A Distributed cultural tourist guide system using HMM-Based speech recognition and speech synthesis

Abstract

In this paper, we propose a cultural tourist guide system based on a voice user interface employing speech recognition and speech synthesis techniques. The proposed system first carries out hidden Markov model (HMM)-based speech recognition for the names of the departure and destination locations. Then, it searches for tourist attractions close to the tourism route and relays this information to the user via a geographic information system. Finally, the tourist information is announced back to the user using an HMM-based speech synthesis technique. In particular, speech recording, feature extraction and speech synthesis are carried out at the client, whereas the tour path search and speech recognition are carried out at the server. The proposed system showed high performance in terms of the average word error rate and intelligibility scores for speech recognition and speech synthesis, respectively. Moreover, a user satisfaction survey indicates that the proposed system provides useful tour guide information for cultural tourists. ©2013 International Information Institute.