Researchers have developed STAG, a novel spatial-temporal network designed to improve micro-expression recognition. This method addresses limitations in existing approaches by jointly modeling motion flow and adaptive facial connectivity, moving beyond apex-onset frames and separate spatial/temporal processing. STAG utilizes optical flow extraction with magnitude-based selection and temporal attention, combined with a dual-branch architecture featuring a graph attention network and a transformer encoder. A bidirectional cross-attention module refines spatial and temporal features, while AU-guided dynamic connectivity adapts region interactions based on muscle activation. AI
IMPACT This research could lead to more accurate and interpretable systems for analyzing subtle human emotions.
RANK_REASON The cluster contains a research paper detailing a new method for micro-expression recognition.
- 4DMEDICAL LIMITED
- arXiv
- CASME II: an improved spontaneous micro-expression database and the baseline evaluation
- Design Failure Mode Effects Analysis
- \NAME
- Sammy
- SMIC-HS
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
