Enhancing phoneme recogniser performance with a simple rule-based language model

Pertti Väyrynen, Johannes Peltola, Tapio Seppänen

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review


The phoneme classification inaccuracy at the acoustic phonetic level is a major weakness in most speech recognition systems. However, the inaccuracy will violate phonotactic constraints at the acoustic phonetic level. A better performance is expected if a language model is adopted in a recognition system for post-processing phoneme estimates and making corrections with a set of explicit rules of the language used. We developed a simple language model with a set of rules to correct phoneme classification errors made by a Hidden Markov Model based phoneme recognizer. The experimental results indicate that about 20% of
recognition errors can be corrected.
Original languageEnglish
Title of host publicationSTeP2000 Millennium of Artificial Intelligence
Subtitle of host publicationAl of Tommorrow
Number of pages7
Publication statusPublished - 2000
MoE publication typeA4 Article in a conference publication
Event9th Finnish Artificial Intelligence Conference : Millennium of Artificial Intelligence - Espoo, Finland
Duration: 28 Aug 200030 Aug 2000
Conference number: 9


Conference9th Finnish Artificial Intelligence Conference
Abbreviated titleSTeP 2000

Fingerprint Dive into the research topics of 'Enhancing phoneme recogniser performance with a simple rule-based language model'. Together they form a unique fingerprint.

Cite this