During the last decade, the most significant advances in the field of continuous speech recognition (CSR) have arisen from the use of hidden Markov models (HMM) for acoustic modeling. These models address one of the major issues for CSR: simultaneous modeling of temporal and frequency distortions in
A probability decision criterion for speech recognition
β Scribed by C.K. Yu; P.C. Ching
- Publisher
- Elsevier Science
- Year
- 1991
- Tongue
- English
- Weight
- 441 KB
- Volume
- 5
- Category
- Article
- ISSN
- 0885-2308
No coin nor oath required. For personal study only.
π SIMILAR VOLUMES
The standard method for making the full content of audio relevant selections, and permit them to be reused effecand video material searchable is to annotate it with hutively. ## man-generated meta-data that describes the content in Through the integration of technologies from the fields a way tha
A switched-capacitor (SC) preprocessing system (preprocessor) which extracts and emphasizes the local peaks of the spectrum in real time is proposed for speech recognition systems. Main components of the system are a specially designed bandpass filter bank, a low-pass decimation filter bank, two-dim