Research map / Detection and segmentation / Open-set and few-shot detection
Prompt-Free Unknown Label Generation for Open World Detection in Remote Sensing
Abdullah Azeem, Ruisheng Wang, Qingquan Li, Abubakar Siddique
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 DHGA hierarchical graph attention to the baseline (ablation) | Deformable DETR + CLIP baseline with ORE-style unknown detection · DOTA-v2 Task 1 | Unknown Recall: 20.5 -> 28.0 | Improved |
| add DHGA hierarchical graph attention to the baseline (ablation) | Deformable DETR + CLIP baseline · DOTA-v2 Task 1 | Known mAP: 44.7 -> 48.9 | Improved |
| add Scene Context Token on top of DHGA (ablation) | HSGDet · DOTA-v2 Task 1 | Unknown Recall: 28.0 -> 33.8 | Improved |
| add CR2T on top of DHGA and SCT (ablation) | HSGDet · DOTA-v2 Task 1 | Unknown Recall: 33.8 -> 41.2 | Improved |
| add CR2T on top of DHGA and SCT (ablation) | HSGDet · DOTA-v2 Task 1 | Wilderness Impact: 08.4 -> 05.8 | Improved |
| add hierarchical parent to CR2T inputs (ablation) | CR2T · DOTA-v2 Task 1 | Unknown Recall: 37.4 -> 38.6 | Improved |
| add sibling context to CR2T inputs (ablation) | CR2T · DOTA-v2 Task 1 | Text Alignment: 0.63 -> 0.66 | Improved |
| use full CR2T (visual, scene context, parent) instead of visual-only (ablation) | CR2T · DOTA-v2 Task 1 | Text Alignment: 0.56 -> 0.79 | Improved |
| raise masked pseudo-unknown class ratio from 10% to 30% (ablation) | CR2T training · DOTA-v2 Task 1 | Unknown Recall: 39.2 -> 41.2 | Improved |
| use HSGDet instead of prior OWOD methods on COCO across incremental tasks | HSGDet · COCO Task 4 | Known mAP: 81.2 vs OW-OVD 64.8 | Improved |
| use HSGDet instead of SkySense-O on remote sensing benchmarks | HSGDet · FAIR1M and DIOR | Known mAP / Unknown Recall: +6.6 and +6.4 K-mAP, +9.9 and +9.2 U-R | Improved |
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.
- OW-OVD — OW-OVD: Unified Open World and Open Vocabulary Object Detection (CVPR 2025) · better on 4 of 4 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
- Locate Anything on Earth: Advancing Open-Vocabulary Object Detection for Remote Sensing Community (AAAI 2025 · similarity 0.90)
- InstructSAM: A Training-free Framework for Instruction-Oriented Remote Sensing Object Recognition (NeurIPS 2025 · similarity 0.88)
- OpenRSD: Towards Open-prompts for Object Detection in Remote Sensing Images (ICCV 2025 · similarity 0.87)
- SkySense-O: Towards Open-World Remote Sensing Interpretation with Vision-Centric Visual-Language Modeling (CVPR 2025 · similarity 0.86)
- Towards Open-Vocabulary Remote Sensing Image Semantic Segmentation (AAAI 2025 · similarity 0.86)
- SkyScript: A Large and Semantically Diverse Vision-Language Dataset for Remote Sensing (AAAI 2024 · similarity 0.85)
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.
