A New Approach For Hand Gestures Recognition Based on Depth Map Captured by RGB-D Camera

Abstract

We present in this paper a new approach for hand gestures recognition based on depth Map captured by a RGB-D camera like kinect. Although this camera provides two types of information "Depth Map" and "RGB Image", we use only the depth data information to analyze and recognize the hand gestures. Given the complexity of this task, we propose in the first step a new method based on edge detection to eliminate the noise and segment the hand. Then we have introduced new descriptors modeling the hand gesture. These features are invariant to scale, rotation and translation. Our approach has been applied on Alphabet French sign language to show its effectiveness and to evaluate the robustness of the proposed descriptors. The experimental results clearly show that the proposed system is very satisfactory and is able to recognize the French alphabet sign with an accuracy more than 93%. Our approach has been also applied on a public dataset in order to be compared to existing works. The results prove that our system can outperform previous methods using the same database.