AI toy personalized interaction generation method and system based on multi-modal emotion recognition
By collecting children's voice and facial images in real time, calculating confidence scores and dynamically weighting and fusing them, and combining them with long-term memory profiles to generate interactive decisions, the system drives a large language model to generate personalized content. This solves the technical challenges of AI toys in emotion recognition and interaction continuity, and enhances the experience of emotional understanding and empathetic companionship.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHENZHEN UASCENT TECH CO LTD
- Filing Date
- 2026-06-08
- Publication Date
- 2026-07-17
AI Technical Summary
Existing AI toys are susceptible to interference in single-modal recognition of emotions, lack multimodal fusion decision-making, and lack personalized interaction continuity, resulting in inaccurate emotion judgment and a lack of continuity in content generation.
By collecting children's voice and facial images in real time, extracting emotional features and calculating confidence scores, and then dynamically weighting and fusing them with long-term memory profiles to generate interactive decision parameters, the large language model is driven to generate personalized interactive content and perform emotional speech synthesis.
It enables multimodal emotion-driven personalized interaction, improves the accuracy of emotion understanding, interaction coherence and empathic companionship experience, and solves the problems of single-modality susceptibility to interference and isolated content generation.
Smart Images
Figure CN122417023A_ABST