面向情感障碍的情绪识别方法、装置、设备和介质
By integrating dialogue text and speech features from different scenarios into emotion recognition, the problem of poor emotion recognition performance is solved, achieving more accurate and reliable emotion recognition, which is applicable to the diagnosis of bipolar disorder and medical record quality inspection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ANHUI IFLYHEALTH CO LTD
- Filing Date
- 2022-11-14
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies do not distinguish between specific emotion recognition scenarios, resulting in the extraction of limited emotional information. This makes it difficult to effectively collect targeted emotional information about users in specific scenarios, leading to poor recognition performance.
By acquiring dialogue text and speech from at least two scenarios, feature fusion is performed based on the correlation between semantic and speech features. Emotion recognition is then performed using the fused semantic and speech features. Multilayer convolutional neural networks and self-attention coding modules are used for feature extraction and fusion.
It improves the reliability and accuracy of emotion recognition, and can extract more complete semantic and speech feature information, ensuring that complete emotion information is referenced during the emotion recognition process. It is suitable for emotion recognition and clinical diagnosis of bipolar disorder.
Smart Images
Figure CN115905532B_ABST