Punjabi speech to text system for connected words

The ultimate goal of research on Speech to Text system is to build machines that are indistinguishable from humans in the ability to communicate in natural spoken language. This paper discusses the implementation of a connected word Speech to Textsystem (STT) for the Punjabi language. Hidden Markov model toolkit (HTK)has been used to develop the system. A Java platform based Graphical User Interface (GUI) has been developed to make the system fast and user friendly. The implemented system performs well with Word Recognition Rate (WR) 95.8% and 95.4%, Word Accuracy Rate (WA) 94.1% and 91.6% and Word Error Rate (WER) 5.9% and 8.35% in class room and open environment respectively.