Research map / Detection and segmentation / CNN features
MCIBI++: Soft Mining Contextual Information Beyond Image for Semantic Segmentation
Zhenchao Jin, Dongdong Yu, Zehuan Yuan, Lequan Yu
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 dataset-level context aggregation (MCIBI++) to FCN (ablation) | FCN ResNet-50 · ADE20K val | mIoU: 36.96 -> 43.39 | Improved |
| fuse by concatenation instead of element-wise add (ablation) | FCN+MCIBI++ · ADE20K val | mIoU: 41.15 -> 43.39 | Improved |
| update the memory module during training (ablation) | FCN+MCIBI (Memory-R) · ADE20K val | mIoU: 21.78 -> 42.18 | Improved |
| store distribution information (mean, variance) instead of raw representations (ablation) | FCN+MCIBI++ · ADE20K val | mIoU: 42.18 -> 43.39 | Improved |
| add coarse-to-fine iterative inference strategy (ablation) | FCN+MCIBI++ · ADE20K val | mIoU: 43.39 -> 43.74 | Improved |
| use fixed momentum 0.1 instead of polynomial annealing (ablation) | FCN+MCIBI++ · ADE20K val | mIoU: 42.45 -> 43.39 | Improved |
| set loss weight alpha to 0.4 instead of 1.0 (ablation) | FCN+MCIBI++ · ADE20K val | mIoU: 42.22 -> 43.39 | Improved |
| increase inference stages from 2 to 4 (ablation) | FCN+MCIBI++ · ADE20K val | mIoU: 43.74 -> 43.91 (49.33 ms -> 74.72 ms) | No clear effect |
| initialize memory by cosine-similarity clustering instead of random pixel (ablation) | FCN+MCIBI++ · ADE20K val | mIoU: 43.39 -> 43.43 | No clear effect |
| add DCA to ASPP (DeepLabV3) as complementary context (ablation) | DeepLabV3 ResNet-50 · ADE20K val | mIoU: 43.19 -> 44.85 | Improved |
| add MCIBI++ to PSPNet | PSPNet ResNet-50 · ADE20K val | mIoU: 42.64 -> 43.89 | Improved |
| add MCIBI++ to UperNet trained on LIP, tested on CIHP | UperNet · CIHP val (trained on LIP) | mIoU: 26.84 -> 27.74 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ISNet — ISNet: Integrate Image-Level and Semantic-Level Context for Semantic Segmentation (ICCV 2021) · 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
- Mining Contextual Information Beyond Image for Semantic Segmentation (ICCV 2021 · similarity 0.91)
- ISNet: Integrate Image-Level and Semantic-Level Context for Semantic Segmentation (ICCV 2021 · similarity 0.88)
- Coarse-To-Fine Feature Mining for Video Semantic Segmentation (CVPR 2022 · similarity 0.87)
- Multi-Scale Context Intertwining for Semantic Segmentation (ECCV 2018 · similarity 0.86)
- Mining Relations among Cross-Frame Affinities for Video Semantic Segmentation (ECCV 2022 · similarity 0.86)
- Integrating Low-Level Visual Cues for Enhanced Unsupervised Semantic Segmentation (AAAI 2025 · similarity 0.85)
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.
