Research map / Detection and segmentation / Open-set and few-shot detection
Multi-modal Segment Anything Model for Camouflaged Scene Segmentation
Guangyu Ren, Hengyan Liu, Michalis Lazarou, Tania Stathaki
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 BLIP caption language prompt (ablation) | SAM baseline · CHAMELEON | weighted F-measure: 0.666 -> 0.706 | Improved |
| add BLIP caption language prompt (ablation) | SAM baseline · COD10K | MAE: 0.054 -> 0.046 | Improved |
| add BLIP caption language prompt (ablation) | SAM baseline · CAMO | S-alpha: 0.790 -> 0.784 | No clear effect |
| add vision (BLIP image embedding) prompt (ablation) | SAM + Language · CAMO | S-alpha: 0.784 -> 0.811 | Improved |
| add vision prompt (ablation) | SAM + Language · COD10K | E-measure: 0.833 -> 0.818 | Mixed |
| add dense embedding from SAM image embedding (ablation) | SAM + Language + Vision · CAMO | weighted F-measure: 0.676 -> 0.712 | Improved |
| add dense embedding from SAM image embedding (ablation) | SAM + Language + Vision · COD10K | MAE: 0.042 -> 0.034 | Improved |
| add multi-level adapter (ablation) | SAM + Language + Vision + Dense · CHAMELEON | S-alpha: 0.872 -> 0.923 | Improved |
| add multi-level adapter (ablation) | SAM + Language + Vision + Dense · COD10K | weighted F-measure: 0.706 -> 0.808 | Improved |
| add multi-level adapter (ablation) | SAM + Language + Vision + Dense · CAMO | MAE: 0.075 -> 0.059 | Improved |
| use MM-SAM instead of SAM-Adapter | SAM-Adapter · COD10K | S-alpha: 0.883 -> 0.896 | Improved |
| use MM-SAM instead of SAM-Adapter | SAM-Adapter · CHAMELEON | S-alpha: 0.896 -> 0.923 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- AlignSAM — AlignSAM: Aligning Segment Anything Model to Open Context via Reinforcement Learning (CVPR 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Relax Image-Specific Prompt Requirement in SAM: A Single Generic Prompt for Segmenting Camouflaged Objects (AAAI 2024 · similarity 0.90)
- Endow SAM with Keen Eyes: Temporal-spatial Prompt Learning for Video Camouflaged Object Detection (CVPR 2024 · similarity 0.88)
- SAM-CP: Marrying SAM with Composable Prompts for Versatile Segmentation (ICLR 2025 · similarity 0.87)
- Camouflage Anything: Learning to Hide using Controlled Out-painting and Representation Engineering (CVPR 2025 · similarity 0.87)
- CamSAM2: Segment Anything Accurately in Camouflaged Videos (NeurIPS 2025 · similarity 0.87)
- SAM 2: Segment Anything in Images and Videos (ICLR 2025 · similarity 0.87)
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.
