Research On The Application Of Camera-based Remote Photoplethysmography In Mental Stress Assessment

Mental stress(or stress)is a concept refers to the internal(personality and thinking mode,etc.)or external(life and work affairs and social relations,etc.)factors those can induce negative emotions.At present,the assessment of mental stress is mainly based on the detection of secretions such as salivary cortisol,noradrenaline and ?-salivary amylase;besides,behavioural bodily patterns and patients' complaints are also adopted.In this paper,a non-contact-mental stress assessment system is designed based on the camera-based remote photoplethysmography(rPPG).By the method,mental stress can be assessed without wearing any sensors on or biological samplings.Firstly,facial videos are recorded with camera.Next,the cardiac pulse wave is extracted from the facial videos by the selection of region if interest(ROI)and denoising.Then methods of feature extraction,such as harmonic analysis,time domain feature recognition and pulse rate variability(PRV)(in time domain,frequency domain and nonlinear analysis)analyzing,are applied on the cardiac pulse wave.Finally,the support vector machine(SVM)is used to generate a model for mental stress assessment.The results show that the average recognition rate of this system reaches 68.32%.Although the method introduced in this paper does not have a better performance than those based on electroencephalogram(EEG)and electrocardiogram(ECG)which have a higher recognition rate of 80%to 90%,its advantages in convenience and cost make it promising.