Hand Gesture Recognition

Abstract

The goal of this research is to detect and recognize hand motions. Photographs of hand gestures are captured using a camera and compared to images in the database; all images are divided in testing and training data set and converted into train. Record and test. record. This. records file trained in TensorFlow Mobile Net pre trained model. Loading of this model using karas gives us the best match of video frame gesture and saved gesture, and the gesture label is shown according to match. One of the most important strategies for removing a dumb person’s communication barrier is gesture recognition. For example, if a person who is unable to speak wishes to communicate with another person, gesture recognition language will enable them to do so.The accuracy of the below proposed model is 93%.This model use Convolution Neural Network. All gesture signs which display on screen will converted into text file and this text file converted into mp3 using gTTs API. so this file can be share to anyone and communication happen smoothly. A hand gesture recognition system allows for nonverbal communication that is natural, inventive, and modern. Its applications in human-computer interaction and sign language are numerous