Research map / Detection and segmentation / Open-set and few-shot detection
LoGoSeg: Integrating Local and Global Features for Open-Vocabulary Semantic Segmentation
Junyang Chen, Xiangbo Lv, Zhiqiang Kou, Xingdong Sheng, Ning Xu, Yiguo Qiao
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 object existence prior (ablation) | LoGoSeg baseline · A-847 / PC-59 | mIoU: 10.8 -> 11.5; 53.9 -> 56.4 | Improved |
| add contextual cross-modal fusion (ablation) | LoGoSeg (OP) · A-847 / PC-59 | mIoU: 11.5 -> 12.3; 56.4 -> 57.8 | Improved |
| add region alignment (ablation) | LoGoSeg (OP+CCF) · A-847 / PC-59 | mIoU: 12.3 -> 12.6; 57.8 -> 58.2 | Improved |
| add global context (GC) to LangQuery (ablation) | fusion module · A-847 | mIoU: 10.9 -> 11.8 | Improved |
| add directional attention (ablation) | fusion module · A-847 | mIoU: 11.8 -> 12.6 | Improved |
| increase fusion depth from 2 to 3 (ablation) | fusion module · A-847 | mIoU: 12.6 -> 11.9 | Hurt |
| use fusion depth 1 instead of 2 (ablation) | fusion module · A-847 | mIoU: 12.6 -> 11.5 | Hurt |
| set object prior weight lambda=1 instead of 0 (ablation) | region alignment · A-847 | mIoU: 11.8 -> 12.6 | Improved |
| use 8x8 instead of 6x6 region grid (ablation) | region alignment (lambda=0) · PAS-20b | mIoU: 76.9 -> 75.8 | Hurt |
| train on PC-59 instead of COCO-Stuff (ablation) | LoGoSeg · A-847 | mIoU: 12.6 -> 10.1 | Hurt |
| LoGoSeg vs SED | LoGoSeg (ViT-L/14) · PC-459 | mIoU: ViT-L 24.5 vs SED 22.6 | Improved |
Closest papers by meaning
- Emergent Open-Vocabulary Semantic Segmentation from Off-the-shelf Vision-Language Models (CVPR 2024 · similarity 0.90)
- Distilling Spectral Graph for Object-Context Aware Open-Vocabulary Semantic Segmentation (CVPR 2025 · similarity 0.90)
- Training-Free Open-Vocabulary Segmentation with Offline Diffusion-Augmented Prototype Generation (CVPR 2024 · similarity 0.90)
- Training-free Open-Vocabulary Semantic Segmentation via Diverse Prototype Construction and Sub-region Matching (AAAI 2025 · similarity 0.90)
- Open-Vocabulary Segmentation with Semantic-Assisted Calibration (CVPR 2024 · similarity 0.89)
- OpenWorldSAM: Extending SAM2 for Universal Image Segmentation with Language Prompts (NeurIPS 2025 · 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.
