Retracted: A hybrid HMM/ANN approach for automatic Gujarati speech recognition

摘要

Speech processing is rising as one of the basic petition district of computerized sign process. Discourse acknowledgment is the working of naturally perceiving the talked expressions of individual in view of information substance in discourse sign. The programmed discovery of discourse is empowering regular and basic managing amongst human and machine. Communication has enormous application in voice dialing, phone correspondence, call directing, local apparatuses control, Speech to Text change, Text to Speech transformation, lip synchronization, robotization frameworks and so on. This paper presents a concise implementation a hybrid HMM/ANN approach on Automatic Gujarati Speech Recognition with the help of different feature techniques, combination classification method. The fundamental difficulties of discourse acknowledgment includes demonstrating the variety of an indistinguishable word from talked by Separate speakers depending on talking styles, vocabulary estimate, social vernaculars, sex etc.