Research map / Continual learning / Class-incremental
Cs2K: Class-specific and Class-shared Knowledge Guidance for Incremental Semantic Segmentation
Wei Cong, Yang Cong, Yuyang Liu, Gan Sun
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 prototype-guided pseudo labeling (ablation) | Cs2K (PLOP) · Pascal VOC 2012 15-1 | mIoU (all): 65.3 -> 70.4 | Improved |
| add prototype-guided class adaptation (ablation) | Cs2K (PLOP) · Pascal VOC 2012 15-1 | mIoU (all): 68.7 -> 70.4 | Improved |
| add weight-guided selective consolidation (ablation) | Cs2K (PLOP) · Pascal VOC 2012 15-1 | mIoU (all): 48.6 -> 70.4 | Improved |
| add self-prototype augmentation (ablation) | Cs2K (PLOP) · Pascal VOC 2012 15-1 | mIoU (all): 69.0 -> 70.4 | Improved |
| add inter-prototype augmentation (ablation) | Cs2K (PLOP) · Pascal VOC 2012 15-1 | mIoU (all): 69.7 -> 70.4 | Improved |
| prototype-guided pseudo labels versus entropy-filtered PL (ablation) | Cs2K · Pascal VOC 2012 10-1 final step | mIoU: 54.0 -> 61.5 | Improved |
| prototype-guided pseudo labels versus naive pseudo labels (ablation) | Cs2K · Pascal VOC 2012 10-1 final step | mIoU: 58.3 -> 61.5 | Improved |
| selective weight consolidation versus uniform weight fusion (ablation) | Cs2K · Pascal VOC 2012 10-1 final step | mIoU: 60.3 -> 61.5 | Improved |
| add Cs2K to MiB | MiB · Pascal VOC 2012 5-3 | mIoU (all): 46.7 -> 56.2 | Improved |
| add Cs2K to MiB vs MiB+EWF on 10-1 old classes | MiB · Pascal VOC 2012 10-1 (classes 0-10) | mIoU: 56.0 -> 43.0 | Hurt |
| Cs2K vs MiB+EWF on ADE20K 100-5 | MiB · ADE20K 100-5 | mIoU (all): 32.1 -> 34.2 | Improved |
Closest papers by meaning
- Replay Without Saving: Prototype Derivation and Distribution Rebalance for Class-Incremental Semantic Segmentation (TPAMI 2025 · similarity 0.93)
- Incrementer: Transformer for Class-Incremental Semantic Segmentation With Knowledge Distillation Focusing on Old Class (CVPR 2023 · similarity 0.93)
- SSUL: Semantic Segmentation with Unknown Label for Exemplar-based Class-Incremental Learning (NeurIPS 2021 · similarity 0.92)
- Preparing the Future for Continual Semantic Segmentation (ICCV 2023 · similarity 0.92)
- Adaptive Prototype Replay for Class Incremental Semantic Segmentation (AAAI 2025 · similarity 0.90)
- CoMBO: Conflict Mitigation via Branched Optimization for Class Incremental Segmentation (CVPR 2025 · similarity 0.90)
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.
