Abstract

At present, subjective questionnaire surveys and physiological parameters measured are mainly used to collect the comprehensive response of humans to the environment, evaluating the environmental comfort subjectively and objectively. However, some measuring instruments make people feel uncomfortable and interfere with the natural state of the human body. Some measuring instruments are too bulky to carry. In addition, the changes in physiological parameters are not synchronized with the changes in human feeling, and there is a lag phenomenon, which cannot immediately reflect the human body current feeling. This study considers that the facial expression is an expression form of human emotion, psychology, and brain activities. When touch, hearing, vision, taste, and smell are stimulated by external factors, the facial expression will change. In view of this, this study examines the thermoacoustic environmental comfort based on facial micro-expression recognition. Firstly, the facial micro-expression database based on environment comfort (FMEEC) is constructed. Secondly, a convolutional neural network (CNN) is used to build the micro-expression recognition model (MERCNN). Finally, the MERCNN model is trained to converge with training and verification samples. The micro-expression images collected in the artificial climate chamber and classroom environment are used to test the model's performance. The evaluation results of the MERCNN are compared with the results of subjective questionnaires and the predicted mean vote (PMV) model. The results show that the MERCNN can accurately evaluate the thermoacoustic environment comfort through face micro-expression recognition.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call