Similarity Matching-Based Extensible Hand Gesture Recognition

This paper presents an accelerometer-based smart ring and a similarity matching-based extensible hand gesture recognition algorithm. Users can wear the ring to perform gestures in 2-D space. The accelerations of hand motions are collected by the three-axis accelerometer, which is integrated in the ring. We divide the gestures into two types, i.e., the basic gesture and the complex gesture, which can be decomposed into a basic gesture sequence. A segmentation algorithm is developed to identify individual gestures in a sequence automatically. To recognize the basic gesture, a simple but effective feature based on the average jerk is extracted. The recognized basic gesture is then encoded by a Johnson code. Finally, the complex gesture is recognized by comparing the similarity between the obtained basic gesture sequence and the stored templates. A library of eight basic gestures and 12 complex gestures is created, and the users can easily define and add their own gestures without pretraining. The model discussed in this paper achieves a basic gesture recognition rate of 98.9% and a complex gesture recognition rate of 97.2%. Compared with complete matching, the proposed algorithm based on similarity matching improves the complex gesture recognition rate ~12%. Experimental results have successfully validated the feasibility and effectiveness of the gesture decomposition and similarity matching-based gesture recognition algorithm.