Abstract

The implementation and evaluation of an automatic speech recognition (ASR) based task for a mobile, hand‐held device is presented. The use of ASR on mobile devices will make environmental noise, and the design of transducers and compensation algorithms for dealing with noise, critical issues for the success of ASR based services on these devices. A description of the task is provided along with a set of compensation techniques that are used to compensate speaker independent hidden Markov models (HMMs) with respect to environment and transducer variability. Data were collected in a prototype application environment for a form‐filling directory access task with speech and pen input and text output (48 subjects in an office environment; 21 in a cafeteria). A technique for combined environment/transducer compensation is presented for this task and is shown to significantly reduce the effects of …

Date

2000

Authors

Richard Rose, Shrikanth Narayanan, S Parthasarathy, Aaron Rosenberg, Bojana Gajic

Journal

The Journal of the Acoustical Society of America

Volume

108

Issue

5_Supplement

Pages

2575-2575

Publisher

Acoustical Society of America