Neural Network Based Emergent Feedback Interface for Articulation of Language Sounds

This paper present an improved feedback system for articulation of Tamil speech sounds. This work trying to enhancing a visual feedback using checking each trained network to show the correct and incorrect articulation basis on input speech with MRI images. In this approach, for the user understandability each input speech will be displayed as text by checking each trained network. Speech Corpus is developed from ten children (5 boys and 5 girls from Normal Hearing) in the age group of 6-10 years. This system based on Neural Networks concepts for improving the Tamil Speech Recognition. The Radial basis Network used for training of individual networks, gender wise with the speech corpus of 25 Tamil phonemes.