tailieunhanh - Báo cáo khoa học: "Transonics: A Practical Speech-to-Speech Translator for English-Farsi Medical Dialogues"

We briefly describe a two-way speech-tospeech English-Farsi translation system prototype developed for use in doctorpatient interactions. The overarching philosophy of the developers has been to create a system that enables effective communication, rather than focusing on maximizing component-level performance. The discussion focuses on the general approach and evaluation of the system by an independent government evaluation team. | Transonics A Practical Speech-to-Speech Translator for English-Farsi Medical Dialogues Emil Ettelaie Sudeep Gandhe Panayiotis Georgiou Kevin Knight Daniel Marcu Shrikanth Narayanan David Traum University of Southern California Los Angeles CA 90089 ettelaie@ gandhe@ georgiou@ knight@ marcu@ shri@ traum@ Robert Belvin HRL Laboratories LLC 3011 Malibu Canyon Rd. Malibu CA 90265 rsbelvin@ Abstract We briefly describe a two-way speech-to-speech English-Farsi translation system prototype developed for use in doctorpatient interactions. The overarching philosophy of the developers has been to create a system that enables effective communication rather than focusing on maximizing component-level performance. The discussion focuses on the general approach and evaluation of the system by an independent government evaluation team. 1 Introduction In this paper we give a brief description of a two-way speech-to-speech translation system which was created under a collaborative effort between three organizations within USC the Speech Analysis and Interpretation Lab of the Electrical Engineering department the Information Sciences Institute and the Institute for Creative Technologies and the Information Sciences Lab of HRL Laboratories. The system is intended to provide a means of enabling communication between monolingual English speakers and monolingual Farsi Persian speakers. The system is targeted at a domain which may be roughly characterized as urgent care medical interactions where the English speaker is a medical professional and the Farsi speaker is the patient. In addition to providing a brief description of the system and pointers to pa pers which contain more detailed information we give an overview of the major system evaluation activities. 2 General Design of the system Our system is comprised of seven speech and language processing components as shown in Fig. 1. Modules communicate using a .