Hand Gesture Recognition Using CNN

Abstract

Computer is useful and important in our life and used in different fields. Interaction between computer and human is accomplished with computer input devices like mouse, printer, keyboard, etc. Using hand gestures as useful medium between human and computer can make the connection easier for disable persons. Disable persons do not have an interaction with normal people because, the normal persons do not know the sign language. We use the CNN for the recognition of hand gestures. As CNN can intake image processing, we use this technique. Recent research has proved the supremacy of convolutional neural network. Whatever, the images that can be captured are compared with datasets and compared the accuracy and give the message of respected dataset. With this software, the disable people can communicate with others. The model of augmented data received accuracy 97.12% which is nearly 4% higher than the model without augmentation (92.87%). So, nonlinearity exists in this type of problem. Also, we added some applications such as mouse movement and volume changes.