Research map / Detection and segmentation / Open-set and few-shot detection
SLIP-RS: Structured-Attribute Language-Image Pre-Training for Remote Sensing Object Detection
Chenxu Wang, Yuxuan Li, Yunheng Li, Xiang Li, Jingyuan Xia, Qibin Hou
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 structured-attribute contrastive learning with attribute annotations (SLIP-RS*) (ablation) | SLIP-RS ConvNeXT-T · Xview fine-tune | AP50: 25.9 -> 27.2 | Improved |
| add attribute supervision (SLIP-RS*) to detector pre-training (ablation) | SLIP-RS ConvNeXT-T · DOTA-v2.0 HBB val | mAP: 45.10 -> 44.43 | Hurt |
| add attribute supervision (SLIP-RS*) to detector pre-training (ablation) | SLIP-RS ConvNeXT-T · DIOR-H test | AP50: 79.97 -> 79.94 | No clear effect |
| add RS-C data to RS-O training data (ablation) | SLIP-RS ConvNeXT-T · DOTA-v2.0 HBB val | mAP: 41.24 -> 45.10 | Improved |
| add RS-C data to RS-O training data (ablation) | SLIP-RS ConvNeXT-T · DIOR-H test | AP50: 76.10 -> 79.97 | Improved |
| scale backbone from ConvNeXT-T to ConvNeXT-L | SLIP-RS* · DOTA-v2.0 HBB val | mAP: 44.43 -> 47.14 | Improved |
| scale backbone from ConvNeXT-T to ConvNeXT-L | SLIP-RS* · Plane Purpose attribute recognition | mAP: 64.85 -> 72.04 | Improved |
| expand attribute training data from RS-Attribute-O to RS-Attribute-15M (ablation) | SLIP-RS ConvNeXT-T · Plane Purpose attribute recognition | mAP: 67.61 -> 64.85 | Hurt |
| expand attribute training data from RS-Attribute-O to RS-Attribute-15M (ablation) | SLIP-RS ConvNeXT-T · Ship Usage attribute recognition | mAP: 34.00 -> 44.52 | Improved |
| replace monolithic category labels with structured attributes versus prior open-vocabulary detectors | SLIP-RS vs OpenRSD · Plane Purpose attribute recognition | mAP: 9.03 -> 72.04 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- OpenRSD — OpenRSD: Towards Open-prompts for Object Detection in Remote Sensing Images (ICCV 2025) · better on 4 of 4 shared results
- Faster-RCNN — Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (TPAMI 2016) · 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
- InstructSAM: A Training-free Framework for Instruction-Oriented Remote Sensing Object Recognition (NeurIPS 2025 · similarity 0.88)
- Remote Sensing Vision-Language Foundation Models without Annotations via Ground Remote Alignment (ICLR 2024 · similarity 0.86)
- Locate Anything on Earth: Advancing Open-Vocabulary Object Detection for Remote Sensing Community (AAAI 2025 · similarity 0.86)
- OW-VAP: Visual Attribute Parsing for Open World Object Detection (ICML 2025 · similarity 0.85)
- Quality-Driven Curation of Remote Sensing Vision-Language Data via Learned Scoring Models (NeurIPS 2025 · similarity 0.85)
- Towards Open-Vocabulary Remote Sensing Image Semantic Segmentation (AAAI 2025 · 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.
