Research map / Label-efficient and robust learning / Noisy labels
See Through the Noise: Improving Domain Generalization in Gaze Estimation
Yanming Peng, Shijing Wang, Yaping Huang, Yi Tian
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 SeeTN (noise-robust semantic manifold regularization) | ResNet-18 · ETH-XGaze to MPIIFaceGaze | mean angular error: 8.07 -> 6.58 | Improved |
| add SeeTN | ResNet-18 · Within Gaze360 | mean angular error: 11.14 -> 10.73 | Improved |
| add L_align^C (clean-sample affinity alignment) (ablation) | SeeTN, ResNet-18 · ETH-XGaze to MPIIFaceGaze | mean angular error: 8.07 -> 7.14 | Improved |
| add L_align^N (noisy-sample affinity regularization) (ablation) | SeeTN, ResNet-18 · ETH-XGaze to MPIIFaceGaze | mean angular error: 7.14 -> 6.58 | Improved |
| add L_align^N (ablation) | SeeTN, ResNet-18 · Gaze360 to EyeDiap | mean angular error: 8.11 -> 7.57 | Improved |
| add prototype-based semantic manifold (ablation) | SeeTN, ResNet-18 · ETH-XGaze to MPIIFaceGaze | mean angular error: 7.88 -> 6.58 | Improved |
| use cosine similarity instead of L2 distance (ablation) | SeeTN, ResNet-18 · ETH-XGaze to MPIIFaceGaze | mean angular error: 7.01 -> 6.58 | Improved |
| use affinity-mismatch indicator eta instead of L1 loss for noisy sample detection (ablation) | SeeTN, ResNet-18 · ETH-XGaze to MPIIFaceGaze | mean angular error: 7.73 -> 6.58 | Improved |
| increase number of prototypes K from 12 to 16 (ablation) | SeeTN, ResNet-18 · ETH-XGaze to MPIIFaceGaze | mean angular error: 6.58 -> 7.59 | Hurt |
| decrease prototypes K from 12 to 8 (ablation) | SeeTN, ResNet-18 · ETH-XGaze to MPIIFaceGaze | mean angular error: 6.58 -> 7.37 | Hurt |
| increase noisy ratio t from 5% to 20% (ablation) | SeeTN, ResNet-18 · ETH-XGaze to MPIIFaceGaze | mean angular error: 6.58 -> 7.63 | Hurt |
| raise loss weight lambda from 0.1 to 1 (ablation) | SeeTN, ResNet-18 · ETH-XGaze to MPIIFaceGaze | mean angular error: 6.58 -> 7.02 | Hurt |
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.
- DivideMix — DivideMix: Learning with Noisy Labels as Semi-supervised Learning (ICLR 2020) · better on 2 of 2 shared results
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
- Gaze Label Alignment: Alleviating Domain Shift for Gaze Estimation (AAAI 2025 · similarity 0.90)
- Suppressing Uncertainty in Gaze Estimation (AAAI 2024 · similarity 0.89)
- Gaze from Origin: Learning for Generalized Gaze Estimation by Embedding the Gaze Frontalization Process (AAAI 2024 · similarity 0.89)
- A Generalized Label Shift Perspective for Cross-Domain Gaze Estimation (NeurIPS 2025 · similarity 0.89)
- CLIP-Gaze: Towards General Gaze Estimation via Visual-Linguistic Model (AAAI 2024 · similarity 0.88)
- OmniGaze: Reward-inspired Generalizable Gaze Estimation in the Wild (NeurIPS 2025 · similarity 0.88)
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.
