Research map / Detection and segmentation / Open-set and few-shot detection
PSALM: Pixelwise Segmentation with Large Multi-modal Model
Zheng Zhang, yeyao ma, Enming Zhang, Xiang Bai
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 |
|---|---|---|---|
| feed mask tokens through the LLM instead of directly to mask generator (ablation) | PSALM · RefCOCO val | cIoU: 74.3 -> 76.1 | Improved |
| feed mask tokens through the LLM instead of directly to mask generator (ablation) | PSALM · A150-OV | mAP: 8.2 -> 9.3 | Improved |
| place mask tokens as suffix rather than prefix (ablation) | PSALM · A150-OV | mAP: 7.8 -> 9.3 | Improved |
| decouple mask prediction and classification (ablation) | PSALM · COCO Semantic Segmentation | mIoU: 42.7 -> 66.5 | Improved |
| use 100 mask proposals instead of a single mask (ablation) | PSALM · RefCOCO val | cIoU: 75.3 -> 76.5 | Improved |
| use [REF] token for sentence condition instead of avg_pooling (ablation) | PSALM · RefCOCO val | cIoU: 75.3 -> 76.1 | Improved |
| use avg_pooling for category condition instead of [REF] (ablation) | PSALM · A150-OV | mAP: 8.4 -> 9.3 | Improved |
| add vision-language alignment stage (ablation) | PSALM · RefCOCO val | cIoU: 71.7 -> 76.1 | Improved |
| add vision-language alignment stage (ablation) | PSALM · COCO-Pan | PQ: 54.9 -> 55.1 | No clear effect |
| joint training across tasks instead of task-specific training (ablation) | PSALM · RefCOCO val | cIoU: 76.5 -> 83.6 | Improved |
| joint training across tasks instead of task-specific training (ablation) | PSALM · COCO-Pan | PQ: 55.6 -> 55.9 | No clear effect |
| joint training across tasks instead of task-specific training (ablation) | PSALM · A150-OV | mAP: 7.0 -> 9.0 | Improved |
Later papers that compared with this method
7 later accepted papers report PSALM as a baseline in their results tables; 7 of them report a better number on most of the results they share with it.
- X-SAM: From Segment Anything to Any Segmentation (AAAI 2026 · better on 5 of 5 shared results)
- CDICS: Delving Into Fine-Grained Attribute for In-Context Segmentation via Compositional Prompts and Phased Decoupling (CVPR 2026 · better on 1 of 1 shared result)
- Rethinking MLLM Itself as a Segmenter with a Single Segmentation Token (CVPR 2026 · better on 1 of 1 shared result)
- HiMTok: Learning Hierarchical Mask Tokens for Image Segmentation with Large Multimodal Model (ICCV 2025 · better on 5 of 6 shared results)
- Refer to Any Segmentation Mask Group With Vision-Language Prompts (ICCV 2025 · better on 2 of 3 shared results)
- LangHOPS: Language Grounded Hierarchical Open-Vocabulary Part Segmentation (NeurIPS 2025 · better on 2 of 2 shared results)
- ObjectRelator: Enabling Cross-View Object Relation Understanding Across Ego-Centric and Exo-Centric Perspectives (ICCV 2025 · better on 1 of 1 shared result)
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.
- OMG-Seg — OMG-Seg: Is One Model Good Enough For All Segmentation? (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
- PixelLM: Pixel Reasoning with Large Multimodal Model (CVPR 2024 · similarity 0.84)
- Test-Time Adaptation of Vision-Language Models for Open-Vocabulary Semantic Segmentation (NeurIPS 2025 · similarity 0.83)
- Effective SAM Combination for Open-Vocabulary Semantic Segmentation (CVPR 2025 · similarity 0.83)
- RankSEG-RMA: An Efficient Segmentation Algorithm via Reciprocal Moment Approximation (NeurIPS 2025 · similarity 0.83)
- Learning Hierarchical Image Segmentation For Recognition and By Recognition (ICLR 2024 · similarity 0.82)
- LQMFormer: Language-aware Query Mask Transformer for Referring Image Segmentation (CVPR 2024 · similarity 0.82)
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.
