Data processing: speech signal processing – linguistics – language – Speech signal processing – Recognition
Reexamination Certificate
1999-02-23
2001-08-14
Dorvil, Richemond (Department: 2741)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Recognition
C704S233000, C704S226000
Reexamination Certificate
active
06275800
ABSTRACT:
FIELD OF THE INVENTION
The present invention pertains to voice recognition.
BACKGROUND OF THE INVENTION
Speaker dependent speech recognition systems use a feature extraction algorithm to perform signal processing on a frame of the input speech and output feature vectors representing each frame. This processing takes place at the frame rate. The frame rate is generally between 10 and 30 ms, and will be exemplified herein as 20 ms in duration. A large number of different features are known for use in voice recognition systems.
Generally speaking, a training algorithm uses the features extracted from the sampled speech of one or more utterances of a word or phrase to generate parameters for a model of that word or phrase. This model is then stored in a model storage memory. These models are later used during speech recognition. The recognition system compares the features of an unknown utterance with stored model parameters to determine the best match. The best matching model is then output from the recognition system as the result.
It is known to use a Hidden Markov Model (HMM) based recognition system for this process. HMM recognition systems allocate frames of the utterance to states of the HMM. The frame-to-state allocation that produces the largest probability, or score, is selected as the best match.
Many voice recognition systems do not distinguish between valid and invalid utterances. Rather, these systems choose one of the stored models which is the closest match. Some systems use an Out-of-Vocabulary rejection algorithm which seeks to detect and reject invalid utterances. This is a difficult problem in small vocabulary, speaker dependent speech recognition systems due to the dynamic size and unknown composition of the vocabulary. These algorithms degrade under noisy conditions, such that the number of false rejections under noisy conditions increases.
In practice, out-of-vocabulary rejection algorithms must balance performance as measured by correct rejections of invalid utterances and false rejections of valid utterances. The false rejection rate can play a critical role in customer satisfaction, as frequent false rejections, like incorrect matches, will cause frustration. Thus, out-of-vocabulary rejection is a balance of meeting user expectations for recognition.
Accordingly it is known to calculate a rejection threshold based upon the noise level. For example, it is known to measure the noise level prior to the detection of the first speech frame. A threshold is calculated from the measurement. An input is rejected if the difference between the word reference pattern and the input speech pattern is greater than the rejection threshold. Such a system is thus dependent upon an arbitrary noise input level. Such measurement can not be relied upon to produce a meaningful rejection decision.
Accordingly, there is a need for an improved method of providing a basis for rejecting utterances in a voice recognition system.
REFERENCES:
patent: 5201004 (1993-04-01), Fujiwara et al.
patent: 5386492 (1995-01-01), Wilson et al.
patent: 5416887 (1995-05-01), Shimada
patent: 5749068 (1998-05-01), Suzuki
patent: 5778342 (1998-07-01), Erell et al.
patent: 5793863 (1998-08-01), Hashimoto
patent: 5832430 (1998-11-01), Lleida et al.
patent: 5864804 (1999-01-01), Kalveram
patent: 5960397 (1999-09-01), Rahim
patent: 5970446 (1999-10-01), Goldberg et al.
patent: 6067513 (2000-05-01), Ishimitsu
Chevalier David Erik
Kazecki Henry L.
Abebe Daniel
Dorvil Richemond
Motorola Inc.
Soldner Michael C.
Vaas Randall S.
LandOfFree
Voice recognition system and method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Voice recognition system and method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Voice recognition system and method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2464221