Research map / Continual learning / Class-incremental
Inherit With Distillation and Evolve With Contrast: Exploring Class Incremental Semantic Segmentation Without Exemplar Memory
Danpei Zhao, Bo Yuan, Zhenwei Shi
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 attenuated layer-aware weight to intermediate-layer distillation (ablation) | IDEC · VOC 15-1 | mIoU (all): 61.44 -> 62.20 | Improved |
| add output-logit distillation after IL-D (ablation) | IDEC · VOC 15-1 | mIoU (all): 62.20 -> 65.80 | Improved |
| add ARCL contrastive learning (ablation) | IDEC · VOC 15-1 | mIoU (all): 65.80 -> 66.35 | Improved |
| add ARCL contrastive learning (ablation) | IDEC · ADE 100-10 | mIoU (all): 31.10 -> 32.56 | Improved |
| add ARCL contrastive learning (ablation) | IDEC · ISPRS 2-1 | mIoU (all): 54.83 -> 57.50 | Improved |
| replace fixed threshold 0.7 with dynamic class-specific pseudo-labelling (DCPL) (ablation) | IDEC · VOC 15-1 | mIoU (all): 66.35 -> 67.32 | Improved |
| DCPL instead of fixed threshold on incremental classes (ablation) | IDEC · VOC 15-1 (classes 16-20) | mIoU: 38.00 -> 36.48 | Hurt |
| DCPL instead of no threshold (ablation) | IDEC · VOC 15-1 | mIoU (all): 61.87 -> 67.32 | Improved |
| DCPL instead of PLOP pseudo-labelling (ablation) | IDEC · VOC 15-1 | mIoU (all): 65.44 -> 67.32 | Improved |
| order IL-D before OL-D rather than OL-D first (ablation) | IDEC · VOC 15-1 | mIoU (all): 61.10 -> 65.80 | Improved |
| replace DeepLabv3 with DeepLabv3+ | IDEC · VOC 15-1 | mIoU (all): 67.32 -> 67.86 | Improved |
| replace ResNet-101 with Swin-T encoder | IDEC · ISPRS 2-1 | mIoU (all): 58.03 -> 59.58 | Improved |
Closest papers by meaning
- Uncertainty-Aware Contrastive Distillation for Incremental Semantic Segmentation (TPAMI 2022 · similarity 0.93)
- Incrementer: Transformer for Class-Incremental Semantic Segmentation With Knowledge Distillation Focusing on Old Class (CVPR 2023 · similarity 0.92)
- Decomposed Knowledge Distillation for Class-Incremental Semantic Segmentation (NeurIPS 2022 · similarity 0.92)
- Modeling the Background for Incremental Learning in Semantic Segmentation (CVPR 2020 · similarity 0.92)
- Adaptive Prototype Replay for Class Incremental Semantic Segmentation (AAAI 2025 · similarity 0.89)
- CoMBO: Conflict Mitigation via Branched Optimization for Class Incremental Segmentation (CVPR 2025 · similarity 0.88)
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.
