Abstract :
— The computer vision based stress identification
system observes the non verbal gestures such as facial
expressions using camera and verbal gestures such as
Speech using microphone. While combining the
observations on speech and gestures, the system produces
high level context sensitive interpretation of human
behaviour. The automatic stress prediction is done based on
a decomposition of stress into a set of intermediate level
variables. This model has an extra edge that uses ARM11
processor and acoustic model in which the processor finds
the contour points based on stress identified from nonverbal gestures of a person and the acoustic model which
stores voice modulation data in various frequencies. System
will compare those observations and send the stress
information about the person to the service desk. The goal
is to provide a surveillance system that notifies the stress
levels
Keyword :
Stress, surveillance, speech, gestures, total harmonic distortion, modulation, affective computing, stress recognition.