Research map / Graph learning / GNN architectures
Hybrid Message Passing With Performance-Driven Structures for Facial Action Unit Detection
Check 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 HMP-PS instead of the strongest reported baseline | BP4D AU detection · BP4D | Avg F1 score (%): 62.9 -> 63.4 | Helped |
| Use HMP-PS instead of the strongest reported baseline | DISFA AU detection · DISFA | Avg F1 score (%): 58.7 -> 61.0 | Helped |
| Add a graph structure to the model (ablation) | HMP-PS on BP4D · BP4D | Avg F1 score (%): 59.6 -> 63.4 | Helped |
| Replace the fully connected graph with performance-driven sampled structures (ablation) | HMP-PS on BP4D · BP4D | Avg F1 score (%): 58.6 -> 63.4 | Helped |
| Replace the FACS prior graph with performance-driven sampled structures (ablation) | HMP-PS on BP4D · BP4D | Avg F1 score (%): 61.3 -> 63.4 | Helped |
| Replace the gradient-based graph with performance-driven sampled structures (ablation) | HMP-PS on BP4D · BP4D | Avg F1 score (%): 61.7 -> 63.4 | Helped |
| Add dynamic weights τ to combine graph structures (ablation) | HMP-PS · BP4D | Avg F1 score (%): 62.5 -> 63.4 | Helped |
| Add dynamic weights τ to combine graph structures (ablation) | HMP-PS · DISFA | Avg F1 score (%): 60.4 -> 61.0 | Helped |
| Add dynamic weights p to combine message types (ablation) | HMP-PS · BP4D | Avg F1 score (%): 61.0 -> 63.4 | Helped |
| Add dynamic weights p to combine message types (ablation) | HMP-PS · DISFA | Avg F1 score (%): 59.7 -> 61.0 | Helped |
| Add summation, differential, and concatenation messages instead of using summation alone (ablation) | HMP-PS · BP4D | Avg F1 score (%): 61.6 -> 63.4 | Helped |
| Add summation, differential, and concatenation messages instead of using summation alone (ablation) | HMP-PS · DISFA | Avg F1 score (%): 58.4 -> 61.0 | Helped |
Closest papers by meaning
- Uncertain Graph Neural Networks for Facial Action Unit Detection (AAAI 2021 · similarity 0.84)
- Knowledge Augmented Deep Neural Networks for Joint Facial Expression and Action Unit Recognition (NeurIPS 2020 · similarity 0.84)
- Dynamic Probabilistic Graph Convolution for Facial Action Unit Intensity Estimation (CVPR 2021 · similarity 0.84)
- Facial Action Unit Detection With Transformers (CVPR 2021 · similarity 0.83)
- Probabilistically Rewired Message-Passing Neural Networks (ICLR 2024 · similarity 0.82)
- Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs (ICML 2024 · 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.