2014 IEEE Global High Tech Congress on Electronics (GHTCE 2014)

103 / 1971-01-01 00:00:00

Hand Gesture Recognition Based On Geometry And Ccm Feature Analysis

HCI ,Hand gesture,CCM

Hand gesture plays a great role in making human computer interaction (HCI) more natural and intimate. The paper addresses geometry and CCM (Color co-occurrence matrix) analysis using both contour and local color features to accomplish an accurate hand gesture recognition system. First, Hand skin is detected by color information with experimental range and adaptive model. After that geometry feature analysis can locate the centroid of hand by Euclidean distance transform (EDT) of binary image, and extreme points of Euclidean distance(ED) between boundary and centroid correspond to fingertips. CCM analysis is performed to determine the complexity of the surface of hand, so it can enrich the gestures because plain hand skins have less complexity than skins with fingers etc. on them. Experimental results show that the proposed system can perform hand gesture recognition correctly.