Method and apparatus for collating images

• US 6,052,474 A
• Filed: 08/23/1996
• Issued: 04/18/2000
• Est. Priority Date: 08/30/1995
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Nakayama, Akihito
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): Werner, Brian P.
• Application Number:
  US08/702,017
• Time in Patent Office: 1,334 Days
• Field of Search: 382/112, 382/124, 382/125, 382/126, 382/127, 382/115, 382/118, 382/149, 382/150, 382/151, 382/217, 382/218, 382/219, 382/220, 382/294, 382/295, 382/194, 382/202, 382/278 , 382/195, 340/825.34, 356/71
• US Class Current: 382/124
• CPC Class Codes: A61B 5/1172 using fingerprinting
• CPC Class Codes: G06V 40/1365 Matching; Classification

Abstract

The present invention provides a method of collating images and an apparatus for collating images capable of collating an image for collation taken in as digital data with an image for reference stored as digital data comparatively simply and surely. There are provided image information extracting means for extracting plural pieces of one-dimensional information from within an image for collation or within an image for reference, positional information group determining means for collating the one-dimensional information with an image for collation or an image for reference of the object, thereby to obtain positional information groups at locations where a coefficient of coincidence of data reaches a set value or higher with respect to respective one-dimensional information, and deciding means for deciding that the image for collation and the image for reference are identical when a combination of data having a relative positional relationship which coincides with the relative positional relationship of the extracted one-dimensional information or is within prescribed dispersion exists among positional information groups obtained with respect to respective one-dimensional information.

Claims

1. A method of comparing working and reference images stored as digital data comprising:

extracting a plurality of line segment data from the digital data representing the working image, the extracted line segments being noncontiguous and having a relative positional relationship among themselves, wherein said extracted data for each line segment comprises data for each pixel in that line segment of said working image;

comparing each extracted line segment with digital data representing the reference image to identify, by location within said reference image, every line segment from said digital data of said reference image which matches each extracted line segment from said working image to a predetermined degree; and

determining that the working and reference images are identical if a match is found for each of said extracted line segments of said working image, wherein the positional relationship among matching line segments from said reference image match said relative positional relationship among said extracted line segments from said working image to a predetermined degree.

2. A method of collating working and reference images stored as digital data according to claim 1, wherein said comparing each extracted line segment from said working image with the digital data representing said reference image is executed in a horizontal direction with overlapping of line segments from said reference image being taken for comparison.

3. A device for comparing collated data from a first image with data from a second image, the device comprising:

an imaging device for generating digital data representing said first image;

a processor for extracting a plurality of collated data strings from said digital data representing a plurality of noncontiguous one-dimensional segments of said first image, wherein each of said data strings comprises data for each pixel in a one-dimensional linear segment of said first image; and

a comparator for comparing each of said plurality of collated data strings with correspondingly sized segments of digital data representing said second image and determining, within a predetermined tolerance, every segment of the second image that matches each collated data string;

wherein, if a match for each of said plurality of collated data strings is found within said digital data representing said second image in a positional relationship with respect to said second image matching that of said plurality of collated data strings with respect to said first image within a second predetermined tolerance, said first image is determined to match said second image.

4. The device of claim 3, wherein said imaging device comprises:

a light source;

a CCD camera; and

a triangular prism optically coupled between said light source and said camera, said prism presenting a surface from which said first image is taken.

5. The device of claim 3, wherein said first and second images are images of fingerprints.

6. The device of claim 3, further comprising a static memory device for storing said digital data representing said second image, wherein said memory device is connected to said imaging device such that said digital data representing said second image may be generated using said imaging device.

7. The device of claim 3, wherein said comparator comprises:

a coefficient of collation detecting logic circuit for counting a number of bits which match between one of said collated data strings and a correspondingly sized segment of said digital data representing said second image; and

a coefficient of collation threshold setting register for setting a minimum count which must be reached by said logic circuit before said collated data string is held to match said segment of said digital data representing said second image.

8. A method of collating working and reference images stored as digital data comprising:

extracting a plurality of line segment data from the digital data representing noncontiguous one-dimensional portions of the working image, the extracted line segments having a relative positional relationship among themselves;

comparing each extracted line segment with digital data representing the reference image to identify, by location within said reference image, line segment data from said digital data of said reference image which matches each extracted line segment from said working image to a predetermined degree; and

determining that the working and reference images are identical if matches are found for each of said extracted line segment of said working image, wherein a positional relationship among matching line segments from said reference image match said relative positional relationship among said extracted line segments from said working image to a predetermined degree;

wherein said plurality of one-dimensional segments are five data strings of 64 bits each.

9. A method for matching first and second images comprising:

extracting a plurality of collated data strings from digital data representing said first image, said collated data strings representing a plurality of noncontiguous one-dimensional segments of said first image, wherein each of said collated data strings comprises data for each pixel in a one-dimensional linear segment of said first image; and

comparing each of said plurality of collated data strings with correspondingly sized segments of digital data representing said second image and determining, within a predetermined tolerance, every segment of the second image that matches each collated data string;

wherein, if a match for each of said plurality of collated data strings is found within said digital data representing said second image in a positional relationship with respect to said second image matching that of said plurality of collated data strings with respect to said first image within a second predetermined tolerance, said first image is determined to match said second image.

10. The method of claim 9, further comprising generating said digital data of said first image by passing light through a triangular prism to a CCD camera so as to image an object on a surface of said prism.

11. The method of claim 9, wherein said first and second images are images of fingerprints.

12. The method of claim 9, further comprising:

generating said digital data of said second image; and

storing said digital data representing said second image in a memory device.

13. The method of claim 9, wherein said comparing comprises:

counting a number of bits which match between one of said collated data strings and a correspondingly sized segment of said digital data representing said second image; and

setting a minimum count which must be reached before said collated data string is held to match said segment of said digital data representing said second image.

14. A method for matching first and second images comprising:

extracting a plurality of collated data strings from digital data representing said first image, said collated data strings representing a plurality of noncontiguous one-dimensional segments of said first image; and

comparing each of said plurality of collated data strings with correspondingly sized segments of digital data representing said second image;

wherein, if, within a predetermined tolerance, matches for said plurality of collated data strings are found within said digital data representing said second image in a positional relationship with respect to said second image matching that of said plurality of collated data strings with respect to said first image within a second predetermined tolerance, said first image is determined to match said second image; and

wherein said plurality of one-dimensional segments are five data strings of 64 bits each.

15. A device for comparing collated data from a first image with data from a second image, the device comprising:

imaging means for generating digital data representing said first image;

processing means for extracting a plurality of collated data strings from said digital data representing a plurality of noncontiguous one-dimensional segments of said first image, wherein each of said collated data strings comprises data for each pixel in a one-dimensional linear segment of said first image; and

comparing means for comparing each of said plurality of collated data strings with correspondingly sized segments of digital data representing said second image and for determining, within a predetermined tolerance, every segment of the second image that matches each collated data string;

wherein, if a match for each of said plurality of collated data strings is found within said digital data representing said second image in a positional relationship with respect to said second image matching that of said plurality of collated data strings with respect to said first image within a second predetermined tolerance, said first image is determined to match said second image.

16. The device of claim 15, wherein said imaging means comprises:

an illumination means;

a camera means; and

a surface optically coupled between said illumination means and said camera means from which said first image is taken.

17. The device of claim 15, wherein said first and second images are images of fingerprints.

18. The device of claim 15, further comprising a memory means for storing said digital data representing said second image, wherein said memory means are connected to said imaging means such that said digital data representing said second image may be generated using said imaging means.

19. The device of claim 15, wherein said comparing means comprises:

counting means for counting a number of bits which match between one of said collated data strings and a correspondingly sized segment of said digital data representing said second image; and

a coefficient of collation threshold setting means for setting a minimum count which must be reached by said counting means before said collated data string is held to match said segment of said digital data representing said second image.

20. A device for comparing collated data from a first image with data from a second image, the device comprising:

imaging means for generating digital data representing said first image;

processing means for extracting a plurality of collated data strings from said digital data representing a plurality of noncontiguous one-dimensional segments of said first image; and

comparing means for comparing each of said plurality of collated data strings with correspondingly sized segments of digital data representing said second image;

wherein, if, within a predetermined tolerance, matches for said plurality of collated data strings are found within said digital data representing said second image in a positional relationship with respect to said second image matching that of said plurality of collated data strings with respect to said first image within a second predetermined tolerance, said first image is determined to match said second image; and

wherein said plurality of one-dimensional segments are five data strings of 64 bits each.