Artificial Intelligence Technology-Driven Teacher Mental State Assessment and Improvement Method

Abstract

With the development of technology, people expect real-time communication with computers. Wearable devices, such as those for monitoring physiological signals, have rapidly developed and are now being applied in college and university evaluation. Due to the non-standard and unscientific practices in teaching, teachers may experience psychological obstacles when evaluating students. To ensure successful evaluation, we must motivate teachers to correctly understand and actively participate in the evaluation process, thus facilitating communication between people and computers. Emotion recognition based on multi-physiological signals, such as ECG, pulse, electromyography, electrodermal, and respiratory signals, is an effective method for achieving this. This dissertation conducts in-depth research on the methods for emotion recognition based on multi-physiological signals. It explores feature extraction methods, feature selection, and fusion to provide objective assessments of physiological and psychological activity states, which are used as a basis for accurate emotional judgments.