Research map / Label-efficient and robust learning / Recognition and long-tail
Weakly-Supervised Deep Convolutional Neural Network Learning for Facial Action Unit Intensity 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 |
|---|---|---|---|
| Use KBSS with peak and valley annotations plus domain-knowledge supervision | AU intensity estimation · FERA 2015 | ICC(3,1): .66 -> .67 | Helped |
| Use KBSS with peak and valley annotations plus domain-knowledge supervision | AU intensity estimation · FERA 2015 | MAE: .68 -> .66 | Helped |
| Use KBSS with peak and valley annotations plus domain-knowledge supervision | AU intensity estimation · DISFA | ICC(3,1): .50 -> .36 | Hurt |
| Use KBSS with peak and valley annotations plus domain-knowledge supervision | AU intensity estimation · DISFA | MAE: .30 -> .33 | Hurt |
| Add temporal intensity ordering (ablation) | KBSS · FERA 2015 | ICC(3,1): .61 -> .67 | Helped |
| Add relative appearance similarity (ablation) | KBSS · FERA 2015 | ICC(3,1): .62 -> .67 | Helped |
| Add facial symmetry (ablation) | KBSS · FERA 2015 | ICC(3,1): .64 -> .67 | Helped |
| Add contrastive appearance difference (ablation) | KBSS · FERA 2015 | ICC(3,1): .65 -> .67 | Helped |
| Use 5-element tuples rather than pairwise intensity relations (ablation) | KBSS · FERA 2015 | ICC(3,1): .62 -> .67 | Helped |
| Use 5-element tuples rather than triplets (ablation) | KBSS · FERA 2015 | ICC(3,1): .62 -> .67 | Helped |
| Add temporal intensity ordering (ablation) | KBSS · DISFA | MAE: .54 -> .33 | Helped |
| Add relative appearance similarity (ablation) | KBSS · DISFA | MAE: .51 -> .33 | Helped |
Closest papers by meaning
- Joint Representation and Estimator Learning for Facial Action Unit Intensity Estimation (CVPR 2019 · similarity 0.93)
- Dynamic Probabilistic Graph Convolution for Facial Action Unit Intensity Estimation (CVPR 2021 · similarity 0.92)
- Facial Action Unit Intensity Estimation via Semantic Correspondence Learning with Dynamic Graph Convolution (AAAI 2020 · similarity 0.92)
- Deep Structured Learning for Facial Action Unit Intensity Estimation (CVPR 2017 · similarity 0.90)
- Trend-Aware Supervision: On Learning Invariance for Semi-supervised Facial Action Unit Intensity Estimation (AAAI 2024 · similarity 0.89)
- AUEditNet: Dual-Branch Facial Action Unit Intensity Manipulation with Implicit Disentanglement (CVPR 2024 · 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.