Parameter estimation apparatus and data collating apparatus

• US 20040022432A1
• Filed: 05/08/2003
• Published: 02/05/2004
• Est. Priority Date: 09/25/2001
• Status: Active Grant

• Current Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Original Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Inventors: Hayata, Keisuke, Nagao, Kenji
• Granted Patent: US 7,224,823 B2
• Time in Patent Office: Days
• US Class Current: 382/159
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 10/431 Frequency domain transformation; Autocorrelation
• CPC Class Codes: G06V 40/16 Human faces, e.g. facial parts, sketches or expressions

Abstract

The present invention enables estimation of desired parameters with less computation cost and with high precision by inputting first training vectors generated from observation patterns and second training vectors generated from estimation targets in order to learn the correlation between observation patterns as inputs and patterns of the estimation targets such that desired outputs are assumed from the inputs, calculating the auto-correlation information of the two training vectors, and cross-correlation information of an average vector, the first training vectors and second training vectors, and using the information, obtaining probable expectation values based on the Bayes theory of the estimation targets with respect to an input pattern.

Claims

1. A parameter estimation apparatus comprising:

an image input section that converts optical data into electronic data;

a parameter input section that inputs coordinates of a first feature point from first electronic data input from the image input section;

a learning section that calculates auto-correlation information from a plurality of items of first electronic data, calculates cross-correlation information from the plurality of items of first electronic data and coordinates of a plurality of first feature points, calculates a feature extraction matrix for estimating coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputs a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from the coordinates of the plurality of first feature points, and the feature extraction matrix;

a correlation information database that stores the first average vector, the second average vector and the feature extraction matrix input from the learning section; and

a parameter estimating section that estimates coordinates of the second feature point, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix.

2. The parameter estimation apparatus according to claim 1, wherein the learning section comprises:

a first training vector calculating section that calculates a first training vector of vector pattern where the first electronic data is arranged in the order of scanning;

a second training vector calculating section that calculates a second training vector of vector pattern where the coordinates of the first feature points are arranged sequentially;

an average vector calculating section that calculates the first average vector by averaging a plurality of first training vectors and the second average vector by averaging a plurality of second training vectors;

an auto-correlation information calculating section that receives as its inputs the plurality of first training vectors and the first average vector, and calculates a pseudo inverse matrix of a first covariance matrix that is auto-correlation information of the first training vectors;

a cross-correlation information calculating section that receives as its inputs the plurality of first training vectors, the first average vector, the plurality of second training vectors and the second average vector, and calculates a second covariance matrix that is cross-correlation information of the first training vectors and the second training vectors; and

a feature extraction matrix calculating section that receives as its inputs the pseudo inverse matrix of the first covariance matrix and the second covariance matrix, and calculates the feature extraction matrix from the pseudo inverse matrix of the first covariance matrix and the second covariance matrix.

3. The parameter estimation apparatus according to claim 1, wherein the parameter estimating section comprises:

a vector calculating section that receives as its input the second electronic data, and calculates an input vector of vector pattern where the second electronic data is arranged in the order of scanning;

a parameter calculating section that calculates an expectation vector including the coordinates of the second feature point from the input vector, the first average vector, the second average vector and the feature extraction matrix; and

a template matching section that calculates the coordinates of the second feature point from the expectation vector.

4. A parameter estimation apparatus comprising:

an image input section that converts optical data into electronic data;

a parameter input section that inputs an image around coordinates of a first feature point from first electronic data input from the image input section;

a learning section that calculates auto-correlation information from a plurality of items of first electronic data, calculates cross-correlation information from the plurality of items of first electronic data and images around coordinates of a plurality of first feature points, calculates a feature extraction matrix for estimating an image around coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputs a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from the images around the coordinates of the plurality of first feature points, and the feature extraction matrix;

a correlation information database that stores the first average vector, the second average vector and the feature extraction matrix input from the learning section; and

a parameter estimating section that estimates coordinates of the second feature point, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix.

5. The parameter estimation apparatus according to claim 4, wherein the learning section comprises:

a first training vector calculating section that calculates a first training vector of vector pattern where the first electronic data is arranged in the order of scanning;

a second training vector calculating section that calculates a second training vector of vector pattern where the images around the coordinates of the first feature points are arranged sequentially;

an average vector calculating section that calculates the first average vector by averaging a plurality of first training vectors and the second average vector by averaging a plurality of second training vectors;

an auto-correlation information calculating section that receives as its inputs the plurality of first training vectors and the first average vector, and calculates a pseudo inverse matrix of a first covariance matrix that is auto-correlation information of the first training vectors;

a cross-correlation information calculating section that receives as its inputs the plurality of first training vectors, the first average vector, the plurality of second training vectors and the second average vector, and calculates a second covariance matrix that is cross-correlation information of the first training vectors and the second training vectors; and

a feature extraction matrix calculating section that receives as its inputs the pseudo inverse matrix of the first covariance matrix and the second covariance matrix, and calculates the feature extraction matrix from the pseudo inverse matrix of the first covariance matrix and the second covariance matrix.

6. The parameter estimation apparatus according to claim 4, wherein the parameter estimating section comprises:

a vector calculating section that receives as its input the second electronic data, and calculates an input vector of vector pattern where the second electronic data is arranged in the order of scanning;

a parameter calculating section that calculates an expectation vector including the image around the coordinates of the second feature point from the input vector, the first average vector, the second average vector and the feature extraction matrix; and

a template matching section that calculates the image around the coordinates of the second feature point from the expectation vector, searches for a matching region that is an image region of the second electronic data matching the image around the coordinates of the second feature point, and calculates the coordinates of the second feature point from the matching region.

7. A parameter estimation apparatus comprising:

an image input section that converts optical data into electronic data;

a parameter input section that inputs coordinates of a first feature point and an image around the coordinates of the first feature point from first electronic data input from the image input section;

a learning section that calculates auto-correlation information from a plurality of items of first electronic data, calculates cross-correlation information from the plurality of items of first electronic data and a combined vector of coordinates of a plurality of first feature points and images around the coordinates of the plurality of first feature points, calculates a feature extraction matrix for estimating coordinates of a second feature point and an image around the coordinates of the second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputs a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from a plurality of combined vectors, and the feature extraction matrix;

a correlation information database that stores the first average vector, the second average vector and the feature extraction matrix input from the learning section; and

a parameter estimating section that estimates the coordinates of the second feature point, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix.

8. The parameter estimation apparatus according to claim 7, wherein the learning section comprises:

a first training vector calculating section that calculates a first training vector of vector pattern where the first electronic data is arranged in the order of scanning;

a second training vector calculating section that calculates a second training vector of vector pattern where combined vectors are arranged sequentially;

an average vector calculating section that calculates the first average vector by averaging a plurality of first training vectors and the second average vector by averaging a plurality of second training vectors;

an auto-correlation information calculating section that receives as its inputs the plurality of first training vectors and the first average vector, and calculates a pseudo inverse matrix of a first covariance matrix that is auto-correlation information of the first training vectors;

a cross-correlation information calculating section that receives as its inputs the plurality of first training vectors, the first average vector, the plurality of second training vectors and the second average vector, and calculates a second covariance matrix that is cross-correlation information of the first training vectors and the second training vectors; and

a feature extraction matrix calculating section that calculates the feature extraction matrix from the pseudo inverse matrix of the first covariance matrix and the second covariance matrix.

9. The parameter estimation apparatus according to claim 7, wherein the parameter estimating section comprises:

a vector calculating section that receives as its input the second electronic data, and calculates an input vector of vector pattern where the second electronic data is arranged in the order of scanning;

a parameter calculating section that calculates an expectation vector including the coordinates of the second feature point and the image around the coordinates of the second feature point from the input vector, the first average vector, the second average vector and the feature extraction matrix; and

a template matching section that estimates the coordinates of the second feature point and the image around the coordinates of the second feature point from the expectation vector, searches for a matching region that is an image region of the second electronic data matching the estimated image around estimated coordinates of the second feature point, and calculates the coordinates of the second feature point from the matching region.

10. A parameter estimation apparatus comprising:

an image input section that converts optical data into electronic data;

a parameter input section that inputs coordinates of a first feature point from first electronic data input from the image input section;

a learning section that divides a set of combined information of the first electronic data and the coordinates of the first feature point into a plurality of distributions, calculates for each distribution auto-correlation information from a plurality of items of first electronic data and cross-correlation information from the plurality of items of first electronic data and coordinates of a plurality of first feature points, calculates for each distribution a feature extraction matrix for estimating coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputs the first average vector calculated from the plurality of items of first electronic data obtained for each distribution, the second average vector calculated from the coordinates of the plurality of first feature points obtained for each distribution, and the feature extraction matrix obtained for each distribution;

a correlation information database that stores the first average vector, the second average vector and the feature extraction matrix obtained for each distribution and input from the learning section; and

a parameter estimating section that estimates coordinates of the second feature point for each distribution, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix.

11. The parameter estimation apparatus according to claim 10, wherein the learning section comprises:

a first training vector calculating section that calculates a first training vector of vector pattern where the first electronic data is arranged in the order of scanning;

a second training vector calculating section that calculates a second training vector of vector pattern where the coordinates of first feature points are arranged sequentially;

a distribution element calculating section that generates a plurality of combined vectors from the plurality of first training vectors and the plurality of second training vectors, and using probability distributions of a set of the combined vectors, divides the first training vectors and the second training vectors into a plurality of distributions;

an average vector calculating section that calculates for each distribution the first average vector by averaging a plurality of first training vectors and the second average vector by averaging a plurality of second training vectors;

an auto-correlation information calculating section that receives as its inputs the plurality of first training vectors and the first average vector for each distribution, and calculates for each distribution a pseudo inverse matrix of a first covariance matrix that is auto-correlation information of the first training vectors;

a cross-correlation information calculating section that receives as its inputs the plurality of first training vectors, the first average vector, the plurality of second training vectors and the second average vector for each distribution, and calculates for each distribution a second covariance matrix that is cross-correlation information of the first training vectors and the second training vectors; and

a feature extraction matrix calculating section that receives as its inputs the pseudo inverse matrix of the first covariance matrix and the second covariance matrix for each distribution, and calculates the feature extraction matrix from the pseudo inverse matrix of the first covariance matrix and the second covariance matrix.

12. The parameter estimation apparatus according to claim 10, wherein the parameter estimating section comprises:

a vector calculating section that receives as its input the second electronic data, and calculates an input vector of vector pattern where the second electronic data is arranged in the order of scanning;

a parameter calculating section that calculates an expectation vector including the coordinates of the second feature point from the input vector, the first average vector classified for each distribution, the second average vector classified for each distribution and the feature extraction matrix classified for each distribution; and

a template matching section that calculates the coordinates of the second feature point from the expectation vector.

13. A parameter estimation apparatus comprising:

an image input section that converts optical data into electronic data;

a parameter input section that inputs an image around coordinates of a first feature point from first electronic data input from the image input section;

a learning section that divides a set of combined information of the first electronic data and the image around the coordinates of the first feature point into a plurality of distributions, calculates for each distribution auto-correlation information from a plurality of items of first electronic data and cross-correlation information from the plurality of items of first electronic data and images around coordinates of a plurality of first feature points, calculates for each distribution a feature extraction matrix for estimating an image around coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputs the first average vector calculated from the plurality of items of first electronic data obtained for each distribution, the second average vector calculated from the images around the coordinates of the plurality of first feature points obtained for each distribution, and the feature extraction matrix obtained for each distribution;

a correlation information database that stores the first average vector, the second average vector and the feature extraction matrix obtained for each distribution and input from the learning section; and

a parameter estimating section that estimates coordinates of the second feature point for each distribution, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix.

14. The parameter estimation apparatus according to claim 13, wherein the learning section comprises:

a first training vector calculating section that calculates a first training vector of vector pattern where the first electronic data is arranged in the order of scanning;

a second training vector calculating section that calculates a second training vector of vector pattern where the coordinates of first feature points are arranged sequentially;

a distribution element calculating section that generates a plurality of combined vectors of the plurality of first training vectors and the plurality of second training vectors, and using probability distributions of a set of the combined vectors, divides the first training vectors and the second training vectors into a plurality of distributions;

an average vector calculating section that calculates for each distribution the first average vector by averaging a plurality of first training vectors and the second average vector by averaging a plurality of second training vectors;

an auto-correlation information calculating section that receives as its inputs the plurality of first training vectors and the first average vector for each distribution, and calculates for each distribution a pseudo inverse matrix of a first covariance matrix that is auto-correlation information of the first training vectors;

a cross-correlation information calculating section that receives as its inputs the plurality of first training vectors, the first average vector, the plurality of second training vectors and the second average vector for each distribution, and calculates for each distribution a second covariance matrix that is cross-correlation information of the first training vectors and the second training vectors; and

a feature extraction matrix calculating section that receives as its inputs the pseudo inverse matrix of the first covariance matrix and the second covariance matrix for each distribution, and calculates the feature extraction matrix from the pseudo inverse matrix of the first covariance matrix and the second covariance matrix.

15. The parameter estimation apparatus according to claim 13, wherein the parameter estimating section comprises:

a vector calculating section that receives as its input the second electronic data, and calculates an input vector of vector pattern where the second electronic data is arranged in the order of scanning;

a parameter calculating section that calculates an expectation vector including the image around the coordinates of the second feature point from the input vector, the first average vector classified for each distribution, the second average vector classified for each distribution and the feature extraction matrix classified for each distribution; and

a template matching section that calculates the image around the coordinates of the second feature point from the expectation vector, searches for a matching region that is an image region of the second electronic data matching the image around the coordinates of the second feature point, and calculates the coordinates of the second feature point from the matching region.

16. A parameter estimation apparatus comprising:

an image input section that converts optical data into electronic data;

a parameter input section that inputs coordinates of a first feature point and an image around the coordinates of the first feature point from first electronic data input from the image input section;

a learning section that divides a set of combined information of the first electronic data, the coordinates of the first feature point and the image around the coordinates of the first feature point into a plurality of distributions, calculates for each distribution auto-correlation information from a plurality of items of first electronic data and cross-correlation information from the plurality of items of first electronic data, coordinates of a plurality of first feature points and images around the coordinates of the plurality of first feature points, calculates for each distribution a feature extraction matrix for estimating coordinates of a second feature point and an image around the coordinates of the second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputs a first average vector calculated from the plurality of items of first electronic data obtained for each distribution, a second average vector calculated from the coordinates of the first feature points and the images around the coordinates of the first feature points obtained for each distribution, and the feature extraction matrix obtained for each distribution;

a correlation information database that stores the first average vector, the second average vector and the feature extraction matrix obtained for each distribution and input from the learning section; and

a parameter estimating section that estimates the coordinates of the second feature point for each distribution, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix.

17. The parameter estimation apparatus according to claim 16, wherein the learning section comprises:

a first training vector calculating section that calculates a first training vector of vector pattern where the first electronic data is arranged in the order of scanning;

a second training vector calculating section that calculates a second training vector of vector pattern where the coordinates of the first feature points and the images around the coordinates of the first feature points are arranged sequentially;

a distribution element calculating section that generates a plurality of combined vectors from the plurality of first training vectors and the plurality of second training vectors, and using probability distributions of a set of the combined vectors, divides the first training vectors and the second training vectors into a plurality of distributions;

an average vector calculating section that calculates for each distribution the first average vector by averaging a plurality of first training vectors and the second average vector by averaging a plurality of second training vectors;

an auto-correlation information calculating section that receives as its inputs the plurality of first training vectors and the first average vector for each distribution, and calculates for each distribution a pseudo inverse matrix of a first covariance matrix that is auto-correlation information of the first training vectors;

a cross-correlation information calculating section that receives as its inputs the plurality of first training vectors, the first average vector, the plurality of second training vectors and the second average vector for each distribution, and calculates for each distribution a second covariance matrix that is cross-correlation information of the first training vectors and the second training vectors; and

a feature extraction matrix calculating section that receives as its inputs the pseudo inverse matrix of the first covariance matrix and the second covariance matrix for each distribution, and calculates the feature extraction matrix from the pseudo inverse matrix of the first covariance matrix and the second covariance matrix.

18. The parameter estimation apparatus according to claim 16, wherein the parameter estimating section comprises:

a vector calculating section that receives as its input the second electronic data, and calculates an input vector of vector pattern where the second electronic data is arranged in the order of scanning;

a parameter calculating section that calculates an expectation vector including the coordinates of the second feature point and an image around the coordinates of the second feature point from the input vector, the first average vector classified for each distribution, the second average vector classified for each distribution and the feature extraction matrix classified for each distribution; and

a template matching section that calculates the coordinates of the second feature point and the image around the coordinates of the second feature point from the expectation vector, searches for a matching region that is an image region of the second electronic data matching the image around the coordinates of the second feature point, and calculates the coordinates of the second feature point from the matching region.

19. A data matching apparatus comprising:

an image input section that converts optical data into electronic data;

a parameter input section that inputs coordinates of a first feature point from first electronic data input from the image input section;

a learning section that calculates auto-correlation information from a plurality of items of first electronic data, calculates cross-correlation information from the plurality of items of first electronic data and coordinates of a plurality of first feature points, calculates a feature extraction matrix for estimating coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputs a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from the coordinates of the plurality of first feature points, and the feature extraction matrix;

a correlation information database that stores the first average vector, the second average vector and the feature extraction matrix input from the learning section;

a parameter estimating section that estimates coordinates of the second feature point, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix;

an image data base that stores the first electronic data; and

a matching section that calculates a matching region that is an image for use in matching from the second electronic data and the coordinates of the second feature point, and matches the matching region with the first electronic data stored in the image database to obtain matching.

20. A data matching apparatus comprising:

an image input section that converts optical data into electronic data;

a parameter input section that inputs an image around coordinates of a first feature point from first electronic data input from the image input section;

a learning section that calculates auto-correlation information from a plurality of items of first electronic data, calculates cross-correlation information from the plurality of items of first electronic data and images around coordinates of a plurality of first feature points, calculates a feature extraction matrix for estimating an image around coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputs a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from the images around the coordinates of the plurality of first feature points, and the feature extraction matrix;

a correlation information database that stores the first average vector, the second average vector and the feature extraction matrix input from the learning section;

a parameter estimating section that estimates coordinates of the second feature point, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix;

an image database that stores the first electronic data; and

a matching section that calculates a matching region that is an image for use in matching from the second electronic data and the coordinates of the second feature point, and matches the matching region with the first electronic data stored in the image database to obtain matching.

21. A data matching apparatus comprising:

an image input section that converts optical data into electronic data;

a parameter input section that inputs coordinates of a first feature point and an image around the coordinates of the first feature point from first electronic data input from the image input section;

a learning section that calculates auto-correlation information from a plurality of items of first electronic data, calculates cross-correlation information from the plurality of items of first electronic data and a combined vector of coordinates of a plurality of first feature points and images around the coordinates of the plurality of first feature points, calculates a feature extraction matrix for estimating coordinates of a second feature point and an image around the coordinates of the second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputs a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from a plurality of combined vectors, and the feature extraction matrix;

a correlation information database that stores the first average vector, the second average vector and the feature extraction matrix input from the learning section;

a parameter estimating section that estimates the coordinates of the second feature point, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix;

an image data base that stores the first electronic data; and

a matching section that calculates a matching region that is an image for use in matching from the second electronic data and the coordinates of the second feature point, and matches the matching region with the first electronic data stored in the image database to obtain matching.

22. A data matching apparatus comprising:

an image input section that converts optical data into electronic data;

a parameter input section that inputs coordinates of a first feature point from first electronic data input from the image input section;

a learning section that divides a set of combined information of the first electronic data and the coordinates of the first feature point into a plurality of distributions, calculates for each distribution auto-correlation information from a plurality of items of first electronic data and cross-correlation information from the plurality of items of first electronic data and coordinates of a plurality of first feature points, calculates for each distribution a feature extraction matrix for estimating coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputs the first average vector calculated from the plurality of items of first electronic data obtained for each distribution, the second average vector calculated from the coordinates of the plurality of first feature points obtained for each distribution, and the feature extraction matrix obtained for each distribution;

a correlation information database that stores the first average vector, the second average vector and the feature extraction matrix obtained for each distribution and input from the learning section;

a parameter estimating section that estimates the coordinates of the second feature point for each distribution, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix;

an image data base that stores the first electronic data; and

a matching section that calculates a matching region that is an image for use in matching from the second electronic data and the coordinates of the second feature point, and matches the matching region with the first electronic data stored in the image database to obtain matching.

23. A data matching method, comprising:

converting optical data into electronic data;

inputting coordinates of a first feature point from first electronic data input from an image input section;

calculating auto-correlation information from a plurality of items of first electronic data, calculating cross-correlation information from the plurality of items of first electronic data and coordinates of a plurality of first feature points, calculating a feature extraction matrix for estimating coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputting a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from the coordinates of the plurality of first feature points, and the feature extraction matrix;

storing the first average vector, the second average vector and the feature extraction matrix input from a learning section in a correlation information database, and estimating coordinates of the second feature point using the second electronic data, first average vector, the second average vector and the feature extraction matrix; and

storing the first electronic data in an image database, calculating a matching region that is an image for use in matching from the second electronic data and the coordinates of the second feature point, and collating the matching region with the first electronic data stored in the image database to obtain matching.

24. A data matching method, comprising:

converting optical data into electronic data;

inputting an image around coordinates of a first feature point from first electronic data input from an image input section;

calculating auto-correlation information from a plurality of items of first electronic data, calculating cross-correlation information from the plurality of items of first electronic data and images around coordinates of a plurality of first feature points, calculating a feature extraction matrix for estimating an image around coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputting a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from the images around the coordinates of the plurality of first feature points, and the feature extraction matrix;

storing the first average vector, the second average vector and the feature extraction matrix input from a learning section in a correlation information database, and estimating the coordinates of the second feature point using the second electronic data, the first average vector, the second average vector and the feature extraction matrix, the second electronic data, the first average vector, the second average vector and the feature extraction matrix; and

storing the first electronic data in an image database, calculating a matching region that is an image for use in matching from the second electronic data and the coordinates of the second feature point, and collating the matching region with the first electronic data stored in the image database to obtain matching.

25. A data matching method, comprising:

converting optical data into electronic data;

inputting coordinates of a first feature point and an image around the coordinates of the first feature point from first electronic data input from an image input section;

calculating auto-correlation information from a plurality of items of first electronic data, calculating cross-correlation information from the plurality of items of first electronic data and a combined vector of coordinates of a plurality of first feature points and images around the coordinates of the plurality of first feature points, calculating a feature extraction matrix for estimating coordinates of a second feature point and an image around the coordinates of the second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputting a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from a plurality of combined vectors, and the feature extraction matrix;

storing the first average vector, the second average vector and the feature extraction matrix input from a learning section in a correlation information database, and estimating coordinates of the second feature point, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix; and

storing the first electronic data in an image database, calculating a matching region that is an image for use in matching the second electronic data and the coordinates of the second feature point, and collating the matching region with the first electronic data stored in the image database to obtain matching.

26. A parameter estimation apparatus comprising:

converting optical data into electronic data;

inputting coordinates of a first feature point from first electronic data input from an image input section;

dividing a set of combined information of the first electronic data and the coordinates of the first feature point into a plurality of distributions, calculating for each distribution auto-correlation information from a plurality of items of first electronic data and cross-correlation information from the plurality of items of first electronic data and coordinates of a plurality of first feature points, calculating for each distribution a feature extraction matrix for estimating coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputting the first average vector calculated from the plurality of items of first electronic data obtained for each distribution, the second average vector calculated from the coordinates of the plurality of first feature points obtained for each distribution, and the feature extraction matrix obtained for each distribution;

storing the first average vector, the second average vector and the feature extraction matrix obtained for each distribution and input from a learning section, and estimating coordinates of the second feature point for each distribution, using the second electronic data, the first average vector, the second average vector and the feature extraction matrix; and

storing the first electronic data in an image database, calculating a matching region that is an image for use in matching from the second electronic data and the coordinates of the second feature point, and collating the matching region with the first electronic data stored in the image database to obtain matching.

27. A program for instructing a computer to execute the processing of:

converting optical data into electronic data;

inputting coordinates of a first feature point from input first electronic data;

calculating auto-correlation information from a plurality of items of first electronic data, calculating cross-correlation information from the plurality of items of first electronic data and coordinates of a plurality of first feature points, calculating a feature extraction matrix for estimating coordinates of a second feature point of input second electronic data input from the auto-correlation information and the cross-correlation information, and outputting a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from the coordinates of the plurality of first feature points, and the feature extraction matrix;

storing the first average vector, the second average vector and the feature extraction matrix input from a learning section in a correlation information database, and estimating coordinates of the second feature point using the second electronic data, the first average vector, the second average vector and the feature extraction matrix; and

storing the first electronic data in an image database, calculating a matching region that is an image for use in matching from the second electronic data and the coordinates of the second feature point, and collating the matching region with the first electronic data stored in the image database to obtain matching.

28. A program for instructing a computer to execute the processing of:

converting optical data into electronic data;

inputting an image around coordinates of a first feature point from first electronic data input from an image input section;

calculating auto-correlation information from a plurality of items of first electronic data, calculating cross-correlation information from the plurality of items of first electronic data and images around coordinates of a plurality of first feature points, calculating a feature extraction matrix for estimating an image around coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputting a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from the images around the coordinates of the plurality of first feature points, and the feature extraction matrix;

storing the first average vector, the second average vector and the feature extraction matrix input from a learning section in a correlation information database, and estimating coordinates of the second feature point using the second electronic data, the first average vector, the second average vector and the feature extraction matrix; and

storing the first electronic data in an image database, calculating a matching region that is an image for use in matching from the second electronic data and the coordinates of the second feature point, and collating the matching region with the first electronic data stored in the image database to obtain matching.

29. A program for instructing a computer to execute the processing of:

converting optical data into electronic data;

inputting coordinates of a first feature point and an image around the coordinates of the first feature point from first electronic data input from an image input section;

calculating auto-correlation information from a plurality of items of first electronic data, calculating cross-correlation information from the plurality of items of first electronic data and a combined vector of coordinates of a plurality of first feature points and images around the coordinates of the plurality of first feature points, calculating a feature extraction matrix for estimating coordinates of a second feature point and an image around the coordinates of the second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputting a first average vector calculated from the plurality of items of first electronic data, a second average vector calculated from a plurality of combined vectors, and the feature extraction matrix;

storing the first average vector, the second average vector and the feature extraction matrix input from a learning section in a correlation information database, and estimating coordinates of the second feature point using the second electronic data, the first average vector, the second average vector and the feature extraction matrix; and

storing the first electronic data in an image database, calculating a matching region that is an image for use in matching from the second electronic data and the coordinates of the second feature point, and collating the matching region with the first electronic data stored in the image database to obtain matching.

30. A program for instructing a computer to execute the processing of:

converting optical data into electronic data;

inputting coordinates of a first feature point from first electronic data input from an image input section;

dividing a set of combined information of the first electronic data and the coordinates of the first feature point into a plurality of distributions, calculating for each distribution auto-correlation information from a plurality of items of first electronic data and cross-correlation information from the plurality of items of first electronic data and coordinates of a plurality of first feature points, calculating for each distribution a feature extraction matrix for estimating coordinates of a second feature point of second electronic data input from the image input section using the auto-correlation information and the cross-correlation information, and outputting the first average vector calculated from the plurality of items of first electronic data obtained for each distribution, the second average vector calculated from the coordinates of the plurality of first feature points obtained for each distribution, and the feature extraction matrix obtained for each distribution;

storing the first average vector, the second average vector and the feature extraction matrix obtained for each distribution and input from a learning section in a correlation information database, and estimating coordinates of the second feature point for each distribution using the second electronic data, the first average vector, the second average vector and the feature extraction matrix; and

storing the first electronic data in an image database, calculating a matching region that is an image for use in matching from the second electronic data and the coordinates of the second feature point, and collating the matching region with the first electronic data stored in the image database to obtain matching.