Toward human arm attention and recognition: Using a computer-vision- and neocognitron-based approach

Abstract

We aim to develop a vision system that focuses on the human arm and recognizes the arm posture in complex scenes. In this paper, we introduce a computer-vision-based approach and a neocognitron-type neural network. In our computer-vision-based approach, we propose a method for arm posture estimation using an arm model with three kinds of small areas in the complex background and clothes patterns. In our neocognitron-based approach, we investigate the functional similarity between the neocognitron and the ventral visual pathway. We examine to the expand neocognitron base on stereo. © 2008 Springer-Verlag Berlin Heidelberg.