Abstract
dc:description.abstractThis research develops an algorithm to perform isolated word recognition. A detailed description of the recognition system is presented in this thesis. We discuss detection of a spoken word from a recording using an end-point algorithm, extraction of the feature vectors from the sampled speech signal, quantization of the feature vectors into a codebook of a particular size, and recognition of the spoken word using discrete Hidden Markov Modeling (HMM) of the words. We discuss the discrete HMM in detail and explain why it is suitable for performing the recognition. A detailed explanation of the training algorithm used to train the HMM to recognize the words is presented. We evaluate the performance of the recognition system under different conditions such as varying the codebook size of the quantizer, the number of states in the HMM, and scaling issues. Experiments showed that a recognition rate of better than 95% was obtained upon using a codebook of size 128 and 4 states in the HMM. The words used for training the HMM were recognized with a rate of better than 99%.
Degree
thesis:*- Name thesis:degree_name
- Master of Science
- Level thesis:degree_level
- masters
- Discipline thesis:degree_discipline
- Electrical Engineering
- Department dc:contributor.department
- Electrical Engineering
- Grantor dc:publisher
- Virginia Tech
- Year dc:date.issued
- 1996
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Rangarajan, Karthik A.
Rights
dc:rights- Statement dc:rights
-
- In Copyright
- Licence dc:rights.uri
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Dc Identifier Other
- etd-11072008-063413
- OAI identifier oai:identifier
- oai:vtechworks.lib.vt.edu:10919/45466