Emotion recognition of human body's posture in open environment

Abstract

This paper proposes and designs an algorithm of human pose emotion recognition based on visible light image in open environment,It mainly summarizes the emotions represented by different postures from the existing body language literature, then extracts a number of key nodes of human body parts from the induced different postures to connect and draw a simplified line map, uses CPM human body key node detection algorithm to extract the coordinates of human body key nodes in the image collected by the camera, draws the simplified line map of human body postures through clustering algorithm, and finally draws the drawn line map The simplified line map human posture map is input to CNN + softmax algorithm for classification, and the emotion information expressed by different posture is recognized. The human posture emotion recognition algorithm designed by this method can effectively monitor the attitude emotion of the target object through the visible light camera in the open environment.