Research map / Detection and segmentation / Weakly supervised detection and segmentation
FeaBoost: Joint Feature and Label Refinement for Semantic Segmentation
Yulei Niu, Zhiwu Lu, Songfang Huang, Xin Gao, Ji-Rong Wen
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 one-vs-all SVM trained on refined labels | FeaBoost · MSRC | accuracy (%): 77 -> 78 | Improved |
| add one-vs-all SVM trained on refined labels | FeaBoost · LabelMe | accuracy (%): 39 -> 42 | Improved |
| weakly supervised FeaBoost versus Niu et al. 2015 low-rank factorization | FeaBoost · MSRC | accuracy (%): 71 -> 77 | Improved |
| FeaBoost+SVM versus fully supervised Russell et al. 2009 | FeaBoost+SVM · MSRC | accuracy (%): 75 -> 78 | Improved |
| FeaBoost+SVM versus Li et al. 2015 and Xu et al. 2015 | FeaBoost+SVM · LabelMe | accuracy (%): 41 -> 42 | Improved |
| add 1 random noisy label to 100% of images | FeaBoost · LabelMe | accuracy (%): 47 -> 35 | Hurt |
| add 2 random noisy labels to 100% of images | FeaBoost · LabelMe | accuracy (%): 47 -> 24 | Hurt |
| add 3 random noisy labels to 100% of images | FeaBoost · LabelMe | accuracy (%): 47 -> 20 | Hurt |
| FeaBoost versus Li et al. 2015 under noise r=1, p=100 | FeaBoost · LabelMe | accuracy (%): 30 -> 35 | Improved |
| FeaBoost versus Niu et al. 2015 under noise r=3, p=100 | FeaBoost · LabelMe | accuracy (%): 18 -> 20 | Improved |
| FeaBoost with clean labels versus Li et al. 2015 | FeaBoost · LabelMe | accuracy (%): 41 -> 47 | Improved |
| FeaBoost with clean labels versus Niu et al. 2015 | FeaBoost · LabelMe | accuracy (%): 33 -> 47 | Improved |
Closest papers by meaning
- Incremental Learning in Semantic Segmentation From Image Labels (CVPR 2022 · similarity 0.88)
- Active Learning for Semantic Segmentation with Multi-class Label Query (NeurIPS 2023 · similarity 0.88)
- Single-Stage Semantic Segmentation From Image Labels (CVPR 2020 · similarity 0.88)
- Language-driven Semantic Segmentation (ICLR 2022 · similarity 0.87)
- Label-Efficient Few-Shot Semantic Segmentation with Unsupervised Meta-Training (AAAI 2024 · similarity 0.87)
- Active Label Correction for Semantic Segmentation with Foundation Models (ICML 2024 · similarity 0.87)
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.
