Research map / Video / Video segmentation and tracking
S2C: A Noise-Resistant Difference Learning Framework for Unsupervised Change Detection in VHR Remote Sensing Images
Lei Ding, Xibing Zuo, Haitao Guo, Jun Lu, Zhihui Gong, Xuanguang Liu, Jicang Lu
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 |
|---|---|---|---|
| use triplet temporal contrast (CTC) instead of spatial contrast (SC) (ablation) | S2C with efficient-SAM ViT-T · CLCD | F1: 31.55 -> 37.86 | Improved |
| add consistency regularization to spatial contrast (CSC vs SC) (ablation) | S2C with efficient-SAM ViT-T · CLCD | F1: 31.55 -> 32.84 | Improved |
| add CSC to CTC (ablation) | S2C with efficient-SAM ViT-T · CLCD | F1: 37.86 -> 40.19 | Improved |
| add grid sparsity loss Lspa (ablation) | S2C with efficient-SAM ViT-T · CLCD | F1: 40.19 -> 42.92 | Improved |
| add IoU refinement (ablation) | S2C with efficient-SAM ViT-S · CLCD | F1: 46.22 -> 47.69 | Improved |
| add IoU refinement (ablation) | S2C with DINOv2 ViT-B · CLCD | F1: 60.12 -> 61.38 | Improved |
| add IoU refinement (ablation) | S2C with DINOv2 ViT-B · SECOND binary | F1: 51.99 -> 53.86 | Improved |
| replace ResNet34 with a VFM backbone (DINOv2 ViT-B) (ablation) | S2C · CLCD | F1: 37.60 -> 58.12 | Improved |
| use fastSAM as backbone instead of efficient-SAM (ablation) | S2C · CLCD | F1: 42.92 -> 17.78 | Hurt |
| use efficient-SAM ViT-S instead of ViT-T (ablation) | S2C · CLCD | F1: 42.92 -> 46.22 | Improved |
| add S2C refinement to DINOv2 on Levir (vs. supervised SAM-CD) | S2C · Levir | F1: 91.82 (SAM-CD supervised) vs 46.69 | Hurt |
| add IoU refinement (ablation) | S2C with DINOv2 ViT-B · SECOND binary | Recall: Rec 42.41 -> 46.02, OA 84.55 -> 84.45 | Mixed |
Closest papers by meaning
- Changen2: Multi-Temporal Remote Sensing Generative Change Foundation Model (TPAMI 2024 · similarity 0.90)
- Exploring Generalizable Remote Sensing Change Detection via Low-Rank Exchange Adaptation of Vision Foundation Model (AAAI 2026 · similarity 0.89)
- The Change You Want To Detect: Semantic Change Detection In Earth Observation With Hybrid Data Generationf (CVPR 2025 · similarity 0.89)
- Feature Spectrum Learning for Remote Sensing Change Detection (CVPR 2025 · similarity 0.87)
- RSRCC: A Remote Sensing Regional Change Comprehension Benchmark Constructed via Retrieval-Augmented Best-of-𝑁 Ranking (NeurIPS 2026 · similarity 0.87)
- FM-ChangeNet: Learning Change through Pathwise Feature Transport (NeurIPS 2026 · 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.
