New LaCoVL-FER network improves facial expression recognition with landmark guidance

Researchers have developed LaCoVL-FER, a novel network designed for facial expression recognition (FER) in challenging real-world conditions. This system employs a landmark-guided adaptive encoder to refine visual features by integrating facial landmark geometry, thereby reducing noise and enhancing expression-relevant representations. Additionally, a vision-language enhancement strategy adapts pre-trained models like CLIP to generate instance-specific visual and textual representations, improving robustness and generalization. Experiments on benchmark datasets such as RAF-DB, FERPlus, and AffectNet demonstrate that LaCoVL-FER surpasses existing state-of-the-art methods. AI

IMPACT This research advances facial expression recognition by improving model robustness and generalization in real-world scenarios.

RANK_REASON The cluster contains a research paper detailing a new model for facial expression recognition. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• AffectNet
• Bi-branch Gated Cross Attention
• Expression-Conditioned Prompting
• FERPlus
• Jiaxin Wang
• LaCoVL-FER
• Landmark-Guided Adaptive Encoder
• RAF-DB
• Vision-Language Enhancement Strategy

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →