Hand Gesture Recognition for Medical Purposes Using CNN.

The presented paper describes implementations of the gesture recognition methods based on the convolutional neural networks. For this purpose, we adopted three CNN structures. The data was obtained from a specially prepared dataset containing images with hand gestures taken on a green screen and downloaded backgrounds depicting hospital and office conditions. During experiments, the precision of recognizing individual gesture classes was measured. Experiments were carried out that showed the performance time of the gesture recognizing in images using a GPU card.