Static hand gesture recognition algorithm based on finger angle characteristics

In this paper, we propose a static gesture recognition algorithm based on finger angle characteristics. Using a fingertip attachment, a gesture is determined by the angle of the fingers relative to the centre of the palm. According to the angle and number, we can define a gesture as one to nine. The algorithm does not significantly consider the direction or scale of the gesture. It considers only the finger angle size. In test experiments, we used 900 gesture images, and the accuracy rate was as high as 96.8%, with an average determination time of less than 0.05 seconds. The experiment results demonstrate that the algorithm is efficient and is able to satisfy practical applications.