Diagnostics of Speech Recognition using Classification Phoneme Diagnostic Trees

M. Cernak and C. Wellekens (France)


Fault diagnosis, speech recognition, intrinsic speech vari abilities.


More than three decades of speech recognition research re sulted in a very sophisticated statistical framework. How ever, less attention was still devoted to diagnostics of speech recognition; most previous research report on re sults in terms of ever-lower WER in various intrinsic or environmental conditions. This paper presents a diagnostics of the decoding pro cess of ASR systems. The purpose of our diagnostics is to go beyond standard evaluation in terms of WERs and confusion matrices, and to look at the recognized output in more details. During the decoding phase, some specific data are collected at the decoder as possible causes of er rors, and later are statistically analyzed using classification and regression trees. Focusing on pure acoustic phone de coding without language modeling, we present and discuss the results of the diagnostics that is used for an analysis of impact of intrinsic speech variabilities on speech recogni tion.

Important Links:

Go Back