Research map / Detection and segmentation / CNN features
UP-DETR: Unsupervised Pre-Training for Object Detection With Transformers
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Pre-train DETR transformers with random query patch detection | DETR · PASCAL VOC test2007 | AP: 49.9 -> 56.1 (150 epochs); 54.1 -> 57.2 (300 epochs) | Helped |
| Pre-train DETR transformers with random query patch detection | DETR · COCO val2017 | AP: 39.7 -> 40.5 (150 epochs); 42.1 -> 42.8 (300 epochs) | Helped |
| Pre-train DETR transformers with random query patch detection | DETR · VOC 2007test one-shot detection | AP50: +22.8 seen; +15.8 unseen | Helped |
| Pre-train DETR transformers with random query patch detection | DETR · COCO val panoptic segmentation | PQ: 44.3 -> 44.5 | Helped |
| Pre-train without freezing the CNN or adding feature reconstruction (ablation) | UP-DETR · PASCAL VOC test2007 | AP50: 74.5 -> 74.0 | Hurt |
| Add a frozen CNN backbone during pre-training (ablation) | UP-DETR · PASCAL VOC test2007 | AP50: 74.0 -> 78.7 | Helped |
| Add feature reconstruction without freezing the CNN (ablation) | UP-DETR · PASCAL VOC test2007 | AP50: 74.0 -> 62.0 | Hurt |
| Add feature reconstruction alongside a frozen CNN (ablation) | UP-DETR · PASCAL VOC test2007 | AP50: 78.7 -> 78.7 | No clear effect |
| Add feature reconstruction alongside a frozen CNN (ablation) | UP-DETR · PASCAL VOC training convergence | AP50 | Helped |
| Use object query shuffling during multi-query pre-training (ablation) | UP-DETR · Pre-training | Effectiveness | No clear effect |
Closest papers by meaning
- MS-DETR: Efficient DETR Training with Mixed Supervision (CVPR 2024 · similarity 0.88)
- DETRs Beat YOLOs on Real-time Object Detection (CVPR 2024 · similarity 0.87)
- QDETRv: Query-Guided DETR for One-Shot Object Localization in Videos (AAAI 2024 · similarity 0.87)
- Frozen-DETR: Enhancing DETR with Image Understanding from Frozen Foundation Models (NeurIPS 2024 · similarity 0.87)
- CP-DETR: Concept Prompt Guide DETR Toward Stronger Universal Object Detection (AAAI 2025 · similarity 0.87)
- MI-DETR: An Object Detection Model with Multi-time Inquiries Mechanism (CVPR 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.