Data processing: artificial intelligence – Neural network – Learning task
Reexamination Certificate
1998-03-18
2001-06-05
Chaki, Kakali (Department: 2122)
Data processing: artificial intelligence
Neural network
Learning task
C713S182000
Reexamination Certificate
active
06243695
ABSTRACT:
FIELD OF THE INVENTION
This invention relates in general to the field of classifiers, in particular to polynomial classifiers and more particularly to tree-structured polynomial classifiers.
BACKGROUND OF THE INVENTION
Software and hardware classifiers are used, among other things, to analyze segments of speech. Classifiers identify which class a particular speech segment belongs. A “class” can be, for example, spoken commands, spoken words, characteristics of a communication channel, modulated signals, biometrics, facial images, and fingerprints.
Modern classifiers use techniques which are highly complex when high accuracy classification is needed. For example, a traditional classifier needing high accuracy also needs large memory and computational resources because of complex polynomial structure. Also, modern classifiers for identifying a user of a system for granting access to the system consume large memory and computational resources while providing modest identification success.
Additionally, higher order polynomial based classifiers are often needed for accurate classification of data which have complicated shapes in feature space. Typical polynomial classifiers have a problem of exponential growth of the number of parameters as the order of the classifier increases. Again, such classifiers have large numbers of parameters which are computationally expensive to compute.
Previous work related to tree-structured classifiers has been in two primary areas, decision trees and neural tree networks (NTN). Typical decision trees consider one element at a time when making a decision. Considering one element at a time constrains the partitioning of the feature space to using discriminants which are perpendicular to the feature axes. Also, considering one element at a time is a severe limitation for problems which require more flexibility in the discriminant positioning (e.g., a diagonal discriminant). NTNs have been proposed as a solution for the limitation of discriminants which are perpendicular to the feature axis since NTNs are not constrained to perpendicular discriminant boundaries. However, NTNs are constrained to linear discriminant boundaries which pose limitations for problems having sophisticated class shapes in feature space.
Thus, what is needed is a system and method requiring less processing and data storage resources to produce improved classification of an unidentified class (e.g., spoken command, communication channel, etc.). What is also needed is a system and method wherein classifiers can achieve high performance classification of sophisticated class shapes in feature space. What is also needed is a system and method for granting access to a system when a user is identified as an approved user of the system.
REFERENCES:
patent: 4720863 (1988-01-01), Li et al.
patent: 4975960 (1990-12-01), Petajan
patent: 5365574 (1994-11-01), Hunt et al.
patent: 5390136 (1995-02-01), Wang
patent: 5412738 (1995-05-01), Brunelli et al.
patent: 5519865 (1996-05-01), Kondo et al.
patent: 5522012 (1996-05-01), Mammone et al.
patent: 5548647 (1996-08-01), Naik et al.
patent: 5592589 (1997-01-01), Poon
patent: 5634087 (1997-05-01), Mammone et al.
patent: 5657424 (1997-08-01), Farrele et al.
patent: 5677989 (1997-10-01), Rabin et al.
patent: 5687287 (1997-11-01), Ghandi et al.
patent: 5790758 (1998-08-01), Streit
patent: 5819222 (1998-10-01), Smyth et al.
patent: 5839103 (1998-11-01), Mammone et al.
patent: 5862519 (1999-01-01), Sharma et al.
patent: 5864807 (1999-01-01), Campbell et al.
patent: 5895447 (1999-04-01), Ittycheriah et al.
patent: 5918223 (1999-06-01), Blum et al.
patent: 5930803 (1999-07-01), Becker et al.
patent: 5946654 (1999-08-01), Newman et al.
patent: 6006184 (1999-12-01), Yamada et al.
patent: 6058205 (2000-05-01), Bahl et al.
patent: 6138115 (2000-10-01), Agrawal et al.
Farrell et al., “Speaker Identification Using Neural Tree Networks”, 1994 IEEE Conf. on Acoustics, Speech and Signal Processing 1/165-1/168, Apr. 1994.*
Farrell et al., “Neural Tree Network/Vector Quantization Probability Estimators for Speaker Recognition”, Proceedings of the 1994 IEEE Workshop on Neural Networks for Signal Processing, 279-88, Sep. 1994.*
Mammone et al., “Speaker Verification Using Phoneme-Based Neural Tree Networks and Phonetic Weighting Scoring Method”, Proceedings of the 1995 IEEE Workshop on Neural Networks for Signal Processing 213-22, Aug. 1995.*
Farrell, “Discriminatory Measures for Speaker Recognition”, Proceedings of the 1995 IEEE Workshop on Neural Networks for Signal Processing, 143-52, Aug. 1995.*
Mammone et al., “A Subword Neural Tree Network Approach to Text-Dependent Speaker Verification”, 1995 International Conf. on Acoustics, Speech, and Signal Processing, 357-60, May, 1995.*
Farrell, “Text-Dependent Speaker Verification Using Data Fusion”, 1995 International Conf. on Acoustics, Speech, and Signal Processing, 349-52, May, 1995.*
M. Sharma et al, “Subword-Based Text-Dependent Speaker Verification System with User-Selectable Passwords”, 1996 IEEE International Conf. on Acoustics, Speech, and Signal Processing, 93-96, May, 1996.*
Farrell et al., “Sub-Word Speaker Verification Using Data Fusion Methods”, Proceedings of the 1997 IEEE Workshop on Neural Networks for Signal Processing, 531-40, Apr. 1994.*
Firesmith et al., Dictionary of Object Technology: The Definitive Desk Reference 57, 89, 90, 184, 1995.*
Merriam Webster's Collegiate Dictionary, Tenth Edition 1129, 1997.*
An article entitled “Speaker Recognition Using Neural Networks And Conventional Classifiers”, by Kevin R. Farrell, Richard J. Mammone, and Khaled T. Assaleh, from IEEE 1994.
Assaleh Khaled
Campbell William Michael
Chaki Kakali
Gorrie Gregory J.
Jones Maurice J.
Motorola Inc.
Whitney Sherry J.
LandOfFree
Access control system and method therefor does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Access control system and method therefor, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Access control system and method therefor will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2525187