Research map / Continual learning / Class-incremental
Palimpsest: Reconciling the CISS Trilemma for Incremental Nuclei Segmentation
Jiajia Li, Huisi Wu
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 Adaptive Residual Shading (logit-space background decoupling) (ablation) | PCS (no SCR) · MoNuSAC 2-1 | Mean Dice: 44.54 -> 68.33 | Improved |
| add Similarity-Aware Centroid Recalibration (ablation) | PCS + ARS · MoNuSAC 2-1 | Mean Dice: 68.33 -> 71.57 | Improved |
| merge adapters via PCS alone (ablation) | baseline adapter · MoNuSAC 2-1 | Mean Dice: 46.94 -> 44.54 | Hurt |
| add margin-based LARS (ablation) | Lnew only · MoNuSAC 2-1 | Mean Dice: 44.54 -> 65.94 | Improved |
| add foreground feature consistency loss LFFC (ablation) | Lnew + LARS · MoNuSAC 2-1 | Mean Dice: 65.94 -> 71.57 | Improved |
| replace uniform recalibration with similarity-aware SCR (ablation) | Uniform recalibration · MoNuSAC 2-1 | Mean Dice: 62.76 -> 71.57 | Improved |
| use no calibration vs SCR (ablation) | No calibration · MoNuSAC 2-1 | Mean Dice: 51.32 -> 71.57 | Improved |
| compare SCR with naive exemplar replay (ablation) | Naive Replay · MoNuSAC 2-1 | Mean Dice: 72.89 vs 71.57 | No clear effect |
| Palimpsest vs InSeg | InSeg · MoNuSAC 1-1 (4 steps) | Mean Dice: 69.42 -> 71.28 | Improved |
| Palimpsest vs InSeg | InSeg · CoNSeP 2-1 | Mean Dice: 71.08 -> 71.84 | Improved |
| Palimpsest vs InSeg on new classes | InSeg · MoNuSAC 3-1 | New-class Dice: 77.81 -> 77.75 | No clear effect |
Closest papers by meaning
- CiNuSeg: Class Incremental Nuclei Segmentation via Anchor-driven Consistency Learning with Dual Region Regularization (AAAI 2026 · similarity 0.89)
- Incremental Nuclei Segmentation from Histopathological Images via Future-class Awareness and Compatibility-inspired Distillation (CVPR 2024 · similarity 0.88)
- MATCH: Multi-faceted Adaptive Topo-Consistency for Semi-Supervised Histopathology Segmentation (NeurIPS 2025 · similarity 0.85)
- Weakly Supervised Gland Segmentation with Class Semantic Consistency and Purified Labels Filtration (AAAI 2025 · similarity 0.84)
- Semantic and Visual Crop-Guided Diffusion Models for Heterogeneous Tissue Synthesis in Histopathology (NeurIPS 2025 · similarity 0.84)
- Why Invariance is Not Enough for Biomedical Domain Generalization and How to Fix It (NeurIPS 2026 · similarity 0.84)
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.
