Research map / Detection and segmentation / Open-set and few-shot detection
MixReorg: Cross-Modal Mixed Patch Reorganization is a Good Mask Learner for Open-World Semantic Segmentation
Kaixin Cai, Pengzhen Ren, Yi Zhu, Hang Xu, Jianzhuang Liu, Changlin Li, Guangrun Wang, Xiaodan Liang
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 contextual mixing to GroupViT-style encoder (ablation) | MixReorg · COCO | mIoU: 18.0 -> 19.3 | Improved |
| add one extra transformer layer to GroupViT (ablation) | GroupViT · COCO | mIoU: 18.4 -> 18.2 | No clear effect |
| add mixed segmentation loss L_seg (ablation) | MixReorg with CM · COCO | mIoU: 19.3 -> 20.5 | Improved |
| add restoration contrastive loss L_re (ablation) | MixReorg with CM · COCO | mIoU: 19.3 -> 20.3 | Improved |
| combine L_seg and L_re on top of CM (ablation) | MixReorg · COCO | mIoU: 20.5 -> 21.3 | Improved |
| increase images mixed M from 16 to 32 (ablation) | MixReorg · COCO | mIoU: 20.5 -> 18.2 | Hurt |
| increase images mixed M up to 16 versus a small M (ablation) | MixReorg · COCO | mIoU: 17.1 -> 20.5 | Improved |
| add progressive mixing modules (P=3 versus P=0) (ablation) | MixReorg · COCO | mIoU: about 7% mIoU | Improved |
| replace GroupViT with MixReorg pretraining on CC12M | GroupViT · PASCAL VOC | mIoU: 41.1 -> 47.9 | Improved |
| replace GroupViT with MixReorg pretraining on CC12M | GroupViT · ADE20K | mIoU: 5.8 -> 8.7 | Improved |
| replace GroupViT with MixReorg pretraining | GroupViT · ImageNet zero-shot | Acc@1: 37.5 -> 38.8 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- GroupViT — GroupViT: Semantic Segmentation Emerges From Text Supervision (CVPR 2022) · better on 9 of 9 shared results
- ViewCo — ViewCo: Discovering Text-Supervised Segmentation Masks via Multi-View Semantic Consistency (ICLR 2023) · 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
- OmniSegmentor: A Flexible Multi-Modal Learning Framework for Semantic Segmentation (NeurIPS 2025 · similarity 0.87)
- Linguistic-Aware Patch Slimming Framework for Fine-grained Cross-Modal Alignment (CVPR 2024 · similarity 0.87)
- Leveraging Depth and Language for Open-Vocabulary Domain-Generalized Semantic Segmentation (NeurIPS 2025 · similarity 0.87)
- Seg4Diff: Unveiling Open-Vocabulary Semantic Segmentation in Text-to-Image Diffusion Transformers (NeurIPS 2025 · similarity 0.87)
- Image-Text Co-Decomposition for Text-Supervised Semantic Segmentation (CVPR 2024 · similarity 0.87)
- LPOSS: Label Propagation Over Patches and Pixels for Open-vocabulary Semantic Segmentation (CVPR 2025 · similarity 0.86)
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.
