In this paper, a novel method is proposed to extract hand gesture features in real-time from RGB-D images captured by the Microsoft's Kinect. A contour length information based de-noise method is introduced for the hand gesture smooth segmentation and edge contour extraction. In addition, a finger earth mover's distance algorithm is applied with a novel approach to locate the palm image and extract fingertip features. Especially the proposed Lasso algorithm can effectively extract the fingertip feature from a hand contour curve correctly with excellent real-time performance.
|Conference||2014 International Conference on Machine Learning and Cybernetics|
|Abbreviated title||ICMLC 2014|
|Period||13/07/14 → 16/07/14|