Research map / Detection and segmentation / CNN features
Mining Contextual Information Beyond Image for Semantic Segmentation
Zhenchao Jin, Tao Gong, Dongdong Yu, Qi Chu, Jian Wang, Changhu Wang, Jie Shao
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 MCIBI with concatenation fusion (ablation) | FCN ResNet-50 · ADE20K val | mIoU: 36.96 -> 42.18 | Improved |
| use concatenation instead of element-wise add for fusion (ablation) | FCN+MCIBI · ADE20K val | mIoU: 40.99 -> 42.18 | Improved |
| add cosine similarity to the memory update (ablation) | FCN+MCIBI · ADE20K val | mIoU: 41.77 -> 42.18 | Improved |
| add poly momentum annealing to the memory update (ablation) | FCN+MCIBI · ADE20K val | mIoU: 41.95 -> 42.18 | Improved |
| initialize memory by clustering instead of random pixel (ablation) | FCN+MCIBI · ADE20K val | mIoU: 42.18 -> 42.20 | No clear effect |
| add representation consistent learning (ablation) | FCN+MCIBI · ADE20K val | mIoU: 42.18 -> 42.84 | Improved |
| add clustering init on top of RCL (ablation) | FCN+MCIBI · ADE20K val | mIoU: 42.84 -> 42.75 | No clear effect |
| add MCIBI to FCN | FCN ResNet-50 · ADE20K val | mIoU: 36.96 -> 42.84 | Improved |
| add MCIBI to stronger frameworks | PSPNet, OCRNet, DeepLabV3 · ADE20K val | mIoU: 42.64 -> 43.77 | Improved |
| add MCIBI to DeepLabV3 | DeepLabV3 ResNet-50 · COCO-Stuff | mIoU: 41.86 -> 42.08 | Improved |
| add MCIBI to FCN, train on LIP test on CIHP | FCN · CIHP val | mIoU: 27.20 -> 27.57 | Improved |
| add MCIBI to DeepLabV3, train on LIP test on CIHP | DeepLabV3 · CIHP val | mIoU: 27.02 -> 27.89 | Improved |
Closest papers by meaning
- Object-Contextual Representations for Semantic Segmentation (ECCV 2020 · similarity 0.92)
- ISNet: Integrate Image-Level and Semantic-Level Context for Semantic Segmentation (ICCV 2021 · similarity 0.91)
- MCIBI++: Soft Mining Contextual Information Beyond Image for Semantic Segmentation (TPAMI 2022 · similarity 0.91)
- Learning Context-Aware Classifier for Semantic Segmentation (AAAI 2023 · similarity 0.88)
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · similarity 0.86)
- Contextrast: Contextual Contrastive Learning for Semantic Segmentation (CVPR 2024 · 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.
