Simulation-Based Model Control Using Static Hand Gestures in MATLAB

Abstract

This paper deals with the domain of simulation-based models control using static hand gestures in the MATLAB environment. The aim of this paper was to design an algorithm for visual static hand gesture recognition with high classification accuracy. For this recognition task, different convolutional neural network models (CNN) were tested. For the successful training of CNN, stochastic backpropagation of error was used. Training of CNN was implemented on the graphic card using toolboxes such as Neural Network and Parallel Computing from the MATLAB program package. For the training and testing of CNN a database of 35 static hand gestures was used. The proposed CNN gesture recognition system has been implemented in the simulation scheme due to the need of setting different model parameters.