On-line Hand Gesture Recognition using the Hypercolumn Neural Network Model

摘要

Gesture recognition is an appealing tool for natural interface with computers especially for physically impaired persons. In this paper, it is proposed to use Hypercolumn model (HCM), which is constructed by hierarchically piling up Self-organizing maps (SOM), as an image recognition system for gesture recognition since the HCM allows alleviating many difficulties associated with , gesture recognition. It is, however, required for on-line systems to reduce the recognition time to the range of normal video camera rates. To achieve this a new competition algorithm to reduce the network recognition time into the real-time range is introduced in this paper. The proposed competition algorithm is based on selecting subset from the most discriminate codebook of the network weights. This could drastically reduce the network recognition time into the range of real video camera rate. The experimental results to recognize on-line hand gestures using HCM are presented.