Effective speech recognition for healthcare industry using phonetic system

Abstract

The automatic speech recognition helps to achieve today’s demands such as flexibility in patient care, efficiency, medical records. ASR allows more effective use and combination of process management devices and systems. Because speech interaction is contactless, they can be seamlessly combined into a current hardware environment. This paper presents the phonetic system that implemented to improve the automatic speech recognition with higher accuracy for increasing performance. The system obtains input speech by a mic then works on the tried speech to recognize the spoken word. After that, it passes the ensuing text to the HMM classifier. The HMM classifier compares occurrence of the accredited word with probability map. The word with the highest probability of occurrence gets selected. It then substitutes accredited word with this utterance; this process is carried out for the entire accredited text. The phonetic system directly obtains and translates speech to text by providing 8% improvement in the accuracy of the system. Smart text independent multi-lingual SMS system is developed using phonetic system, which allows the user to convert their voice into text and send message. STIM SMS system can offer a very spirited substitute to traditional keyboard.