Abstract: This paper addresses pose gesture from image sequences by converting them as classification problem. The first step of this work consist of a method to Extract edge features of image and classify images using optimum(bayes) statistical classifier method. We label the extracted data by viewpoints and actions. Next we define classes. Pose detection is based on optimum (bayes) statistical classifier method. We input data using CNN for image for grayscales conversion, smoothening of resultant image. Our proposed work gives results about the difference between pose detected during training and expected pose.

Keywords: pose gesture, Baysian Classifier, CNN,

| DOI: 10.17148/IARJSET.2021.8921