Research map / Detection and segmentation / CNN features
Exploring Cross-Image Pixel Contrast for Semantic Segmentation
Wenguan Wang, Tianfei Zhou, Fisher Yu, Jifeng Dai, Ender Konukoglu, Luc Van Gool
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 inter-image (cross-image) pixel contrast instead of no contrast (ablation) | HRNetV2-W48 segmentation · Cityscapes val | mIoU: 78.1 -> 81.0 | Improved |
| restrict contrast to pixels within the same image (intra-image) rather than across images (ablation) | HRNetV2-W48 segmentation · Cityscapes val | mIoU: 81.0 -> 78.9 | Hurt |
| add a memory bank to mini-batch-only contrast (ablation) | pixel contrast with HRNetV2-W48 · Cityscapes val | mIoU: 79.8 -> 80.5 (pixel memory), 79.8 -> 80.2 (region memory) | Improved |
| add region memory to pixel memory (ablation) | pixel contrast with HRNetV2-W48 · Cityscapes val | mIoU: 80.5 -> 81.0 | Improved |
| use semi-hard rather than random positive/negative sampling (ablation) | pixel contrast with random anchors · Cityscapes val | mIoU: 79.3 -> 80.1 | Improved |
| use hardest rather than random positive/negative sampling (ablation) | pixel contrast with random anchors · Cityscapes val | mIoU: 79.3 -> 79.4 | No clear effect |
| use semi-hard rather than hardest sampling with hard anchors (ablation) | pixel contrast with seg.-aware hard anchors · Cityscapes val | mIoU: 80.5 -> 81.0 | Improved |
| add segmentation-aware hard anchor sampling to semi-hard sampling (ablation) | pixel contrast · Cityscapes val | mIoU: 80.1 -> 81.0 | Improved |
| add contrastive loss to cross-entropy, versus adding Lovasz loss (ablation) | HRNetV2-W48 segmentation · Cityscapes val | mIoU: 80.3 -> 81.0 | Improved |
| add contrastive loss on top of AAF loss (ablation) | HRNetV2-W48 + CE + AAF · Cityscapes val | mIoU: 81.0 -> 81.0 | No clear effect |
| add contrastive loss on top of RMI loss (ablation) | HRNetV2-W48 + CE + RMI · Cityscapes val | mIoU: 79.8 -> 81.3 | Improved |
| add contrastive loss on top of Lovasz loss (ablation) | HRNetV2-W48 + CE + Lovasz · Cityscapes val | mIoU: 80.3 -> 81.5 | Improved |
Closest papers by meaning
- Cross-Image Pixel Contrasting for Semantic Segmentation (TPAMI 2024 · similarity 0.99)
- Looking Beyond Single Images for Contrastive Semantic Segmentation Learning (NeurIPS 2021 · similarity 0.91)
- Region-Aware Contrastive Learning for Semantic Segmentation (ICCV 2021 · similarity 0.91)
- Contrastive Learning for Label Efficient Semantic Segmentation (ICCV 2021 · similarity 0.91)
- Pixel Contrastive-Consistent Semi-Supervised Semantic Segmentation (ICCV 2021 · similarity 0.90)
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · 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.
