Model parameter estimation for mixture density polynomial segment models

Toshiaki Fukada, Yoshinori Sagisaka, Kuldip K. Paliwal

Research output: Contribution to journalConference article

20 Citations (Scopus)

Abstract

In this paper, we propose parameter estimation techniques for mixture density polynomial segment models (henceforth MDPSM) where their trajectories are specified with an arbitrary regression order. MDPSM parameters can be trained in one of three different ways: (1) segment clustering, (2) expectation maximization (EM) training of mean trajectories, or (3) EM training of mean and variance trajectories. These parameter estimation methods were evaluated in TIMIT vowel classification experiments. The experimental results showed that modeling both the mean and variance trajectories are consistently superior to modeling only the mean trajectory. We also found that modeling both trajectories results in significant improvements over the conventional HMM.

Original languageEnglish
Pages (from-to)1403-1406
Number of pages4
JournalICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
Volume2
Publication statusPublished - 1997 Jan 1
EventProceedings of the 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP. Part 1 (of 5) - Munich, Ger
Duration: 1997 Apr 211997 Apr 24

    Fingerprint

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this