Research map / Detection and segmentation / CNN features
DQ-DETR: Dual Query Detection Transformer for Phrase Extraction and Grounding
Shilong Liu, Shijia Huang, Feng Li, Hao Zhang, Yaoyuan Liang, Hang Su, Jun Zhu, Lei Zhang
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 |
|---|---|---|---|
| use dual queries to separate object localization and phrase localization (ablation) | DQ-DETR · Flickr30k Entities | CMAP50: 66.68 -> 69.86 | Improved |
| add text mask attention (ablation) | DQ-DETR · Flickr30k Entities | CMAP50: 66.68 -> 68.26 | Improved |
| add positional query sharing (ablation) | DQ-DETR with dual queries and text mask attention · Flickr30k Entities | CMAP50: 69.86 -> 70.63 | Improved |
| add text mask attention (ablation) | DQ-DETR · Flickr30k Entities | Recall@1: 75.44 -> 76.58 | Improved |
| add dual queries (ablation) | DQ-DETR with text mask attention · Flickr30k Entities | Recall@1: 76.58 -> 78.87 | Improved |
| add positional query sharing (ablation) | DQ-DETR with dual queries and text mask attention · Flickr30k Entities | Recall@1: 78.87 -> 79.16 | Improved |
| increase the number of queries to 200 (ablation) | MDETR · Flickr30k Entities | CMAP50: 61.49 -> 56.89 | Hurt |
| train with DQ-DETR rather than MDETR | Flickr30k Entities, ResNet-101 pre-training · Flickr30k Entities | CMAP50: +5.8 CMAP$_{50}$ | Improved |
| train with DQ-DETR rather than MDETR | Flickr30k Entities, Any-Box protocol · Flickr30k Entities | Recall@1: 82.5 -> 83.2 | Improved |
| train with DQ-DETR rather than MDETR | COCO detection · COCO detection | AP: 36.5 -> 41.2 | Improved |
| use DQ-DETR instead of OFA for referring expression comprehension | RefCOCO testA · RefCOCO testA | Recall@1: 90.67 -> 91.04 | Improved |
| use DQ-DETR instead of SeqTR | RefCOCO testB · RefCOCO testB | Recall@1: 83.59 -> 83.51 | Hurt |
Later papers that compared with this method
4 later accepted papers report DQ-DETR as a baseline in their results tables; 4 of them report a better number on most of the results they share with it.
- Slim-DETR: Real-Time Tiny Object Detection with Efficient Interaction and Gaussian Query (ECCV 2026 · better on 3 of 3 shared results)
- Reviving the Discarded High-Resolution Feature for Transformer-Based Tiny Object Detection (NeurIPS 2026 · better on 1 of 1 shared result)
- Uncertainty-Aware Gradient Stabilization for Small Object Detection (ICCV 2025 · better on 1 of 1 shared result)
- Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection (ECCV 2024 · better on 1 of 1 shared result)
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
- Zero-Shot Grounding of Objects From Natural Language Queries (ICCV 2019 · similarity 0.90)
- Referencing Where to Focus: Improving Visual Grounding with Referential Query (NeurIPS 2024 · similarity 0.89)
- Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding (NeurIPS 2025 · similarity 0.88)
- Connecting the Dots: Training-Free Visual Grounding via Agentic Reasoning (AAAI 2026 · similarity 0.87)
- GroundVLP: Harnessing Zero-Shot Visual Grounding from Vision-Language Pre-training and Open-Vocabulary Object Detection (AAAI 2024 · similarity 0.87)
- SimVG: A Simple Framework for Visual Grounding with Decoupled Multi-modal Fusion (NeurIPS 2024 · 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.
