Research map / Detection and segmentation / Open-set and few-shot detection
CorrCLIPv2: Harnessing Unlabeled Images for Open-Vocabulary Semantic Segmentation
D H Zhang, Fagui Liu, Quan Tang, C. L. Philip Chen
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 Scope Reconstruction (confine patch attention within class-agnostic masks) (ablation) | ClearCLIP baseline · A150 | mIoU: 18.1 -> 24.8 | Improved |
| add Value Reconstruction using VFM features (ablation) | CorrCLIP with SR · A150 | mIoU: 24.8 -> 24.9 | Improved |
| add Feature Refinement (spatial and semantic branches) (ablation) | CorrCLIP with SR+VR · A150 | mIoU: 24.9 -> 26.5 | Improved |
| add Map Correction (majority voting within masks) (ablation) | CorrCLIP with SR+VR+FR · City | mIoU: 45.2 -> 53.7 | Improved |
| add only the semantic branch or only the spatial branch of Feature Refinement (ablation) | CorrCLIP with SR+VR · City | mIoU: 44.2 -> 43.5 (semantic only), 44.2 -> 45.8 (spatial only) | Mixed |
| use larger mask scale factor in Scope Reconstruction (ablation) | CorrCLIP SR · V21 | mIoU: 62.6 -> 69.2 (scale 1 to 4) | Improved |
| add VFM-Indexed Semantic Retrieval (memory bank of unlabeled COCO images) (ablation) | CorrCLIP-B · 10 standard benchmarks (mean) | mIoU: 44.2 -> 47.7 | Improved |
| add Decoupled Multi-CLIP Ensemble (DFN auxiliary bank) (ablation) | CorrCLIP-B +VISR · 10 standard benchmarks (mean) | mIoU: 47.7 -> 49.6 | Improved |
| add VISR and DMCE (ablation) | CorrCLIP-L · 10 standard benchmarks (mean) | mIoU: 47.8 -> 50.0 -> 51.0 | Improved |
| index memory bank with DINOv2 or RADIO instead of SAM (ablation) | CorrCLIPv2 · 10 standard benchmarks (mean) | mIoU: 39.7 -> 49.4 (DINOv2), 39.7 -> 49.6 (C-RADIOv3) | Improved |
| use EntitySeg masks instead of SAM masks (ablation) | CorrCLIPv2 · A150 | mIoU: 30.2 -> 33.2 | Improved |
| store CorrCLIP features in memory bank instead of ClearCLIP or ProxyCLIP features (ablation) | CorrCLIPv2 memory bank · A150 | mIoU: 28.3 -> 33.4 (vs ClearCLIP), 30.5 -> 33.4 (vs ProxyCLIP) | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- CorrCLIP — CorrCLIP: Reconstructing Patch Correlations in CLIP for Open-Vocabulary Semantic Segmentation (ICCV 2025) · better on 2 of 2 shared results
- ReME — ReME: A Data-Centric Framework for Training-Free Open-Vocabulary Segmentation (ICCV 2025) · better on 1 of 1 shared result
- ProxyCLIP — ProxyCLIP: Proxy Attention Improves CLIP for Open-Vocabulary 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
- Emergent Open-Vocabulary Semantic Segmentation from Off-the-shelf Vision-Language Models (CVPR 2024 · similarity 0.90)
- Towards Open-Vocabulary Semantic Segmentation Without Semantic Labels (NeurIPS 2024 · similarity 0.89)
- Training-free Open-Vocabulary Semantic Segmentation via Diverse Prototype Construction and Sub-region Matching (AAAI 2025 · similarity 0.88)
- Parameter-efficient Fine-tuning in Hyperspherical Space for Open-vocabulary Semantic Segmentation (CVPR 2025 · similarity 0.88)
- SegEarth-OV: Towards Training-Free Open-Vocabulary Segmentation for Remote Sensing Images (CVPR 2025 · similarity 0.88)
- OpenWorldSAM: Extending SAM2 for Universal Image Segmentation with Language Prompts (NeurIPS 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.
