Research map / Detection and segmentation / Weakly supervised detection and segmentation
SegSort: Segmentation by Discriminative Sorting of Segments
Jyh-Jing Hwang, Stella X. Yu, Jianbo Shi, Maxwell D. Collins, Tien-Ju Yang, Xiao Zhang, Liang-Chieh 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 |
|---|---|---|---|
| replace pixel-wise softmax with SegSort metric learning | DeepLabv3+ / MNV2 · PASCAL VOC 2012 val | mIoU: 72.51 -> 74.94 | Improved |
| replace pixel-wise softmax with SegSort metric learning | PSPNet / ResNet-101 · PASCAL VOC 2012 val | mIoU: 80.12 -> 81.77 | Improved |
| replace pixel-wise softmax with SegSort metric learning | DeepLabv3+ / MNV2 · PASCAL VOC 2012 val | boundary f-measure: 50.90 -> 58.83 | Improved |
| replace pixel-wise softmax with SegSort | DeepLabv3+ / ResNet-101 · PASCAL VOC 2012 val | mIoU: 78.24 -> 78.85 | Improved |
| replace pixel-wise softmax with SegSort | PSPNet / ResNet-101 · Cityscapes val | mIoU: 76.72 -> 78.15 | Improved |
| replace ASM with SegSort | PSPNet / ResNet-101 · PASCAL VOC 2012 val | mIoU: 81.43 -> 81.77 | Improved |
| train unsupervised SegSort instead of clustering pretrained features (ablation) | HED-owt-ucm baseline · PASCAL VOC 2012 val | mIoU: 49.50 -> 55.86 | Improved |
| add unsupervised pretraining before supervised training on train (ablation) | SegSort · PASCAL VOC 2012 val | mIoU: 71.86 -> 72.47 | Improved |
| add unsupervised pretraining before supervised training on train (ablation) | SegSort · PASCAL VOC 2012 val | boundary f-measure: 55.70 -> 56.52 | Improved |
| increase k-means iterations from 4 to 10 at inference (ablation) | SegSort (11-NN) · Cityscapes val | mIoU: 78.50 -> 78.69 | Improved |
| increase nearest neighbors from 1 to 11 at inference (ablation) | SegSort (4 iter) · Cityscapes val | mIoU: 78.05 -> 78.50 | Improved |
| vary the number of nearest neighbors from 1 to 31 (ablation) | SegSort PSPNet · PASCAL VOC 2012 val | mIoU: 0.4% | No clear effect |
Later papers that compared with this method
2 later accepted papers report SegSort as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Unsupervised Hierarchical Semantic Segmentation With Multiview Cosegmentation and Clustering Transformers (CVPR 2022 · better on 10 of 10 shared results)
- Self-Supervised Visual Representation Learning from Hierarchical Grouping (NeurIPS 2020 · better on 2 of 2 shared results)
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
- Seg-ReSearch: Segmentation with Interleaved Reasoning and External Search (ICML 2026 · similarity 0.89)
- SegGPT: Towards Segmenting Everything in Context (ICCV 2023 · similarity 0.89)
- Improving Semi-Supervised Semantic Segmentation with Sliced-Wasserstein Feature Alignment and Uniformity (CVPR 2025 · similarity 0.88)
- OMG-Seg: Is One Model Good Enough For All Segmentation? (CVPR 2024 · similarity 0.88)
- Integrating Low-Level Visual Cues for Enhanced Unsupervised Semantic Segmentation (AAAI 2025 · similarity 0.87)
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · 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.
