METHOD AND APPARATUS FOR RECOGNIZING IMAGE PATTERN, METHOD AND APPARATUS FOR JUDGING IDENTITY OF IMAGE PATTERNS, RECORDING MEDIUM FOR RECORDING THE PATTERN RECOGNIZING METHOD AND RECORDING MEDIUM FOR RECORDING THE PATTERN IDENTITY JUDGING METHOD

• US 6,628,811 B1
• Filed: 03/18/1999
• Issued: 09/30/2003
• Est. Priority Date: 03/19/1998
• Status: Expired due to Fees

• Current Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Original Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Inventors: Soma, Masaki, Nagao, Kenji
• Primary Examiner(s): WERNER, BRIAN P
• Application Number:
  US09/271,387
• Time in Patent Office: 1,657 Days
• Field of Search: 382/115, 382/118, 382/116, 382/206, 382/190, 382/224, 382/228, 382/276, 356/71, 340/5.53, 340/5.83, 713/186, 713/200, 707/6, 902/3, 902/6, 706/20
• US Class Current: 382/118
• CPC Class Codes: G06F 18/2135 based on approximation criteria, e.g. principal component analysis
• CPC Class Codes: G06V 10/7715 Feature extraction, e.g. by transforming the feature space, e.g. multi-dimensional scaling [MDS]; Mappings, e.g. subspace methods
• CPC Class Codes: G06V 40/169 Holistic features and representations, i.e. based on the facial image taken as a whole

Abstract

A plurality of teaching identification face image patterns of teaching persons are obtained in a first pattern obtaining process using an image scanner, and a plurality of teaching video face image patterns of the teaching persons are obtained in a second pattern obtaining process using a video camera. A feature extraction matrix, which minimizes an overlapping area between a pattern distribution of the teaching identification face image patterns and a perturbation distribution between a group of teaching identification face image patterns and a group of teaching video face image patterns, is calculated. In cases where a feature extraction using the feature extraction matrix is performed for referential face image patterns of registered persons obtained in the first pattern obtaining process, referential feature pattern vectors independent of any pattern obtaining process are obtained. When an input face image pattern of a specific person obtained in the second pattern obtaining process is received, the feature extraction is performed for the input face image pattern to obtain an input feature pattern vector independent of any pattern obtaining process. Therefore, the specific person can be recognized as a specific registered person by selecting a specific referential feature pattern vector most similar to the input feature pattern vector.

Claims

1. A pattern recognizing method, comprising the steps of:

obtaining a set of first teaching patterns of a plurality of teaching samples according to a first pattern obtaining process;

obtaining a set of second teaching patterns of the teaching samples according to a second pattern obtaining process differing from the first pattern obtaining process;

calculating a teaching pattern distribution from the set of first teaching patterns or the set of second teaching patterns;

calculating a teaching distribution of a perturbation between the set of first teaching patterns and the set of second teaching patterns;

calculating a feature extraction matrix, which minimizes an overlapping area between the teaching pattern distribution and the teaching perturbation distribution, from the teaching pattern distribution and the teaching perturbation distribution;

obtaining a set of referential patterns of a plurality of referential samples according to the first pattern obtaining process;

calculating a set of referential feature patterns of the referential samples from the set of referential patterns according to the feature extraction matrix, the set of referential feature patterns being independent of the first pattern obtaining process and the second pattern obtaining process;

receiving an input pattern of an input sample according to the second pattern obtaining process;

calculating an input feature pattern of the input sample from the input pattern according to the feature extraction matrix;

selecting a specific referential feature pattern most similar to the input feature pattern from the set of referential feature patterns; and

recognizing a specific referential sample corresponding to the specific referential feature pattern as the input sample.

2. A pattern recognizing method according to claim 1 in which the step of calculating a teaching pattern distribution comprises the steps of

calculating a teaching pattern covariance matrix of a pattern sample space from the set of first teaching patterns or the set of second teaching patterns; and

assuming the teaching pattern covariance matrix as the teaching pattern distribution,

the step of calculating a teaching distribution of a perturbation comprises the steps of

calculating a teaching pattern perturbation between one first teaching pattern of one teaching sample and one second teaching pattern of the teaching sample for each teaching sample;

calculating a teaching perturbation covariance matrix from the teaching pattern perturbations of the teaching samples; and

assuming the teaching perturbation covariance matrix as the teaching perturbation distribution, and

the step of calculating a feature extraction matrix comprises the steps of

calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix and the teaching perturbation covariance matrix, from the teaching pattern covariance matrix and the teaching perturbation covariance matrix;

diagonalizing the teaching pattern covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching pattern covariance matrix;

diagonalizing the teaching perturbation covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching perturbation covariance matrix;

calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix to be diagonalized, from the diagonal matrices; and

calculating the feature extraction matrix from the both-diagonalizing matrix and the amplitude re-transformation matrix.

3. A pattern recognizing method according to claim 2 in which the teaching pattern covariance matrix Cs is expressed in an equation $\mathrm{Cs}=\left(1/N\right)$

∑

i=1N

 



(bi-m)

(bi-m)T

where bi denotes one first teaching pattern or one second teaching pattern of an i-th teaching sample, N denotes the number of teaching samples, and m denotes an average pattern of the set of first teaching patterns or the set of second teaching pattern,

the teaching pattern perturbation Pi is expressed in an equation

4. A pattern recognizing method according to claim 1 in which the step of calculating a teaching distribution of a perturbation comprises the step of

calculating a teaching perturbation distribution between one first teaching pattern of one teaching sample and one second teaching pattern of the teaching sample for each teaching sample,

the step of calculating a feature extraction matrix comprises the step of

calculating a feature extraction matrix, which minimizes an overlapping area between a teaching pattern distribution of one teaching sample and the teaching perturbation distribution of the teaching sample, from the teaching pattern distribution and the teaching perturbation distribution of the teaching sample for each teaching sample,

the step of obtaining a set of referential patterns comprises the step of

obtaining a set of referential patterns of the teaching samples according to the first pattern obtaining process or the second pattern obtaining process,

the step of calculating a set of referential feature patterns comprises the step of

calculating one referential feature pattern of one teaching sample from one referential pattern of the teaching sample according to the feature extraction matrix of the teaching sample for each teaching sample,

the step of calculating an input feature pattern comprises the step of

calculating an input feature pattern of the input sample from the input pattern according to the feature extraction matrix of one teaching sample for each teaching sample, and

the step of selecting a specific referential feature pattern comprises the steps of

estimating a similarity between one input feature pattern corresponding to one teaching sample and one referential feature pattern of the same teaching sample;

and

selecting a specific referential feature pattern of a specific teaching sample most similar to the input feature pattern corresponding to the teaching sample from the set of referential feature patterns.

5. A pattern recognizing method according to claim 4 in which the step of calculating a teaching pattern distribution comprises the steps of

calculating a teaching pattern covariance matrix of a pattern sample space from the set of first teaching patterns or the set of second teaching patterns;

and

assuming the teaching pattern covariance matrix as the teaching pattern distribution,

the step of calculating a teaching perturbation distribution comprises the steps of

calculating a teaching pattern perturbation between one first teaching pattern of one teaching sample and one second teaching pattern of the teaching sample for each teaching sample;

calculating a teaching perturbation covariance matrix from one teaching pattern perturbation of one teaching sample for each teaching sample; and

assuming the teaching perturbation covariance matrix as the teaching perturbation distribution for each teaching sample, and

the step of calculating a feature extraction matrix comprises the steps of

calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix and the teaching perturbation covariance matrix of one teaching sample, from the teaching pattern covariance matrix and the teaching perturbation covariance matrix for each teaching sample;

diagonalizing the teaching pattern covariance matrix according to the both-diagonalizing matrix of each teaching sample to produce a diagonal matrix of the teaching pattern covariance matrix corresponding to one teaching sample for each teaching sample;

diagonalizing the teaching perturbation covariance matrix of one teaching sample according to the both-diagonalizing matrix of the teaching sample for each teaching sample to produce a diagonal matrix of the teaching perturbation covariance matrix for each teaching sample;

calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix to be diagonalized, from the diagonal matrices corresponding to one teaching sample for each teaching sample; and

calculating the feature extraction matrix corresponding to one teaching sample from the both-diagonalizing matrix corresponding to the teaching sample and the amplitude re-transformation matrix corresponding to the teaching sample for each teaching sample.

6. A pattern recognizing method according to claim 5 in which the teaching pattern covariance matrix Cs is expressed in an equation $\mathrm{Cs}=\left(1/N\right)$

∑

i=1N

 



(bi-m)

(bi-m)T

where bi denotes one first teaching pattern or one second teaching pattern of an i-th teaching sample, N denotes the number of teaching samples, and m denotes an average pattern of the set of first teaching patterns or the set of second teaching pattern,

the teaching pattern perturbation Pi is expressed in an equation

7. A pattern recognizing method according to claim 1 in which the step of obtaining a set of second teaching patterns comprises the step of

obtaining a group of second teaching patterns for each teaching sample,

the step of calculating a teaching distribution of a perturbation comprises the step of

calculating a teaching perturbation distribution between one first teaching pattern of one teaching sample and one group of second teaching patterns of the teaching sample for each teaching sample,

the step of calculating a feature extraction matrix comprises the step of

calculating a feature extraction matrix, which minimizes an overlapping area between a teaching pattern distribution of one teaching sample and the teaching perturbation distribution of the teaching sample, from the teaching pattern distribution and the teaching perturbation distribution of the teaching sample for each teaching sample,

the step of obtaining a set of referential patterns comprises the step of

obtaining a set of referential patterns of the teaching sample according to the first pattern obtaining process,

the step of receiving an input pattern comprises the step of

obtaining a group of input patterns of the input sample according to the second pattern obtaining process,

the step of calculating an input feature pattern comprises the steps of

calculating a group of input feature patterns of the input sample from the group of input patterns according to the feature extraction matrix of one teaching sample for each teaching sample;

and

calculating an average input feature pattern from the group of input feature patterns for each teaching sample, and

the step of selecting a specific referential feature pattern comprises the steps of

estimating a similarity between one average input feature pattern corresponding to one teaching sample and one average referential feature pattern of the same teaching sample; and

selecting a specific average referential feature pattern of a specific teaching sample most similar to the average input feature pattern corresponding to the teaching sample from the set of average referential feature patterns.

8. A pattern recognizing method according to claim 7 in which the step of calculating a teaching pattern distribution comprises the steps of

calculating a teaching pattern covariance matrix of a pattern sample space from the set of first teaching patterns or the set of second teaching patterns;

and

assuming the teaching pattern covariance matrix as the teaching pattern distribution,

the step of calculating a teaching perturbation distribution comprises the steps of

calculating a group of teaching pattern perturbations between one first teaching pattern of one teaching sample and one group of second teaching patterns of the teaching sample for each teaching sample;

calculating a teaching perturbation covariance matrix from one group of teaching pattern perturbations of one teaching sample for each teaching sample; and

assuming the teaching perturbation covariance matrix as the teaching perturbation distribution for each teaching sample, and

the step of calculating a feature extraction matrix comprises the steps of

calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix and the teaching perturbation covariance matrix of one teaching sample, from the teaching pattern covariance matrix and the teaching perturbation covariance matrix for each teaching sample;

diagonalizing the teaching pattern covariance matrix according to the both-diagonalizing matrix of each teaching sample to produce a diagonal matrix of the teaching pattern covariance matrix corresponding to one teaching sample for each teaching sample;

diagonalizing the teaching perturbation covariance matrix of one teaching sample according to the both-diagonalizing matrix of the teaching sample for each teaching sample to produce a diagonal matrix of the teaching perturbation covariance matrix for each teaching sample;

calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix to be diagonalized, from the diagonal matrices corresponding to one teaching sample for each teaching sample; and

calculating the feature extraction matrix corresponding to one teaching sample from the both-diagonalizing matrix corresponding to the teaching sample and the amplitude re-transformation matrix corresponding to the teaching sample for each teaching sample.

9. A pattern recognizing method according to claim 8 in which the teaching pattern covariance matrix Cs is expressed in an equation $\mathrm{Cs}=\left(1/N\right)$

∑

i=1N

 



(bi-m)

(bi-m)T

where bi denotes one first teaching pattern or one second teaching pattern of an i-th teaching sample, N denotes the number of teaching samples, and m denotes an average pattern of the set of first teaching patterns or the set of second teaching pattern,

each teaching pattern perturbation Pji is expressed in an equation

10. A pattern recognizing apparatus, comprising:

first pattern obtaining means for obtaining a set of first teaching patterns of a plurality of teaching samples according to a first pattern obtaining process;

second pattern obtaining means for obtaining a set of second teaching patterns of the teaching samples according to a second pattern obtaining process differing from the first pattern obtaining process;

feature extracting means for calculating a teaching pattern distribution from the set of first teaching patterns obtained by the first pattern obtaining means or the set of second teaching patterns obtained by the second pattern obtaining means, calculating a teaching distribution of a perturbation between the set of first teaching patterns and the set of second teaching patterns, and calculating a feature extraction matrix, which minimizes an overlapping area between the teaching pattern distribution and the teaching perturbation distribution, from the teaching pattern distribution and the teaching perturbation distribution;

referential feature pattern calculating means for obtaining a set of referential patterns of a plurality of referential samples according to the first pattern obtaining process, and calculating a set of referential feature patterns of the referential samples from the set of referential patterns according to the feature extraction matrix calculated by the feature extracting means to make the set of referential feature patterns independent of the first pattern obtaining process and the second pattern obtaining process; and

input pattern recognizing means for receiving an input pattern of an input sample according to the second pattern obtaining process, calculating an input feature pattern of the input sample from the input pattern according to the feature extraction matrix calculated by the feature extracting means, selecting a specific referential feature pattern most similar to the input feature pattern from the set of referential feature patterns calculated by the referential feature pattern calculating means, and recognizing a specific referential sample corresponding to the specific referential feature pattern as the input sample.

11. A pattern recognizing apparatus according to claim 10 in which the feature extracting means comprises

pattern covariance assuming means for calculating a teaching pattern covariance matrix of a pattern sample space from the first teaching patterns or the second teaching patterns and assuming the teaching pattern covariance matrix as the teaching pattern distribution;

pattern perturbation calculating means for calculating a teaching pattern perturbation between one first teaching pattern of one teaching sample and one second teaching pattern of the teaching sample for each teaching sample;

perturbation covariance assuming means for assuming a teaching perturbation covariance matrix from the teaching pattern perturbations calculated by the pattern perturbation calculating means as the teaching perturbation distribution;

both-diagonalizing matrix calculating means for calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix assumed by the pattern covariance assuming means and the teaching perturbation covariance matrix assumed by the perturbation covariance assuming means, from the teaching pattern covariance matrix and the teaching perturbation covariance matrix;

diagonal matrix producing means for diagonalizing the teaching pattern covariance matrix assumed by the pattern covariance assuming means according to the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means to produce a diagonal matrix of the teaching pattern covariance matrix, and diagonalizing the teaching perturbation covariance matrix assumed by the perturbation covariance assuming means according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching perturbation covariance matrix;

amplitude re-transformation matrix calculating means for calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to be calculated by the referential feature pattern calculating means to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means to be diagonalized, from the diagonal matrices; and

calculating the feature extraction matrix from the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means and the amplitude re-transformation matrix calculated by the amplitude re-transformation matrix calculating means.

12. A pattern recognizing apparatus, comprising:

first pattern obtaining means for obtaining a set of first teaching patterns of a plurality of registered samples according to a first pattern obtaining process;

second pattern obtaining means for obtaining a set of second teaching patterns of the registered samples according to a second pattern obtaining process differing from the first pattern obtaining process;

feature extracting means for calculating a teaching pattern distribution from the first teaching patterns obtained by the first pattern obtaining means or the second teaching patterns obtained by the second pattern obtaining means, calculating a teaching distribution of a perturbation between one first teaching pattern of one registered sample and one second teaching pattern of the registered sample for each registered sample, and calculating a feature extraction matrix, which minimizes an overlapping area between the teaching pattern distribution of one registered sample and the teaching perturbation distribution of the registered sample, from the teaching pattern distribution and the teaching perturbation distribution for each registered sample;

referential feature pattern calculating means for obtaining a set of referential patterns of the registered samples according to the first pattern obtaining process, and calculating a referential feature pattern of one registered sample from one referential pattern of the registered sample according to the feature extraction matrix of the registered sample calculated by the feature extracting means for each registered sample to make each referential feature pattern independent of the first pattern obtaining process and the second pattern obtaining process; and

input pattern recognizing means for receiving an input pattern of an input sample according to the second pattern obtaining process, calculating an input feature pattern corresponding to one registered sample from the input pattern according to the feature extraction matrix of the registered sample calculated by the feature extracting means for each registered sample, estimating a similarity between one referential feature pattern of one registered sample and the input feature pattern corresponding to the registered sample for each registered sample, selecting a specific referential feature pattern most similar to the input feature pattern from the referential feature patterns calculated by the referential feature pattern calculating means, and recognizing a specific registered sample corresponding to the specific referential feature pattern as the input sample.

13. A pattern recognizing apparatus according to claim 12 in which the feature extracting means comprises

pattern covariance assuming means for calculating a teaching pattern covariance matrix of a pattern sample space from the first teaching patterns or the second teaching patterns and assuming the teaching pattern covariance matrix as the teaching pattern distribution;

pattern perturbation calculating means for calculating a teaching pattern perturbation between one first teaching pattern of one registered sample and one second teaching pattern of the registered sample for each registered sample;

perturbation covariance assuming means for calculating a teaching perturbation covariance matrix from one teaching pattern perturbation calculated by the pattern perturbation calculating means for each registered person and assuming each teaching perturbation covariance matrix as the teaching perturbation distribution;

both-diagonalizing matrix calculating means for calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix assumed by the pattern covariance assuming means and the teaching perturbation covariance matrix assumed by the perturbation covariance assuming means, from the teaching pattern covariance matrix and the teaching perturbation covariance matrix for each registered person;

diagonal matrix producing means for diagonalizing the teaching pattern covariance matrix assumed by the pattern covariance assuming means according to the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means for each registered person to produce a diagonal matrix of the teaching pattern covariance matrix for each registered person, and diagonalizing the teaching perturbation covariance matrix assumed by the perturbation covariance assuming means according to the both-diagonalizing matrix for each registered person to produce a diagonal matrix of the teaching perturbation covariance matrix for each registered person;

amplitude re-transformation matrix calculating means for calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to be calculated by the referential feature pattern calculating means to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means to be diagonalized, from the diagonal matrices corresponding one registered sample for each registered person; and

calculating the feature extraction matrix from the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means and the amplitude re-transformation matrix calculated by the amplitude re-transformation matrix calculating means for each registered person.

14. A pattern identity judging method, comprising the steps of:

obtaining a set of first teaching patterns from a plurality of teaching samples according to a first pattern obtaining process;

obtaining a set of second teaching patterns from the teaching samples according to a second pattern obtaining process differing from the first pattern obtaining process;

calculating a teaching pattern distribution from the set of first teaching patterns or the set of second teaching patterns;

calculating a teaching distribution of a perturbation between the set of first teaching patterns and the set of second teaching patterns;

calculating a feature extraction matrix, which minimizes an overlapping area between the teaching pattern distribution and the teaching perturbation distribution, from the teaching pattern distribution and the teaching perturbation distribution;

receiving a first input pattern of a first input sample according to the first pattern obtaining process;

calculating a first input feature pattern of the first input sample from the first input pattern according to the feature extraction matrix, the first input feature pattern being independent of the first pattern obtaining process and the second pattern obtaining process;

receiving a second input pattern of a second input sample according to the second pattern obtaining process;

calculating a second input feature pattern of the second input sample from the second input pattern according to the feature extraction matrix, the second input feature pattern being independent of the first pattern obtaining process and the second pattern obtaining process;

collating the first input feature pattern with the second input feature pattern to estimate a similarity between the first input sample and the second input sample; and

judging that the first input sample is identical with the second input sample in cases where the similarity is high.

15. A pattern identity judging method according to claim 14 in which the step of calculating a teaching pattern distribution comprises the step of

assuming a teaching pattern covariance matrix of a pattern sample space as the teaching pattern distribution,

the step of calculating a teaching distribution of a perturbation comprises the steps of

calculating a teaching pattern perturbation between one first teaching pattern of one teaching sample and one second teaching pattern of the teaching sample for each teaching sample; and

assuming a teaching perturbation covariance matrix from the teaching pattern perturbations as the teaching perturbation distribution, and

the step of calculating a feature extraction matrix comprises the steps of

calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix and the teaching perturbation covariance matrix, from the teaching pattern covariance matrix and the teaching perturbation covariance matrix;

diagonalizing the teaching pattern covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching pattern covariance matrix;

diagonalizing the teaching perturbation covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching perturbation covariance matrix;

calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix to be diagonalized, from the diagonal matrices; and

calculating the feature extraction matrix from the both-diagonalizing matrix and the amplitude re-transformation matrix.

16. A pattern identity judging method according to claim 15 in which

the teaching pattern covariance matrix Cs is expressed in an equation $\mathrm{Cs}=\left(1/N\right)$

∑

i=1N

 



(bi-m)

(bi-m)T

where bi denotes one first teaching pattern or one second teaching pattern of an i-th teaching sample, N denotes the number of teaching samples, and m denotes an average pattern of the set of first teaching patterns or the set of second teaching pattern,

the teaching pattern perturbation Pi is expressed in an equation

17. A pattern identity judging method according to claim 14 in which

the step of obtaining a set of second teaching patterns comprises the step of

obtaining a group of second teaching patterns for each teaching sample,

the step of calculating a teaching pattern distribution comprises the steps of

calculating a teaching pattern covariance matrix of a pattern sample space from the set of first teaching patterns or the set of second teaching patterns;

and

assuming the teaching pattern covariance matrix as the teaching pattern distribution,

the step of calculating a teaching distribution of a perturbation comprises the steps of

calculating a group of teaching pattern perturbations between one first teaching pattern of one teaching sample and one group of second teaching patterns of the teaching sample for each teaching sample;

calculating a teaching perturbation covariance matrix from one group of teaching pattern perturbations of one teaching sample for each teaching sample;

calculating an average teaching perturbation covariance matrix from the teaching perturbation covariance matrices; and

assuming the average teaching perturbation covariance matrix as the teaching perturbation distribution,

the step of calculating a feature extraction matrix comprises the steps of

calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix and the teaching perturbation covariance matrix, from the teaching pattern covariance matrix and the teaching perturbation covariance matrix;

diagonalizing the teaching pattern covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching pattern covariance matrix;

diagonalizing the teaching perturbation covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching perturbation covariance matrix;

calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix to be diagonalized, from the diagonal matrices; and

calculating the feature extraction matrix from the both-diagonalizing matrix and the amplitude re-transformation matrix,

the step of receiving a second input pattern comprises the step of

obtaining a group of second input patterns of the second input sample, and

the step of calculating a second input feature pattern comprises the steps of

calculating an average of the group of second input patterns as a second input pattern newly defined; and

calculating a second input feature pattern of the second input sample from the second input pattern newly defined.

18. A pattern identity judging method according to claim 17 in which

the teaching pattern covariance matrix Cs is expressed in an equation $\mathrm{Cs}=\left(1/N\right)$

∑

i=1N

 



(bi-m)

(bi-m)T

where bi denotes one first teaching pattern of an i-th teaching sample, N denotes the number of teaching samples, and m denotes an average pattern of the set of first teaching patterns,

each teaching pattern perturbation Pji is expressed in an equation

19. A pattern identity judging apparatus, comprising:

first pattern obtaining means for obtaining a set of first teaching patterns of a plurality of teaching samples according to a first pattern obtaining process;

second pattern obtaining means for obtaining a set of second teaching patterns of the teaching samples according to a second pattern obtaining process differing from the first pattern obtaining process;

feature extracting means for calculating a teaching pattern distribution from the set of first teaching patterns obtained by the first pattern obtaining means or the set of second teaching patterns obtained by the second pattern obtaining means, calculating a teaching distribution of a perturbation between the set of first teaching patterns and the set of second teaching patterns, and calculating a feature extraction matrix, which minimizes an overlapping area between the teaching pattern distribution and the teaching perturbation distribution, from the teaching pattern distribution and the teaching perturbation distribution;

feature pattern calculating means for receiving a first input pattern of a first input sample according to the first pattern obtaining process, receiving a second input pattern of a second input sample according to the second pattern obtaining process, calculating a first input feature pattern of the first input sample from the first input pattern according to the feature extraction matrix calculated by the feature extracting means to make the first input feature pattern independent of the first pattern obtaining process and the second pattern obtaining process, and calculating a second input feature pattern of the second input sample from the second input pattern according to the feature extraction matrix to make the second input feature pattern independent of the first pattern obtaining process and the second pattern obtaining process; and

identity judging means for collating the first input feature pattern calculated by the feature pattern calculating means with the second input feature pattern calculated by the feature pattern calculating means to estimate a similarity between the first input sample and the second input sample, and judging that the first input sample is identical with the second input sample in cases where the similarity is high.

20. A pattern identity judging apparatus according to claim 19 in which the feature extracting means comprises

pattern covariance assuming means for calculating a teaching pattern covariance matrix of a pattern sample space from the set of first teaching patterns or the set of second teaching patterns and assuming the teaching pattern covariance matrix as the teaching pattern distribution;

pattern perturbation calculating means for calculating a teaching pattern perturbation between one first teaching pattern of one teaching sample and one second teaching pattern of the teaching sample for each teaching sample;

perturbation covariance assuming means for assuming a teaching perturbation covariance matrix from the teaching pattern perturbations calculated by the pattern perturbation calculating means as the teaching perturbation distribution;

both-diagonalizing matrix calculating means for calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix assumed by the pattern covariance assuming means and the teaching perturbation covariance matrix assumed by the perturbation covariance assuming means, from the teaching pattern covariance matrix and the teaching perturbation covariance matrix;

diagonal matrix producing means for diagonalizing the teaching pattern covariance matrix assumed by the pattern covariance assuming means according to the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means to produce a diagonal matrix of the teaching pattern covariance matrix, and diagonalizing the teaching perturbation covariance matrix assumed by the perturbation covariance assuming means according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching perturbation covariance matrix;

amplitude re-transformation matrix calculating means for calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to be calculated by the referential feature pattern calculating means to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means to be diagonalized, from the diagonal matrices; and

calculating the feature extraction matrix from the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means and the amplitude re-transformation matrix calculated by the amplitude re-transformation matrix calculating means.

21. A pattern identity judging apparatus, comprising:

first pattern obtaining means for obtaining a set of first teaching patterns of a plurality of teaching samples according to a first pattern obtaining process;

second pattern obtaining means for obtaining a group of second teaching patterns according to a second pattern obtaining process differing from the first pattern obtaining process for each teaching sample;

feature extracting means for calculating a teaching pattern distribution from the set of first teaching patterns obtained by the first pattern obtaining means or the group of second teaching patterns obtained by the second pattern obtaining means comprising;

means for calculating a teaching distribution of a perturbation between one first teaching pattern of one teaching sample and the group of second teaching patterns of the teaching sample for each teaching sample,

means for calculating an average teaching perturbation distribution from the teaching perturbation distributions, and

means for calculating a feature extraction matrix, which minimizes an overlapping area between the teaching pattern distribution and the average teaching perturbation distribution, from the teaching pattern distribution and the average teaching perturbation distribution;

feature pattern calculating means comprising;

means for receiving a first input pattern of a first input sample according to the first pattern obtaining process,

means for receiving a second input pattern of a second input sample according to the second pattern obtaining process,

means for calculating a first input feature pattern of the first input sample from the first input pattern according to the feature extraction matrix calculated by the feature extracting means to make the first input feature pattern independent of the first pattern obtaining process and the second pattern obtaining process, and

means for calculating a second input feature pattern of the second input sample from the second input pattern according to the feature extraction matrix to make the second input feature pattern independent of the first pattern obtaining process and the second pattern obtaining process; and

identity judging means for collating the first input feature pattern calculated by the feature pattern calculating means with the second input feature pattern calculated by the feature pattern calculating means to estimate a similarity between the first input sample and the second input sample, and judging that the first input sample is identical with the second input sample in cases where the similarity is high.

22. A pattern identity judging apparatus according to claim 21 in which the feature extracting means comprises

pattern covariance assuming means for calculating a teaching pattern covariance matrix of a pattern sample space from the set of first teaching patterns or the groups of second teaching patterns and assuming the teaching pattern covariance matrix as the teaching pattern distribution;

pattern perturbation calculating means for calculating a group of teaching pattern perturbations between one first teaching pattern of one teaching sample and one group of second teaching patterns of the teaching sample for each teaching sample;

perturbation covariance assuming means for calculating a teaching perturbation covariance matrix from one group of teaching pattern perturbations calculated by the pattern perturbation calculating means for each teaching sample, calculating an average teaching perturbation covariance matrix from the teaching perturbation covariance matrices and assuming the average teaching perturbation covariance matrix as the average teaching perturbation distribution;

both-diagonalizing matrix calculating means for calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix assumed by the pattern covariance assuming means and the average teaching perturbation covariance matrix assumed by the perturbation covariance assuming means, from the teaching pattern covariance matrix and the average teaching perturbation covariance matrix;

diagonal matrix producing means for diagonalizing the teaching pattern covariance matrix assumed by the pattern covariance assuming means according to the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means to produce a diagonal matrix of the teaching pattern covariance matrix, and diagonalizing the average teaching perturbation covariance matrix assumed by the perturbation covariance assuming means according to the both-diagonalizing matrix to produce a diagonal matrix of the average teaching perturbation covariance matrix;

amplitude re-transformation matrix calculating means for calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to be calculated by the referential feature pattern calculating means to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means to be diagonalized, from the diagonal matrices; and

calculating the feature extraction matrix from the both-diagonalizing matrix calculated by the both-diagonalizing matrix calculating means and the amplitude re-transformation matrix calculated by the amplitude re-transformation matrix calculating means.

23. A recording medium for recording a software program of a pattern recognizing method executed in a computer, the pattern recognizing method, comprising the steps of:

obtaining a set of first teaching patterns of a plurality of teaching samples according to a first pattern obtaining process;

obtaining a set of second teaching patterns of the teaching samples according to a second pattern obtaining process differing from the first pattern obtaining process;

calculating a teaching pattern distribution from the set of first teaching patterns or the set of second teaching patterns;

calculating a teaching distribution of a perturbation between the set of first teaching patterns and the set of second teaching patterns;

calculating a feature extraction matrix, which minimizes an overlapping area between the teaching pattern distribution and the teaching perturbation distribution, from the teaching pattern distribution and the teaching perturbation distribution;

obtaining a set of referential patterns of a plurality of referential samples according to the first pattern obtaining process;

calculating a set of referential feature patterns of the referential samples from the set of referential patterns according to the feature extraction matrix, the set of referential feature patterns being independent of the first pattern obtaining process and the second pattern obtaining process;

receiving an input pattern of an input sample according to the second pattern obtaining process;

calculating an input feature pattern of the input sample from the input pattern according to the feature extraction matrix;

selecting a specific referential feature pattern most similar to the input feature pattern from the set of referential feature patterns; and

recognizing a specific referential sample corresponding to the specific referential feature pattern as the input sample.

24. A recording medium according to claim 23 in which the step of calculating a teaching pattern distribution comprises the steps of

calculating a teaching pattern covariance matrix of a pattern sample space from the set of first teaching patterns or the set of second teaching patterns; and

assuming the teaching pattern covariance matrix as the teaching pattern distribution,

the step of calculating a teaching distribution of a perturbation comprises the steps of

calculating a teaching pattern perturbation between one first teaching pattern of one teaching sample and one second teaching pattern of the teaching sample for each teaching sample;

calculating a teaching perturbation covariance matrix from the teaching pattern perturbations of the teaching samples; and

assuming the teaching perturbation covariance matrix as the teaching perturbation distribution, and

the step of calculating a feature extraction matrix comprises the steps of

calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix and the teaching perturbation covariance matrix, from the teaching pattern covariance matrix and the teaching perturbation covariance matrix;

diagonalizing the teaching pattern covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching pattern covariance matrix;

diagonalizing the teaching perturbation covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching perturbation covariance matrix;

calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix to be diagonalized, from the diagonal matrices; and

calculating the feature extraction matrix from the both-diagonalizing matrix and the amplitude re-transformation matrix.

25. A recording medium according to claim 23 in which the step of calculating a teaching distribution of a perturbation comprises the step of

calculating a teaching perturbation distribution between one first teaching pattern of one teaching sample and one second teaching pattern of the teaching sample for each teaching sample,

the step of calculating a feature extraction matrix comprises the step of

calculating a feature extraction matrix, which minimizes an overlapping area between a teaching pattern distribution of one teaching sample and the teaching perturbation distribution of the teaching sample, from the teaching pattern distribution and the teaching perturbation distribution of the teaching sample for each teaching sample,

the step of obtaining a set of referential patterns comprises the step of

obtaining a set of referential patterns of the teaching samples according to the first pattern obtaining process or the second pattern obtaining process,

the step of calculating a set of referential feature patterns comprises the step of

calculating one referential feature pattern of one teaching sample from one referential pattern of the teaching sample according to the feature extraction matrix of the teaching sample for each teaching sample,

the step of calculating an input feature pattern comprises the step of

calculating an input feature pattern of the input sample from the input pattern according to the feature extraction matrix of one teaching sample for each teaching sample, and

the step of selecting a specific referential feature pattern comprises the steps of

estimating a similarity between one input feature pattern corresponding to one teaching sample and one referential feature pattern of the same teaching sample;

and

selecting a specific referential feature pattern of a specific teaching sample most similar to the input feature pattern corresponding to the teaching sample from the set of referential feature patterns.

26. A recording medium according to claim 23 in which the step of obtaining a set of second teaching patterns comprises the step of

obtaining a group of second teaching patterns for each teaching sample,

the step of calculating a teaching distribution of a perturbation comprises the step of

calculating a teaching perturbation distribution between one first teaching pattern of one teaching sample and one group of second teaching patterns of the teaching sample for each teaching sample,

the step of calculating a feature extraction matrix comprises the step of

calculating a feature extraction matrix, which minimizes an overlapping area between a teaching pattern distribution of one teaching sample and the teaching perturbation distribution of the teaching sample, from the teaching pattern distribution and the teaching perturbation distribution of the teaching sample for each teaching sample,

the step of obtaining a set of referential patterns comprises the step of

obtaining a set of referential patterns of the teaching sample according to the first pattern obtaining process,

the step of receiving an input pattern comprises the step of

obtaining a group of input patterns of the input sample according to the second pattern obtaining process,

the step of calculating an input feature pattern comprises the steps of

calculating a group of input feature patterns of the input sample from the group of input patterns according to the feature extraction matrix of one teaching sample for each teaching sample;

and

calculating an average input feature pattern from the group of input feature patterns for each teaching sample, and

the step of selecting a specific referential feature pattern comprises the steps of

estimating a similarity between one average input feature pattern corresponding to one teaching sample and one average referential feature pattern of the same teaching sample; and

selecting a specific average referential feature pattern of a specific teaching sample most similar to the average input feature pattern corresponding to the teaching sample from the set of average referential feature patterns.

27. A recording medium for recording a software program of a pattern identity judging method executed in a computer, the pattern identity judging method, comprising the steps of:

obtaining a set of first teaching patterns from a plurality of teaching samples according to a first pattern obtaining process;

obtaining a set of second teaching patterns from the teaching samples according to a second pattern obtaining process differing from the first pattern obtaining process;

calculating a teaching pattern distribution from the set of first teaching patterns or the set of second teaching patterns;

calculating a teaching distribution of a perturbation between the set of first teaching patterns and the set of second teaching patterns;

calculating a feature extraction matrix, which minimizes an overlapping area between the teaching pattern distribution and the teaching perturbation distribution, from the teaching pattern distribution and the teaching perturbation distribution;

receiving a first input pattern of a first input sample according to the first pattern obtaining process;

calculating a first input feature pattern of the first input sample from the first input pattern according to the feature extraction matrix, the first input feature pattern being independent of the first pattern obtaining process and the second pattern obtaining process;

receiving a second input pattern of a second input sample according to the second pattern obtaining process;

calculating a second input feature pattern of the second input sample from the second input pattern according to the feature extraction matrix, the second input feature pattern being independent of the first pattern obtaining process and the second pattern obtaining process;

collating the first input feature pattern with the second input feature pattern to estimate a similarity between the first input sample and the second input sample; and

judging that the first input sample is identical with the second input sample in cases where the similarity is high.

28. A recording medium according to claim 27, in which the step of calculating a teaching pattern distribution comprises the step of

assuming a teaching pattern covariance matrix of a pattern sample space as the teaching pattern distribution,

the step of calculating a teaching distribution of a perturbation comprises the steps of

calculating a teaching pattern perturbation between one first teaching pattern of one teaching sample and one second teaching pattern of the teaching sample for each teaching sample; and

assuming a teaching perturbation covariance matrix from the teaching pattern perturbations as the teaching perturbation distribution, and

the step of calculating a feature extraction matrix comprises the steps of

calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix and the teaching perturbation covariance matrix, from the teaching pattern covariance matrix and the teaching perturbation covariance matrix;

diagonalizing the teaching pattern covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching pattern covariance matrix;

diagonalizing the teaching perturbation covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching perturbation covariance matrix;

calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix to be diagonalized, from the diagonal matrices; and

calculating the feature extraction matrix from the both-diagonalizing matrix and the amplitude re-transformation matrix.

29. A recording medium according to claim 27, in which the step of obtaining a set of second teaching patterns comprises the step of

obtaining a group of second teaching patterns for each teaching sample,

the step of calculating a teaching pattern distribution comprises the steps of

calculating a teaching pattern covariance matrix of a pattern sample space from the set of first teaching patterns or the set of second teaching patterns;

and

assuming the teaching pattern covariance matrix as the teaching pattern distribution,

the step of calculating a teaching distribution of a perturbation comprises the steps of

calculating a group of teaching pattern perturbations between one first teaching pattern of one teaching sample and one group of second teaching patterns of the teaching sample for each teaching sample;

calculating a teaching perturbation covariance matrix from one group of teaching pattern perturbations of one teaching sample for each teaching sample;

calculating an average teaching perturbation covariance matrix from the teaching perturbation covariance matrices; and

assuming the average teaching perturbation covariance matrix as the teaching perturbation distribution,

the step of calculating a teaching distribution of a perturbation comprises the steps of

calculating a both-diagonalizing matrix, which diagonalizes both the teaching pattern covariance matrix and the teaching perturbation covariance matrix, from the teaching pattern covariance matrix and the teaching perturbation covariance matrix;

diagonalizing the teaching pattern covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching pattern covariance matrix;

diagonalizing the teaching perturbation covariance matrix according to the both-diagonalizing matrix to produce a diagonal matrix of the teaching perturbation covariance matrix;

calculating an amplitude re-transformation matrix, which again transforms a referential pattern covariance matrix indicated by the set of referential feature patterns to adjust amplitudes of diagonal elements of the referential pattern covariance matrix after the referential pattern covariance matrix is transformed by the both-diagonalizing matrix to be diagonalized, from the diagonal matrices; and

calculating the feature extraction matrix from the both-diagonalizing matrix and the amplitude re-transformation matrix.