Research map / Detection and segmentation / Open-set and few-shot detection
Exposing Semantic Segmentation Failures via Maximum Discrepancy Competition
Jiebin Yan, Yu Zhong, Yuming Fang, Zhangyang Wang, Kede Ma
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 category constraint to MAD selection (ablation) | MAD image selection · PASCAL VOC MAD set | image diversity | Improved |
| add scale constraint on top of category constraint (ablation) | MAD image selection · PASCAL VOC MAD set | image diversity | Improved |
| switch concordance measure from mIoU to FWIoU (ablation) | MAD ranking · PASCAL VOC | SRCC of aggressiveness ranking: 0.891 | No clear effect |
| switch concordance measure from mIoU to MPA (ablation) | MAD ranking · PASCAL VOC | SRCC of aggressiveness ranking: 0.878 | No clear effect |
| switch concordance measure from mIoU to MPA (ablation) | MAD ranking · PASCAL VOC | SRCC of resistance ranking: 0.830 | No clear effect |
| evaluate on MAD set instead of PASCAL VOC test | ten segmentation models · PASCAL VOC | mIoU: all above 0.6 -> 84.7% below 0.6 | Hurt |
| detection auxiliary task (BlitzNet) vs attention modules (EMANet, PSANet) | segmentation models · PASCAL VOC MAD | aggressiveness rank: BlitzNet mIoU rank 7 -> A rank 3; EMANet 2 -> 5 | Mixed |
| multi-resolution fusion and chained residual pooling (RefineNet) | segmentation model · PASCAL VOC MAD | aggressiveness rank: mIoU rank 6 -> A rank 2 | Improved |
| more training data (Cityscapes coarse + Mapillary) for HANet | HANet · WildDash MAD | rank: A rank 1, R rank 1 | Improved |
| graph convolution modules in DGCNet | DGCNet · WildDash MAD | rank: mIoU rank 2 -> A rank 4 | Hurt |
Closest papers by meaning
- Reliability in Semantic Segmentation: Are We on the Right Track? (CVPR 2023 · similarity 0.89)
- Direct Segmentation without Logits Optimization for Training-Free Open-Vocabulary Semantic Segmentation (CVPR 2026 · similarity 0.89)
- Exploring Cross-Image Pixel Contrast for Semantic Segmentation (ICCV 2021 · similarity 0.89)
- Emergent Open-Vocabulary Semantic Segmentation from Off-the-shelf Vision-Language Models (CVPR 2024 · similarity 0.89)
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · similarity 0.88)
- Test-Time Adaptation of Vision-Language Models for Open-Vocabulary Semantic Segmentation (NeurIPS 2025 · similarity 0.88)
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.
