Research map / Detection and segmentation / Weakly supervised detection and segmentation
Scribble-Supervised Semantic Segmentation Inference
Jingshan Xu, Chuanwei Zhou, Zhen Cui, Chunyan Xu, Yuge Huang, Pengcheng Shen, Shaoxin Li, Jian Yang
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 CPP module (ablation) | DeepLabV3+ baseline · PASCAL VOC 2012 val (single-scale) | mIoU (%): 66.6 -> 70.1 | Improved |
| add SLD module (ablation) | DeepLabV3+ baseline · PASCAL VOC 2012 val (single-scale) | mIoU (%): 66.6 -> 69.4 | Improved |
| add SLD on top of CPP (ablation) | CPP model · PASCAL VOC 2012 val (single-scale) | mIoU (%): 70.1 -> 72.9 | Improved |
| add multi-scale testing (ablation) | PSI (CPP+SLD) · PASCAL VOC 2012 val | mIoU (%): 72.9 -> 74.9 | Improved |
| use CPP instead of UP for pattern aggregation (ablation) | PSI · PASCAL VOC 2012 val | mIoU (%): 72.6 -> 74.9 | Improved |
| use CPP instead of SF for pattern aggregation (ablation) | PSI · PASCAL VOC 2012 val | mIoU (%): 73.3 -> 74.9 | Improved |
| use higher-level destination layer X(4) instead of X(1) (ablation) | CPP · PASCAL VOC 2012 val | mIoU (%): 65.33 -> 70.09 | Improved |
| increase SLD steps from 1 to 3 (ablation) | PSI · PASCAL VOC 2012 val | mIoU (%): 72.50 -> 72.88 | Improved |
| increase SLD steps from 4 to 6 (ablation) | PSI · PASCAL VOC 2012 val | mIoU (%): 72.88 -> 72.74 | Hurt |
| use PSI instead of plain DeepLabV3+ with scribbles | DeepLabV3+ · PASCAL Context | mIoU (%): 37.1 -> 43.1 | Improved |
| use PSI instead of BPG (no CRF) | PSI vs BPG · PASCAL VOC 2012 val | mIoU (%): 73.2 -> 74.9 | Improved |
Closest papers by meaning
- Progressive Bayesian Inference for Scribble-Supervised Semantic Segmentation (AAAI 2023 · similarity 0.92)
- Scribbles for All: Benchmarking Scribble Supervised Segmentation Across Datasets (NeurIPS 2024 · similarity 0.91)
- Scribble Hides Class: Promoting Scribble-Based Weakly-Supervised Semantic Segmentation with Its Class Label (AAAI 2024 · similarity 0.91)
- Scribble-Supervised Semantic Segmentation with Prototype-based Feature Augmentation (ICML 2024 · similarity 0.90)
- Diffusion-Driven Two-Stage Active Learning for Low-Budget Semantic Segmentation (NeurIPS 2025 · similarity 0.88)
- Active Label Correction for Semantic Segmentation with Foundation Models (ICML 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.
