Sign Language Identification Using Deep Learning Methods

Hearing impaired persons communicate via hand gestures, so regular people have difficulty understanding their language. Due to this, technologies that recognize various gestures and communicate the information to ordinary people are required. In this Identification technique, we use the Convolution Neural Network (CNN). These are strong Image Processing and AI technologies that employ Deep Learning to carry out expressive and rewarding duties. It uses a system that is like the Multilayer Perceptron which has the function of reducing processing requirements. The CNN is mainly comprised of an Input layer, an Output layer, and a Hidden layer (which comprises a large no of convolution layers), Pooling layers, Fully-Connected layers, etc. The person will give their hand sign as an image in our project. The letters and words can then be converted using images and inputs. The dataset used in our project has been created by us solely. The hand signal will now be given as input, and the neural network will anticipate what it means and provide the results. It will be given a dataset and, based on the analysis, will attempt to provide the desired outcome. It is anticipated that a 98% accuracy rate is achieved at the end. It can be scaled up much further by adding inputs and enhancing our Neural Net’s efficiency.