Design of Real-Time Psychological State Evaluation System Based on Fusion Detection of Facial Emotion and Pain Expressions
Xinmei Liu, Hao Yin, Xiangwei BaiAbstract
Objective
Single facial emotion detection cannot capture hidden pain-induced psychological pressure, leading to inaccurate recognition of subtle abnormal mental states and low accuracy of daily mental health monitoring. To solve this defect, this paper designs a real-time psychological state evaluation system integrating facial emotion and pain expression detection. It aims to realize comprehensive identification of implicit psychological pressure, track dynamic mental state changes, and provide reliable technical support for real-time early warning of potential mental health risks.
Subjects and Methods
This study collects facial image data of volunteers with different pain levels and emotional fluctuations. It adopts improved feature fusion neural network algorithm to realize joint detection of common emotional expressions and pain facial features. It establishes a quantitative evaluation model to map fusion facial features to individual psychological pressure values, dynamic mental states and overall mental health levels for real-time online assessment.
Results
Compared with single emotion detection technology, the proposed fusion detection system improves the recognition accuracy of hidden psychological pressure significantly. It can timely capture subtle deteriorations of mental states caused by physical pain and negative emotions, realize real-time quantitative scoring of mental health conditions, and reduce missing detection and misjudgment of potential psychological problems.
Conclusions
The fusion detection system breaks the limitation of traditional single-modal facial recognition. It achieves synchronous monitoring of external pain signals, internal psychological pressure and real-time mental states, supports efficient non-contact mental health screening, and provides an effective intelligent tool for daily psychological intervention and mental health early prevention.
Acknowledgement
Shanxi Key Research and Development Project "Development of Complex Circuit Board Fault Detection System" (No. 201903D121058).
Corresponding Author
Xinmei Liu, School of Artificial Intelligence, HeBei University of Engineering Science, Hebei, 050091, China; School of Information and Communication Engineering, North University of China, Shanxi, 030051, China.