Synthesis-based recognition of continuous speech

Publisher Website
Google Scholar
Abstract
An acoustic phonemic recognition system for continuous [human] speech is presented. The system uses both steady state and transition segments of the speech signal to achieve recognition. The information contained in formant transitions is used by the system by using a synthesis-based recognition approach. This improves the performance of the system considerably. Recognition of continuous speech is accomplished here in 3 stages: segmentation, steady-state recognition and synthesis-based recognition. The system was tried out on 40 test utterances, each 3-4 s in duration, spoken by a single male speaker and the following results are obtained: 5.4% missed segment error, 8.3% extra segment error, 52.3% correct recognition using only steady-state segments and 62.0% correct recognition using both steady-state and transition segments.

This publication has 1 reference indexed in Scilit: