Albanian Dynamic Dactyls Recognition using Kinect Technology and DTW
Eriglen Gani, Alda Kika · 2017
Sign language is the primary way of communication for the people with hearing impairment. Every region and country has its own sign language. Albanian finger-spelling gestures for expressing the alphabet letters are composed of static and dynamic gestures. This paper presents the study that is conducted to recognize the gestures in the dactyls dynamic set. Depth and skeleton features offered by Kinect sensor are used for the segmentation of the interested area which corresponds to the hand of the user. A pre-processing technique has been applied on segmented hand to extract the contour of the hand and the center of the palm. These data will serve as similarity measurement between the registered sign and the test sign. DTW is used to compare the similarity between a reference and a captured gesture based on the assumption that all the gestures are dynamic. The proposed system is capable of successfully detecting dynamic dactyls in AlbSL with an accuracy of 91.25%.