Beena Ahmed, Fatima A. Raja, Ernesto D. Nuguid, and Shatha Al-Shmary
assistive software technology, assistive technology for learning disabilities, assistive technology for speech therapy
The purpose of this study was to identify acoustic features that can provide effective automatic discrimination between samples of speech with errors produced by children with childhood apraxia of speech (CAS) and error free speech for possible future use in an assistive speech therapy system. Voice samples from seven children previously diagnosed with CAS were used for acoustic analysis that involved derivation of the zero-crossing rate, energy and sub-band energies for a series of words the subjects were asked to repeat. The work in this project shows that a simple energy tracking method can be used to differentiate between words with errors and no errors as spoken by a child undergoing speech therapy.
Important Links:
Go Back
