Research map / Detection and segmentation / CNN features
Looking Beyond the Window: Global-Local Aligned CLIP for Training-free Open-Vocabulary Semantic Segmentation
ByeongCheol Lee, Hyun Seok Seong, Sangeek Hyun, Gilhan Park, WonJun Moon, Jae-Pil Heo
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 key-value extension + dynamic normalization to inner-window baseline (ablation) | ProxyCLIP-style baseline · 8 OVSS datasets avg | mIoU: 30.8 -> 43.1 | Improved |
| add proxy anchor attention (ablation) | GLA-CLIP · 8 OVSS datasets avg | mIoU: 43.1 -> 44.0 | Improved |
| add key-value extension (ablation) | GLA-CLIP with proxy · 8 OVSS datasets avg | mIoU: 43.0 -> 44.0 | Improved |
| dataset-specific tuned normalization instead of dynamic normalization (ablation) | GLA-CLIP · 8 OVSS datasets avg | mIoU: 44.0 vs 44.3 | No clear effect |
| add GLA to ProxyCLIP | ProxyCLIP · ProxyCLIP setting avg | mIoU: 42.3 -> 42.9 | Improved |
| add GLA to SCLIP | SCLIP · SCLIP setting avg | mIoU: 38.2 -> 39.8 | Improved |
| add GLA to ClearCLIP | ClearCLIP · ClearCLIP setting avg | mIoU: 38.1 -> 39.3 | Improved |
| add GLA to ClearCLIP | ClearCLIP · VOC20 | mIoU: 80.9 -> 79.7 | Hurt |
| GLA-CLIP vs CLIP-DINOiser on ADE20K | GLA-CLIP · ADE20K (CLIP-DINOiser setting) | mIoU: 20.0 -> 19.1 | Hurt |
| add GLA to ProxyCLIP with DINOv3 VFM | ProxyCLIP · VOC21/PC59/C-Stf avg | mIoU: 42.6 -> 42.9 | Improved |
| use GLA with weaker CLIP ViT-B/32 backbone | ProxyCLIP/CASS · VOC21/PC59/C-Stf avg | mIoU: 38.7 -> 41.4 | Improved |
| raise proxy repetition steps T (ablation) | proxy anchor · avg mIoU | mIoU | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- ProxyCLIP — ProxyCLIP: Proxy Attention Improves CLIP for Open-Vocabulary Segmentation (ECCV 2024) · better on 5 of 5 shared results
- DIH-CLIP — DIH-CLIP: Unleashing the Diversity of Multi-Head Self-Attention for Training-Free Open-Vocabulary Semantic Segmentation (ICCV 2025) · better on 1 of 1 shared result
- ClearCLIP — ClearCLIP: Decomposing CLIP Representations for Dense Vision-Language Inference (ECCV 2024) · better on 1 of 1 shared result
- CLIP-DINOiser — CLIP-DINOiser: Teaching CLIP a few DINO tricks for open-vocabulary semantic segmentation (ECCV 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
- Enhancing Local Cognition of CLIP for Training-Free Open Vocabulary Semantic Segmentation (TPAMI 2026 · similarity 0.92)
- Unveiling the Knowledge of CLIP for Training-Free Open-Vocabulary Semantic Segmentation (AAAI 2025 · similarity 0.91)
- Open-Vocabulary Segmentation with Semantic-Assisted Calibration (CVPR 2024 · similarity 0.90)
- Understanding Fine-tuning CLIP for Open-vocabulary Semantic Segmentation in Hyperbolic Space (CVPR 2025 · similarity 0.90)
- Emergent Open-Vocabulary Semantic Segmentation from Off-the-shelf Vision-Language Models (CVPR 2024 · similarity 0.89)
- Dual Semantic Guidance for Open Vocabulary Semantic Segmentation (CVPR 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.
