Fusing Multimodal features for Recognizing Hand Gestures

Abstract

Human computer interaction systems based on hand gestures are getting much attention from the research community for establishing realistic communication between human and machines. However, different persons perform the same gestures in different manner in terms of velocity and motion scale. This poses a challenging problem to reduce the variations between different persons and maximize the coherence of same gestures. In this paper, the original pyramid histogram of gradients in three orthogonal planes combining with optical flow to build dynamic descriptor is explored to discriminate features for recognition of hand gestures. The shape and motion features of images in a video sequence are captured to obtain the geometric and illumination invariant dynamic feature vector for classification. A multiclass support vector machine classifier is utilized here to detect the hand gestures. The proposed method gives excellent recognition rate and outperformed the existing approaches.