Research map / Continual learning / Class-incremental
Incremental Few-Shot Semantic Segmentation via Multi-Level Switchable Visual Prompts
Maoxian Wan, Kaige Li, Qichuan Geng, Weimin Shi, Zhong Zhou
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 the focus decomposition decoder (FDD) instead of a single head (ablation) | MSVP · PASCAL VOC IFSS 1-shot | mIoU-N: 44.89 -> 49.08 | Improved |
| add the background aggregation loss Lba instead of the vanilla loss Lvan (ablation) | MSVP · PASCAL VOC IFSS 1-shot | mIoU-B: 70.53 -> 73.04 | Improved |
| add the cross-modal query engine (CMQE) (ablation) | MSVP · PASCAL VOC IFSS 1-shot | mIoU-N: 22.20 -> 49.08 | Improved |
| add the CMQE (ablation) | MSVP · PASCAL VOC IFSS 1-shot | HM: 33.76 -> 58.71 | Improved |
| add the prompt switching strategy to the multi-level prompts (ablation) | MSVP · PASCAL VOC IFSS 1-shot | HM: 25.68 -> 58.71 | Improved |
| add the multi-level prompts (TP, SP, RP with switching) over vanilla prompt extension (ablation) | MSVP · PASCAL VOC IFSS 1-shot | mIoU-N: 47.32 -> 49.08 | Improved |
| add the multi-level prompts over vanilla prompt extension (ablation) | MSVP · PASCAL VOC IFSS 1-shot | HM: 55.05 -> 58.71 | Improved |
| vary number of task-persistent prompts Lg (peak at 24) (ablation) | MSVP · PASCAL VOC IFSS 1-shot | mIoU-B / mIoU-N | Mixed |
| vary vectors per stage-specific prompt Ls (best at 8) (ablation) | MSVP · PASCAL VOC IFSS 1-shot | mIoU | Mixed |
| vary cluster number h for region-unique prompts (best at 2) (ablation) | MSVP · PASCAL VOC IFSS 1-shot | mIoU-N | Mixed |
| vary similarity threshold for region-unique prompts (best at 0.7) (ablation) | MSVP · PASCAL VOC IFSS 1-shot | mIoU-N | Mixed |
| use MSVP instead of the prior best IFSS method SRAA | IFSS on VOC · PASCAL VOC IFSS 1-shot | mIoU-N: 18.8 -> 49.08 | Improved |
Closest papers by meaning
- Deconfound Semantic Shift and Incompleteness in Incremental Few-shot Semantic Segmentation (AAAI 2025 · similarity 0.90)
- SAM-Aware Graph Prompt Reasoning Network for Cross-Domain Few-Shot Segmentation (AAAI 2025 · similarity 0.89)
- Visual Prompting for Generalized Few-shot Segmentation: A Multi-scale Approach (CVPR 2024 · similarity 0.89)
- APSeg: Auto-Prompt Network for Cross-Domain Few-Shot Semantic Segmentation (CVPR 2024 · similarity 0.89)
- Few-Shot Incremental Learning via Foreground Aggregation and Knowledge Transfer for Audio-Visual Semantic Segmentation (AAAI 2025 · similarity 0.89)
- Bridge the Points: Graph-based Few-shot Segment Anything Semantically (NeurIPS 2024 · similarity 0.89)
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.
