Research map / Label-efficient and robust learning / Noisy labels
SILT: Shadow-Aware Iterative Label Tuning for Learning to Detect Shadows from Noisy Labels
Han Yang, Tianyu Wang, Xiaowei Hu, Chi-Wing Fu
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 global-local fusion to label tuning (ablation) | SILT self-training · SBU (relabeled test) | BER: 5.82 -> 4.67 | Improved |
| add shadow-aware filtering to self-training (ablation) | ST · SBU (relabeled test) | BER: 5.82 -> 5.44 | Improved |
| add shadow counterfeiting augmentation (ablation) | ST+LT+SDA · SBU (relabeled test) | BER: 4.46 -> 4.37 | Improved |
| naive self-training vs plain training on noisy labels (ablation) | U-Net PVT v2-B5 · SBU (relabeled test) | BER: 6.62 -> 5.82 | Improved |
| add 89 non-shadow dark-object images (ablation) | SILT · SBU | BER: 4.37 -> 4.19 | Improved |
| add 89 non-shadow dark-object images (ablation) | SILT · UCF and ISTD | BER: UCF 7.23 -> 7.60; ISTD 1.16 -> 1.18 | Hurt |
| train existing SOTA detectors on SILT-refined labels | BDRAR, DSC, FSD, FDRnet, SDCM · SBU (relabeled test) | BER: SDCM 5.71 -> 4.87; FSD 10.87 -> 6.13 | Improved |
| use more self-training rounds than seven | SILT · SBU | mask quality | Hurt |
| set R_corr to 0.95 rather than 0.91 or 0.99 (ablation) | SILT · SBU | BER: 4.59 / 4.52 -> 4.19 | Improved |
| train on small well-labeled subset vs original noisy SBU | SDCM · SBU split | BER: 5.30 -> 4.79 | Improved |
Closest papers by meaning
- Large Scale Shadow Annotation and Detection Using Lazy Annotation and Stacked CNNs (TPAMI 2019 · similarity 0.90)
- Noisy Label Recovery for Shadow Detection in Unfamiliar Domains (CVPR 2016 · similarity 0.89)
- A Multi-Task Mean Teacher for Semi-Supervised Shadow Detection (CVPR 2020 · similarity 0.88)
- Direction-Aware Spatial Context Features for Shadow Detection and Removal (TPAMI 2019 · similarity 0.87)
- DC-ShadowNet: Single-Image Hard and Soft Shadow Removal Using Unsupervised Domain-Classifier Guided Network (ICCV 2021 · similarity 0.87)
- Bidirectional Feature Pyramid Network with Recurrent Attention Residual Modules for Shadow Detection (ECCV 2018 · 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.
