Research map / Detection and segmentation / CNN features
DAPE: Harmonizing Content-Position Encoding for Versatile Dense Visual Prediction
Xiuquan Hou, Meiqin Liu, Senlin Zhang, Shaoyi Du
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 Shifted Query Sampler to align content and position queries (ablation) | DINO · COCO val2017 (12 epochs) | AP: 49.9 -> 50.6 | Improved |
| add Low-Rank Position Encoder (ablation) | DINO · COCO val2017 (12 epochs) | AP: 49.9 -> 50.9 | Improved |
| add one-to-many matching in the encoder (ablation) | DINO · COCO val2017 (12 epochs) | AP: 49.9 -> 50.9 | Improved |
| combine SQS, LRPE and one-to-many matching (ablation) | DINO · COCO val2017 (12 epochs) | AP: 49.9 -> 51.8 | Improved |
| use LRPE instead of 2D relation encoding (ablation) | DAPE · COCO val2017 (12 epochs) | training memory (G): 7.1G -> 5.3G | Improved |
| use LRPE instead of 2D relation encoding (ablation) | DAPE · COCO val2017 (12 epochs) | AP: 51.7 -> 51.8 | No clear effect |
| use LRPE instead of 2D relation encoding (ablation) | Relation-DETR · COCO val2017 (12 epochs) | training memory (G): 7.2G -> 5.5G | Improved |
| use LRPE instead of 2D relation encoding (ablation) | Relation-DETR · COCO val2017 (12 epochs) | AP: 51.7 -> 51.6 | No clear effect |
| increase LRPE rank from 4 to 16 (ablation) | DAPE · COCO val2017 (12 epochs) | AP: 51.5 -> 51.8 | Improved |
| increase LRPE rank beyond 16 (32, 64, 128) (ablation) | DAPE · COCO val2017 (12 epochs) | AP: 51.8 -> 51.7 (rank 32), 52.0 (rank 128) | No clear effect |
| fine-tune with LoRA instead of Meta-DAPE meta-learning (5-shot) | DAPE few-shot · COCO few-shot detection (5-shot) | nAP: nAP 12.9 -> 5.7; bAP 31.5 -> 39.8 | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- Relation-DETR — Relation DETR: Exploring Explicit Position Relation Prior for Object Detection (ECCV 2024) · better on 3 of 3 shared results
- DI-MaskDINO — DI-MaskDINO: A Joint Object Detection and Instance Segmentation Model (NeurIPS 2024) · better on 1 of 1 shared result
- MaskDINO — Mask DINO: Towards a Unified Transformer-Based Framework for Object Detection and Segmentation (CVPR 2023) · better on 1 of 1 shared result
- Meta-DETR — Meta-DETR: Image-Level Few-Shot Detection with Inter-Class Correlation Exploitation (TPAMI 2022) · 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
- Parabolic Position Encoding: Vision-Centric, Principled, Extrapolatable, General (NeurIPS 2026 · similarity 0.86)
- Perception Encoder: The best visual embeddings are not at the output of the network (NeurIPS 2025 · similarity 0.84)
- DAPE: Data-Adaptive Positional Encoding for Length Extrapolation (NeurIPS 2024 · similarity 0.84)
- Dense Video Object Captioning from Disjoint Supervision (ICLR 2025 · similarity 0.83)
- DPPE: Rethinking Camera-Based Positional Encoding for Scaling Multi-View Transformers (NeurIPS 2026 · similarity 0.83)
- Refining CLIP's Spatial Awareness: A Visual-Centric Perspective (ICLR 2025 · similarity 0.83)
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.
