Research map / Label-efficient and robust learning / Recognition and long-tail
Order Regularization on Ordinal Loss for Head Pose, Age and Gaze Estimation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 order regularization to ordinal loss | Ordinal loss · BIWI (300W-LP/BIWI protocol), MORPH, MPII, and GazeCapture | Error: 3% to 8%; 4.7%; 4.7%; 3.9% | Helped |
| Add order regularization to ordinal loss | Ordinal loss · BIWI (300W-LP/BIWI protocol), MORPH, MPII, and GazeCapture | Error relative to state of the art: 14.4%, 2.2% and 6.5% | Helped |
| Allow limited weight-norm variation with a variance penalty (ablation) | Order-regularized ordinal loss · MPII | Err (°): 4.74; 4.49 | Helped |
| Allow limited weight-norm variation with a variance penalty (ablation) | Order-regularized ordinal loss · GazeCapture | Err (cm.): 1.80; 1.73 | Helped |
| Parameterize bias using differences Δb instead of directly storing b (ablation) | Order-regularized ordinal loss · MPII | Err (°): 4.57; 4.57 | No clear effect |
| Parameterize bias using differences Δb instead of directly storing b (ablation) | Order-regularized ordinal loss · GazeCapture | Err (cm.): 1.81; 1.76 | Helped |
| Remove the upper bound mmax from the differential-bias constraint (ablation) | Order-regularized ordinal loss · MPII | Err (°): 4.49; 4.49 | No clear effect |
| Remove the upper bound mmax from the differential-bias constraint (ablation) | Order-regularized ordinal loss · GazeCapture | Err (cm.): 1.73; 1.73 | No clear effect |
| Use only the bias-order constraint (ablation) | Order-regularized ordinal loss · MPII | Err (°): 4.49; 4.66 | Hurt |
| Use only the bias-order constraint (ablation) | Order-regularized ordinal loss · GazeCapture | Err (cm.): 1.73; 1.82 | Hurt |
| Use only the weight-order constraint (ablation) | Order-regularized ordinal loss · MPII | Err (°): 4.49; 4.60 | Hurt |
| Use only the weight-order constraint (ablation) | Order-regularized ordinal loss · GazeCapture | Err (cm.): 1.73; 1.75 | Hurt |
Closest papers by meaning
- Unimodal-Concentrated Loss: Fully Adaptive Label Distribution Learning for Ordinal Regression (CVPR 2022 · similarity 0.86)
- Ordinal Regression With Multiple Output CNN for Age Estimation (CVPR 2016 · similarity 0.86)
- Improving Deep Regression with Ordinal Entropy (ICLR 2023 · similarity 0.85)
- Neural Collapse in Cumulative Link Models for Ordinal Regression: An Analysis with Unconstrained Feature Model (NeurIPS 2025 · similarity 0.84)
- Improving Deep Regression with Tightness (ICLR 2025 · similarity 0.82)
- SLACE: A Monotone and Balance-Sensitive Loss Function for Ordinal Regression (AAAI 2025 · similarity 0.82)
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.