Research map / Detection and segmentation / CNN features
Relation DETR: Exploring Explicit Position Relation Prior for Object Detection
Xiuquan Hou, Meiqin Liu, Senlin Zhang, Ping Wei, Badong Chen, Xuguang Lan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add position relation encoder to DINO (ablation) | DINO ResNet-50 · COCO val2017 1x | AP: 49.9 -> 50.3 | Improved |
| add VariFocal loss (ablation) | DINO · COCO val2017 1x | AP: 49.9 -> 50.9 | Improved |
| add position relation on top of VFL (ablation) | DINO + VFL · COCO val2017 1x | AP: 50.9 -> 51.2 | Improved |
| add contrast relation pipeline (ablation) | DINO + VFL + Rel · COCO val2017 1x | AP: 51.2 -> 51.7 | Improved |
| add relation encoder to Deformable-DETR | Deformable-DETR · COCO val2017 1x | AP: 45.4 -> 47.0 | Improved |
| add relation encoder to DAB-Deformable-DETR | DAB-Deformable-DETR · COCO val2017 1x | AP: 46.3 -> 48.3 | Improved |
| add relation encoder to DN-Deformable-DETR | DN-Deformable-DETR · COCO val2017 1x | AP: 47.2 -> 47.3 | No clear effect |
| apply hybrid matching to DINO | DINO · COCO val2017 1x | AP: 49.9 -> 49.5 | Hurt |
| add relation encoder plus contrast pipeline to DINO instead of hybrid matching | DINO · COCO val2017 1x | AP: 49.9 -> 51.2 | Improved |
| replace DINO with Relation-DETR on defect dataset MSSD | DINO · MSSD | AP@50:95: 51.0 -> 57.4 | Improved |
| replace DINO with Relation-DETR on CSD | DINO · CSD | AP@50:95: 53.0 -> 54.4 | Improved |
| use Relation-DETR vs DINO with VFL | DINO with VFL · SA-Det-100k | AP@50:95: 43.7 -> 45.0 | Improved |
Later papers that compared with this method
6 later accepted papers report Relation-DETR as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.
- DAPE: Harmonizing Content-Position Encoding for Versatile Dense Visual Prediction (AAAI 2026 · better on 3 of 3 shared results)
- Integrating Diverse Assignment Strategies into DETRs (AAAI 2026 · better on 2 of 2 shared results)
- Order Matters: Competition-Guided Query Ordering for RNN-Based Object Detection (NeurIPS 2026 · better on 1 of 1 shared result)
- PaQ-DETR: Learning Pattern and Quality-Aware Dynamic Queries for Object Detection (CVPR 2026 · better on 0 of 1 shared result)
- Mr. DETR++: Instructive Multi-Route Training for Detection Transformers with Mixture-of-Experts (TPAMI 2026 · better on 1 of 1 shared result)
- MI-DETR: An Object Detection Model with Multi-time Inquiries Mechanism (CVPR 2025 · better on 3 of 3 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Salience-DETR — Salience DETR: Enhancing Detection Transformer with Hierarchical Salience Filtering Refinement (CVPR 2024) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Relation DETR+: Exploring Explicit Position Relation Prior for Dense Prediction (TPAMI 2026 · similarity 0.97)
- FP-DETR: Detection Transformer Advanced by Fully Pre-training (ICLR 2022 · similarity 0.89)
- RelationNet++: Bridging Visual Representations for Object Detection via Transformer Decoder (NeurIPS 2020 · similarity 0.88)
- EGTR: Extracting Graph from Transformer for Scene Graph Generation (CVPR 2024 · similarity 0.88)
- CP-DETR: Concept Prompt Guide DETR Toward Stronger Universal Object Detection (AAAI 2025 · similarity 0.87)
- UN-DETR: Promoting Objectness Learning via Joint Supervision for Unknown Object Detection (AAAI 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
