$ MAE	$E_{\mathrm{axis}}$ MAE	Axis switch (%)	Angle entropy
$1.0\leq r<1.2$	27,035	92.64	72.26	0.817	$25.71^{\circ}$	$5.29^{\circ}$	20.69%	0.515
$1.2\leq r<1.5$	11,521	91.53	71.30	0.814	$7.83^{\circ}$	$3.46^{\circ}$	4.14%	0.451
$r\geq 1.5$	85,882	90.62	72.68	0.822	$1.87^{\circ}$	$1.81^{\circ}$	0.05%	0.380

Appendix

Table 9: Detection and orientation statistics on DIOR-R grouped by GT aspect ratio. R@50 and R@75 use independent class-aware one-to-one matching.

E_{\mathrm{axis}}

is a diagnostic measure for near-

90^{\circ}

axis exchange, not a replacement for the standard angular error.

Figure 7: Error-retention curves on DIOR-R. Predictions are retained in ascending order of estimated uncertainty. Left: complete OBB localization risk ranked by boundary entropy. Right: angular error ranked by mean angular entropy. Lower curves indicate more effective uncertainty ranking.

Score estimator	ECE 50	NLL 50	Brier 50	AURC 50	Q-ECE	$\rho(\mathrm{conf.},q_{50})$
Boundary LQE (run 1)	$0.0494$	$0.0698$	$0.0113$	$0.8607$	$0.0552$	$0.3011$
Boundary LQE (run 2)	$0.0490$	$0.0691$	$0.0114$	$0.8596$	$0.0548$	$0.3025$
Angle-aware LQE	$\mathbf{0.0455}$	$\mathbf{0.0649}$	$\mathbf{0.0106}$	$0.8599$	$\mathbf{0.0514}$	$0.3021$

Appendix

Table 10: Calibration over all top-300 predictions on DIOR-R. The two boundary-only rows correspond to independent training runs. The angle-aware estimator introduces only 113 additional parameters. Lower is better for ECE, NLL, Brier, AURC, and Q-ECE.

Object Detection

Object detection locates and classifies objects in images or video, balancing localization accuracy, robustness, and inference speed.

See all 624 papers

RiO-DETR: DETR for Real-time Oriented Object Detection
Mar 10, 2026Zhangchi Hu, Yifan Zhao, Yansong Peng +8Object Detection Detection Transformer
University of Science and Technology of China · Tsinghua University · Huawei Technologies Co., Ltd. +1
Real-Time Oriented Object Detection Transformer in Remote Sensing Images
Mar 16, 2026Zeyu Ding, Yong Zhou, Jiaqi Zhao +4Object Detection Detection Transformer
School of Computer Science and Technology / School of Artificial Intelligence, China University of Mining and Technology, Xuzhou 221116, China · Mine Digitization Engineering Research Center of the Ministry of Education, Xuzhou 221116, China · Jiangsu Provincial Industrial Technology Engineering Center for Intelligent Sensing and Emergency IoT in Underground Space, Xuzhou 221116, China +1
OrientedFormer: An End-to-End Transformer-Based Oriented Object Detector in Remote Sensing Images
Sep 29, 2024Zeyu Ding, Yong Zhou, Jiaqi Zhao +4Remote Sensing Image Understanding Detection Transformer
School of Computer Science and Technology, China University of Mining and Technology, and with Mine Digitization Engineering Research Center of the Ministry of Education, and also with Innovation Research Center of Disaster Intelligent Prevention and Emergency Rescue, China University of Mining and Technology, Xuzhou 221116, China · School of Electrical Engineering and Computer Science, University of Ottawa, Ottawa, ON K1N 6N5, Canada

Mar 10, 2026J/K move · Enter open · S save

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

Zhangchi Hu, Yifan Zhao, Yansong Peng, Wenzhang Sun+7

University of Science and Technology of China · Tsinghua University · Huawei Technologies Co., Ltd.+1

Object Detection Detection Transformer Real-Time Object Detection Aerial Object Detection

Open paper AskPDF arXiv

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. GitHub Repository: https://github.com/RicePasteM/RiO-DETR.

Abstract

Object Detection

Explore similar work

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

Real-Time Oriented Object Detection Transformer in Remote Sensing Images

OrientedFormer: An End-to-End Transformer-Based Oriented Object Detector in Remote Sensing Images