Touch Talk

T-Talk aims at providing a fast and easy way for the speech impaired to communicate. It essentially converts the inputs of the user into efficient speech output. This being the era of “touch”, we aim at replacing the discrete and slow keyboard inputs with a much faster, intuitive and continuous input medium, i.e., THE TOUCH SCREEN. It is related to the idea of projecting the sign language in 3D into a 2D sheet and processing it into speech. The problem of efficient human-like speech synthesis is yet to be solved completely. The usage of neural networks and touch screen interface to achieve this is quite an innovative approach. The project provides a great platform to integrate stroke recognition, neural network systems, text-to-speech engines and most importantly the essence of speech itself! As of now we are developing it for the Devanagiri script, but the idea is very easily extensible to all Indian languages. Subsequent improvements can involve multi-touch interfaces and highly versatile speech engines and embedded systems. Ultimately the aim is to make everyone talk!