Submitted in fulfillment of the academic requirements for the degree of Master of Technology in Electrical Engineering in the Department of Electronic Engineering, Faculty of Engineering, ML Sultan Technikon of Durban in South Africa, March 2002. / The use of speaker recognition technology in interactive voice response and electronic commerce systems has been limited. This is due to the lack of research attention and published results when compared to all the other areas of speech recognition technologies / M

This study considered the evaluation of commercially available isolated word, speaker dependent, speech recognition units, and preprocessing techniques that may be used for improving their performance. The problem was considered in three separate stages. A series of tests were designed to exercise an isolated word, speaker dependent, speech recognition unit. These tests provided a sound basis for determining a given unit's strengths and weaknesses. This knowledge permits a more informed decision on the best recognition device for a given price range. As well, this knowledge may be used in the design of a robust vocabulary, and creation of guidelines for best performance. The test vocabularies were based on the forty English phonemes identified by Rabiner and Schafer [28] and the test variations were representative of common variations which may be expected in normal use. A digital archive system was implemented for storing the voice input of test subjects. This facility provided a data base for an investigation of preprocessing techniques. As well, it permits the testing of different speech recognition units with the same voice input, providing a platform for device comparison. Several speech preprocessing and performance improvement techniques were then investigated. Specifically, two types of time normalization, the enhancement of low energy phonemes and a change in training technique were investigated. These techniques permit a more accurate analysis of the failure mechanism of the speech recognition unit. They may also provide the basis for a speech preprocessor design which could be placed in front of a commercial speech recognition unit. A commercially available speech recognition unit, the NEC SR100, was used as a measure of the effectiveness of the tests and of the improvements. Results of the study indicated that the designed tests and the preprocessing & performance improvement techniques investigated were useful in identifying the speech recognition unit's weaknesses. Also, depending on the economics of implementation, it was found that preprocessing may provide a cost effective solution to some of the recognition unit's shortcomings. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

This dissertation presents a design of a robust, automatic speaker recognition (ASR) system. The ASR system is designed to work with both text-independent and text-dependent speaker recognition. Several speaker spectral features are studied to determine their contribution in term of accuracy to the system. A new algorithm is designed to label a speaker voice as either male-type voice or female-type voice. Following this division, the processing time of the speaker identification for the ASR system will be reduced by about half. Rectangular window, Hamming window, first order preemphasis filter, and many proposed spectral distances are also investigated. The principal components analysis is used to achieve high degree of female-type and male-type separation as well as the speaker recognition accuracy. Spectral features are combined to improve the recognition performance of the system. In addition, many other system components such as speech endpoint detection, automatic noise thresholds, etc. are required to build correctly in order to achieve high speaker recognition accuracy. Multi-stage decision process is used both to improve and to speed up the decision if certain criteria are met. Finally, TIMIT acoustic continuous speech corpus is used to evaluate the speaker recognition performance and the robustness of the system. / Ph. D.

This feasibility study examined the possibility of using an independent voice recognition system as the input device during a training delivery requirement. The intent was to determine whether the voice recognition system could be incorporated into a training delivery system designed to train students how to use the Communications Electronics Operating Instructions manual, a tool used for communicating over the radio network during military operations. This study showed how the voice recognition system worked in an integrated voice based delivery system for the purpose of delivering instruction. An added importance of the study was that the voice system was an independent speech recognition system. At the time this study was conducted, there did not exist a reasonably priced speech recognition system that interfaced with both graphics and authoring software which allowed any student to speak to the system without training the system to recognize the individual student's voice. This feature increased the usefulness and flexibility of the system. The methodology for this feasibility study was a development and evaluation model. This required a market analysis, development of the voice system and instructional course ware, testing the system using a sample population from the Armor School at Ft. Knox, Kentucky, and making required alterations. The data collection approach was multifaceted. There were surveys to be completed by each subject: a student profile survey, a pretest, a posttest, and an opinion survey about how well the instruction met expectations. Data was also collected concerning how often the recognition system recognized, did not recognize, or misrecognized the voice of each subject. The information gathered was analyzed to determine how well the voice recognition system performs in a training delivery application. The findings of this feasibility study indicated that an effective voice based training delivery system could be developed by integrating an IBM clone personal computer with a graphics board and supporting software, signal processing board and supporting software for audio output and input, and instructional authoring software. Training was delivered successfully since all students completed the course, 85% performed better on the posttest than on the pretest, and the mean gain scores more than satisfied the expected criterion for the training course. The misrecognition factor was 12%. An important finding of this study is that the misrecognition factor did not affect the students' opinion of how well the voice system operated or the students' learning gain. / Ed. D.