Development of Dynamic Vocalization Customer Service Software for Multiple Languages for Telephone Environment

Abstract

This paper aims to design and develop a dynamic vocalization interface as telephone agent assistant. This interface uses tools such as, speech-to-text and text-to-speech software to translate the users' requests using a telephone. It consists of a number of libraries for recognition of multiple languages. This application aims for non-native individuals or groups who have accents and therefore have difficulties to utilize the speech recognition applications effectively. Since, the speech recognition engines utilized in telephone systems are generally developed for native language speakers. Although recent speech recognition systems have very intelligent speech recognition engine which can be tuned to any specific accent. They are not good candidates for the service industry, since a variety of users with a variety of accents use these applications for very short period of time and the speech recognition engines of these applications do not have enough time to adapt to many varieties of accents or pronunciations. In this paper, we develop a dynamic vocalization interface which can overcome this shortcoming and allows users with a variety of accents to effectively use the automated telephone systems and request services for their daily requirements. This application can also be utilized in various domains such as: health care, automobile industries to facilitate easier human to computer interactions.