Research map / Detection and segmentation / CNN features
Information-Bottleneck Driven Binary Neural Network for Change Detection
Kaijie Yin, Zhiyuan Zhang, Shu Kong, Tian Gao, Cheng-Zhong Xu, Hui Kong
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 BiCD auxiliary objective on backbone and 1-bit change generator (ablation) | 1-bit C-3PO · PCD TSUNAMI | F1-score (%): 84.7 -> 86.5 | Improved |
| place full auxiliary objective (with separability) at backbone only (ablation) | 1-bit C-3PO · PCD TSUNAMI | F1-score (%): 84.7 -> 84.8 | No clear effect |
| remove separability term in backbone auxiliary (ablation) | 1-bit C-3PO · PCD TSUNAMI | F1-score (%): 84.8 -> 85.2 | Improved |
| place full auxiliary objective at the 1-bit change generator (ablation) | 1-bit C-3PO · PCD TSUNAMI | F1-score (%): 84.7 -> 85.8 | Improved |
| add BiCD to 1-bit DR-TANet vs ReActNet | DR-TANet · PCD TSUNAMI | F1-score (%): 83.4 -> 85.1 | Improved |
| add BiCD to 1-bit C-3PO vs ReActNet | C-3PO · PCD GSV | F1-score (%): 71.2 -> 74.1 | Improved |
| add BiCD to 1-bit C-3PO vs ReActNet | C-3PO · VL_CMU_CD | F1-score (%): 69.9 -> 71.9 | Improved |
| add BiCD to 1-bit DR-TANet vs ReActNet | DR-TANet · VL_CMU_CD | F1-score (%): 62.6 -> 65.9 | Improved |
| add BiCD to 1-bit C-3PO vs ReActNet | C-3PO · LEVIR-CD | F1-score (%): 88.8 -> 89.9 | Improved |
| add feature-separability weight beta2 > 0 (best 0.08) (ablation) | 1-bit C-3PO · PCD TSUNAMI | F1-score (%): 84.7 -> 86.5 | Improved |
| binarize C-3PO (1-bit with BiCD) vs real-valued | C-3PO · edge latency | latency (ms): 392.8 -> 158.4 | Improved |
| binarize C-3PO with BiCD vs real-valued | C-3PO · PCD GSV | F1-score (%): 77.8 -> 74.1 | Hurt |
Closest papers by meaning
- MapFormer: Boosting Change Detection by Using Pre-change Information (ICCV 2023 · similarity 0.85)
- FM-ChangeNet: Learning Change through Pathwise Feature Transport (NeurIPS 2026 · similarity 0.85)
- The Change You Want To Detect: Semantic Change Detection In Earth Observation With Hybrid Data Generationf (CVPR 2025 · similarity 0.83)
- MUCD: Unsupervised Point Cloud Change Detection via Masked Consistency (AAAI 2025 · similarity 0.82)
- Feature Spectrum Learning for Remote Sensing Change Detection (CVPR 2025 · similarity 0.82)
- Towards Generalizable Scene Change Detection (CVPR 2025 · similarity 0.81)
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.
