An Incremental Learning Framework for Skeletal-based Hand Gesture Recognition with Leap Motion

Hand gesture recognition has become the focus of researchers lately because of its manifold applications in various fields. Leap Motion (LM) is a device to obtain useful and accurate information of the hand action, which is suitable for collecting the three-dimensional (3D) human hand gesture. In this paper, a novel framework which consists of an incremental learning (IL) algorithm without deep structure is proposed and applied to hand gestures classification that explicitly aimed to the LM data. The same datasets are used to train the proposed framework and the conventional Long Short Term Memory Recurrent Neural Network (LSTM-RNN). Due to the structural advantage of the proposed model, the recognition performance is improved distinctly in robustness and training time than the LSTM network. Moreover, convincing experiment results are given to illustrate that the solution is more efficient in static gesture classification.