Research map / Detection and segmentation / Open-set and few-shot detection
OpenDPR: Open-Vocabulary Change Detection via Vision-Centric Diffusion-Guided Prototype Retrieval for Remote Sensing Imagery
Qi Guo, Jue Wang, Yinhe Liu, Yanfei Zhong
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 |
|---|---|---|---|
| replace CLIP image-text identification with diffusion-guided visual prototype retrieval | M-C-I · WHU-CD (unsupervised) | IoU: 40.6 -> 54.3 | Improved |
| replace CLIP image-text identification with visual prototype retrieval | M-C-I · Hi-UCD mini (multi-class) | mIoU: 13.9 -> 17.5 | Improved |
| use DINOv2 encoder instead of CLIP (ablation) | OpenDPR (LLM prompts, max) · WHU-CD | IoU: 47.40 -> 54.31 | Improved |
| use DINOv2 encoder instead of CLIP (ablation) | OpenDPR (LLM prompts, max) · SECOND | mIoU: 26.98 -> 28.55 | Improved |
| use global-max matching instead of category-mean (ablation) | OpenDPR (DINOv2) · SECOND | mIoU: 21.27 -> 28.55 | Improved |
| use LLM-generated prompts instead of naive prompts (ablation) | OpenDPR (DINOv2, max) · WHU-CD | IoU: 48.92 -> 54.31 | Improved |
| use LLM-generated prompts instead of naive prompts (ablation) | OpenDPR (CLIP, mean) · SECOND | mIoU: 25.76 -> 25.16 | Hurt |
| add VFM-based pseudo-label refinement to S2C (ablation) | S2C · WHU-CD | IoU: 61.84 -> 64.83 | Improved |
| add S2C weak supervision to OpenDPR | OpenDPR · WHU-CD | IoU: 54.31 -> 68.28 | Improved |
| add S2C weak supervision to OpenDPR | OpenDPR · Hi-UCD mini | mIoU: 17.52 -> 19.69 | Improved |
| vary number of prototypes (ablation) | OpenDPR / OpenDPR-W · four datasets | IoU | No clear effect |
Closest papers by meaning
- UniChange: Unifying Change Detection with Multimodal Large Language Model (CVPR 2026 · similarity 0.89)
- RSRCC: A Remote Sensing Regional Change Comprehension Benchmark Constructed via Retrieval-Augmented Best-of-𝑁 Ranking (NeurIPS 2026 · similarity 0.89)
- The Change You Want To Detect: Semantic Change Detection In Earth Observation With Hybrid Data Generationf (CVPR 2025 · similarity 0.88)
- Towards Generalizable Scene Change Detection (CVPR 2025 · similarity 0.87)
- Locate Anything on Earth: Advancing Open-Vocabulary Object Detection for Remote Sensing Community (AAAI 2025 · similarity 0.87)
- The STVchrono Dataset: Towards Continuous Change Recognition in Time (CVPR 2024 · similarity 0.86)
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.
