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 language | English |
---|---|
Pages (from-to) | 1403-1406 |
Number of pages | 4 |
Journal | ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings |
Volume | 2 |
Publication status | Published - 1997 Jan 1 |
Externally published | Yes |
Event | Proceedings of the 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP. Part 1 (of 5) - Munich, Ger Duration: 1997 Apr 21 → 1997 Apr 24 |
ASJC Scopus subject areas
- Software
- Signal Processing
- Electrical and Electronic Engineering