Increasing Accuracy of Hand Gesture Recognition using Convolutional Neural Network

Human gestures play important roles in the interaction between humans and machines. These human gestures are becoming more important, yet complex gesture input and noise induced by external elements are important problems to solve in order to improve the accuracy of hand gesture recognition methods. Convolutional Neural Networks (CNN) are offered as a technology that can solve this problem in this research. CNN has the advantage of being able to learn image data, and this technology will greatly improve human-machine interaction accuracy. Data was extracted using Vivaldi antennas with a frequency bandwidth of 7.4-9.0 GHz and gain characteristics of 8 dB in five sign language operations, and data that went through the preprocessing process was learned through CNN. The classification results of the proposed CNN showed about 90% accuracy.