Speaker-Dependent Based Speech Recognition

摘要

This is the first part of the two parts of a qualitative focused R&D research aimed at designing an application to assist students with visual impairment (VI) in learning English writing and reading skills. The designed application was a speaker-dependent based speech recognition. Conducting alpha and beta testings, it was revealed that MAKTUM, the name of the application, exposed weaknesses on the selection of Ogden\u27s Basic English as the linguistic resources for the application and on the recording complexities. On the other hand, MAKTUM displayed strengths in individualized pronunciation and simple interfaces to operate.