視覺化語音服務系統

Abstract

本論文提出了VSXML格式及其相關技術,使傳統語音服務可以迅速的被轉換成為手機上的應用程式，達成使用者不需全程聆聽語音服務的目標。整個系統包含了VSXML的格式定義、VSXML Designer和VSXML Renderer. VSXML是一個可將語音服務中各個語音問題單元結構化敘述的標記語言，VSXML Designer則為一個方便語音服務設計者將傳統語音服務視覺化，同時生成行動裝置適用的應用程式。VSXML Renderer則根據不同的VSXML檔案，在行動裝置上呈現出相對應的控制項，同時也負責記錄整個語音流程的使用者回應並執行自動撥出。 透過在Motorola和Nokia手機上的實作測試，我們證明了VSXML格式對於傳統語音服務的轉換能力與整體系統的可行性及簡易性。整體而言，經由本系統及VSXML格式，不需要中途查詢資料庫的語音服務都可以完全的被視覺化。

The thesis presents Voice Service XML (VSXML) technologies to transform the traditional voice services into the visualized ones for mobile phone users. The technologies contain VSXML formatted files, VSXML Designer, and VSXML Renderer. The VSXML file encapsulates the interactive Question Units plus the communication protocols during the voice services conversation. The VSXML Designer is a toolkit can provide visualizable and handy applications for voice service designers. The VSXML Renderer generates the corresponding controls of the voice service conversion according to the VSXML file. The techniques have been designed and implemented to Motorola and Nokia mobile phones. The thesis concludes the technologies are flexible and easy to adapt. In general, the voice services can be visualizable completely if there is no database accessing. We also addressed the future work of the technologies.