Research map / Detection and segmentation / Weakly supervised detection and segmentation
Learning from Weak and Noisy Labels for Semantic Segmentation
Zhiwu Lu, Zhenyong Fu, Tao Xiang, Peng Han, Liwei Wang, Xin Gao
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 |
|---|---|---|---|
| Increase noisily tagged images from 0% to 100% | WSSL · MSRC | accuracy (%): 76 -> 55 | Hurt |
| Increase noisily tagged images from 0% to 100% | WSSL · VOC2007 | accuracy (%): 49 -> 33 | Hurt |
| Use WSSL instead of best sparse baseline (100% noise) | Baseline 3 · MSRC | accuracy (%): 47 -> 55 | Improved |
| Use WSSL with two sparsity regularizers instead of one | Baseline 2 · MSRC | accuracy (%): 70 -> 76 | Improved |
| Use WSSL instead of best baseline (clean tags) | Baseline 2 · VOC2007 | accuracy (%): 34 -> 49 | Improved |
| Use WSSL instead of best baseline (100% noise) | Baseline 2 · VOC2007 | accuracy (%): 20 -> 33 | Improved |
| Use WSSL with noisy tags vs clean-tag weakly supervised method | weakly supervised method B · VOC2007 | accuracy (%): 24 -> 33 | Improved |
| Use WSSL with 0% noise vs fully supervised on VOC2007 | fully supervised method · VOC2007 | accuracy (%): 31 -> 49 | Improved |
| Use WSSL with 0% noise vs fully supervised on MSRC | fully supervised method B · MSRC | accuracy (%): 76 -> 76 | No clear effect |
| Use WSSL instead of weakly supervised method on MSRC with 75% noise | weakly supervised method A · MSRC | accuracy (%): 67 -> 60 | No clear effect |
| Use WSSL instead of slower baselines | Baseline 2 · VOC2007 | running time (s): 145 -> 40 | Improved |
Closest papers by meaning
- Universal Weakly Supervised Segmentation by Pixel-to-Segment Contrastive Learning (ICLR 2021 · similarity 0.91)
- Weakly-Supervised Semantic Segmentation by Iterative Affinity Learning (IJCV 2020 · similarity 0.89)
- Combining Bottom-Up, Top-Down, and Smoothness Cues for Weakly Supervised Image Segmentation (CVPR 2017 · similarity 0.89)
- Saliency Guided Dictionary Learning for Weakly-Supervised Image Parsing (CVPR 2016 · similarity 0.89)
- Simple Does It: Weakly Supervised Instance and Semantic Segmentation (CVPR 2017 · similarity 0.89)
- Label-efficient Segmentation via Affinity Propagation (NeurIPS 2023 · similarity 0.89)
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.
