Speech Processing Algorithm for Automatic Speaker Recognition-Based Telephone Voice Dialing in Yorùbá

Abstract

The commonly used method for telephone dialing is through the use of the phone key pad. This form of input is cumbersome and not user friendly to some people (drivers, handicaps etc.). These problems have led to a lot of researches in development of natural and user friendly input such as speech. Most of the works in this area are in English, French, Chinese and other languages. This makes very important a research in using standard Yorùbá as an input in telephone voice dialing. The telephone voice dialing in Yorùbá is achieved through speech recognition training phase and recognition phase of the algorithm. The raw speech captured were digitalized and processed to achieve the speech signal which was used in the pattern matching of the system to allow authorized users and disallow the unauthorized users from using the system. The telephone voice dialing algorithm was developed through pre-emphasis filter, framing and windowing and dynamic time warp. With the design, telephone voice dialing in other languages could also be developed following the same procedures.