Method and device for matching fingerprints with precise minutia pairs selected from coarse pairs

• US 4,646,352 A
• Filed: 06/28/1983
• Issued: 02/24/1987
• Est. Priority Date: 06/28/1982
• Status: Expired due to Term

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Asai, Koh, Owada, Katsuaki, Matsuno, Shunji, Izumisawa, Hiroyuki, Kinoshita, Seiichiro
• Primary Examiner(s): Boudreau, Leo H.
• Application Number:
  US06/508,759
• Time in Patent Office: 1,337 Days
• Field of Search: 356/71, 382/5, 382/21, 382/30, 382/34
• US Class Current: 382/125
• CPC Class Codes: G06V 40/12 Fingerprints or palmprints
• CPC Class Codes: G07C 9/37 using biometric data, e.g. fingerprints, iris scans or voice recognition

Abstract

A pair candidate list (73) is formed by selecting minutia pairs with reference to a minutia list (71) showing original position and direction data given for minutiae by principal coordinate systems preliminarily selected on a search and a file fingerprint and those relation data of the minutiae which are substantially independent of the coordinate systems. It is very likely that the pair candidate list shows coarse pairs because the coordinate systems may not yet be optimally matched to each other. One of the coordinate systems is transformed by those optimum amounts to provide transformed position and direction data which are decided by the original position and direction data of the minutia pairs of the pair candidate list. A pair list (86) is formed by precisely selecting minutiae from the pair candidate list with reference to the transformed position and direction data and the original position and direction data given by the other principal coordinate system and to the relation data. On forming the pair list, an additional minutia list (81) is preferably formed which shows the transformed position and direction data and the last-mentioned original position and direction data together with the relation data.

Claims

1. A method of deciding a degree of match between a search and a file fingerprint, comprising the steps of:

selecting preliminarily matched coordinate systems on the respective fingerprints;

selecting a search and a file fingerprint area where minutiae are clear;

forming a minutia list showing those original position and direction data of the minutiae which are given by the respective preliminarily matched coordinate systems, said minutia list furthermore showing those relation data of the minutiae which are substantially independent of said preliminarily matched coordinate systems;

selecting pairs of minutiae from the minutiae of the respective fingerprints by comparing the original position and direction data given for the respective fingerprints and furthermore comparing the relation data given for the respective fingerprints;

forming a pair candidate list which shows said pairs of minutiae as pair candidates, respectively;

deciding optimum amounts of coordinate match by referring to the original position and direction data of the minutiae of said pair candidates, said optimum amounts being for precisely matching said preliminarily matched coordinate systems with each other;

transforming the original position and direction data given by a predetermined one of said preliminarily matched coordinate systems for the minutiae of said pair candidates into transformed position and direction data given by a transformed coordinate system into which said predetermined one of the preliminarily matched coordinate systems is transformed by said optimum amounts;

forming a pair list by referring to said minutia list, said pair candidate list, and said search and said file fingerprint areas, said pair list showing precise pairs of minutiae selected from those minutiae of said pair candidates which are present in an area common to said search and said file fingerprint areas, the minutiae of each precise pair having a precise local similarity between the transformed position and direction data and the original position and direction data given by a different one of said preliminarily matched coordinate systems and between the relation data, said different one of the preliminarily matched coordinate systems being different from said predetermined one of the preliminarily matched coordinate systems, said pair list furthermore showing evaluations, each evaluation being representative of said local similarity; and

deciding said degree of match using the minutiae of said pair list and by referring to said evaluations.

2. A method as claimed in claim 1, wherein said minutia list forming step comprises the steps of:

selecting primary local coordinate systems which having primary local origins at the respective minutiae;

selecting primary proximate minutiae with reference to the respective primary local coordinate systems, said primary proximate minutiae being proximate to the respective local origins;

deciding primary data of said primary proximate minutiae with reference to the respective primary coordinate systems;

selecting secondary local coordinate systems which have secondary local origins at the respective primary proximate minutiae;

selecting secondary proximate minutiae with reference to the respective secondary local coordinate systems, said secondary proximate minutiae being proximate to the respective secondary local origins;

deciding secondary data of said secondary proximate minutiae with reference to the respective primary local coordinate systems; and

listing said primary and said secondary data as said relation data.

3. A method as claimed in claim 1, wherein said pair list forming step comprises the steps of:

forming an additional minutia list which shows said transformed position and direction data, the original position and direction data given by said different one of the preliminarily matched coordinate systems for the minutiae of said pairs of candidates, and the relation data of the minutiae of said pair candidates; and

forming said pair list by referring to said additional minutia list, said pair candidate list, and said search and said file fingerprint areas, the transformed position and direction data given in said additional minutia list being used in evaluating said precise local similarity.

4. A fingerprint matching device for deciding a degree of match between a search and a file fingerprint, comprising:

minutia list producing means (71, 72,

     81) for producing a minutia list showing those original position and direction data of minutiae which are given by preliminarily matched coordinate systems selected on said search and said file fingerprints, respectively, said minutia list furthermore showing those relation data of said minutiae which are substantially independent of said preliminarily matched coordinate systems;

fingerprint area memory means (87) for memorizing a search and a file fingerprint area where the minutiae are clear;

selecting means (77) coupled to said minutia list producing means for selecting pairs of minutiae from the minutiae of the respective fingerprints, said pairs of minutiae having similar original position and direction data given for the respective fingerprints and furthermore similar relation data given for the respective fingerprints;

pair candidate list memory means (73) coupled to said selecting means for memorizing a pair candidate list which shows said pairs of minutiae as pair candidates, respectively;

optimum amount deciding means (78,

     75) coupled to said minutia list producing means and said pair candidate list memory means for deciding optimum amounts of coordinate match by referring to the original position and direction data of the minutiae of said pair candidates, said optimum amounts being for precisely matching said preliminarily matched coordinate systems with each other;

coordinate transforming means (72) coupled to said minutia list producing means, said pair candidate list memory means, and said optimum amount deciding means for transforming a predetermined one of said preliminarily matched coordinate systems into a transformed coordinate system by said optimum amounts, said coordinate transforming means thereby transforming the original position and direction data given by said predetermined one of the preliminarily matched coordinate systems for the minutiae of said pair candidates into transformed position and direction data;

pair list memory means (86,

     83) coupled to said minutia list producing means, said pair candidate list memory means, and said fingerprint area memory means for memorizing a pair list showing precise pairs of minutiae selected from those minutiae of said pair candidates which are present in an area common to said search and said file fingerprint areas, the minutiae of each precise pair having a precise local similarity between the transformed position and direction data and the original position and direction data given by a different one of said preliminarily matched coordinate systems and between the relation data, said different one of the preliminarily matched coordinate systems being different from said predetermined one of the preliminarily matched coordinate systems, said pair list furthermore showing evaluations, each evaluation being representative of said local similarity; and

deciding means (89) coupled to said pair list memory means for deciding said degree of match by using the minutiae of said pair list and by referring to said evaluations.

5. A fingerprint matching device as claimed in claim 4, wherein said minutia list producing means comprises:

first means (in

     197) for producing a remaining part of said minutia list, said part showing the original position and direction data given by the respective preliminarily matched coordinate systems and furthermore showing that part of said relation data which gives primary data represented by primary local coordinate systems selected to have primary local origins at the respective minutiae, said primary data being given to primary proximate minutiae with reference to the respective primary local coordinate systems, said primary proximate minutiae being selected proximate to the respective primary local origins with reference to the respective primary local coordinate systems; and

second means (in

     197) for producing a remaining part of said minutia list, said remaining part showing that remaining part of said relation data which gives secondary data represented by the respective primary local coordinate systems, said secondary data being given to secondary proximate minutiae which are selected proximate to the respective primary proximate minutiae with reference to the respective ones of secondary local coordinate systems selected to have secondary local origins at the respective primary proximate minutiae.

6. A fingerprint matching device as claimed in claim 5, wherein:

said primary proximate minutiae are selected proximate to each primary local origin in a predetermined number of sectors which have a common vertex at said each primary local origin and into which the primary local coordinate system having said each primary local origin is divided;

said secondary proximate minutiae being selected proximate to each of the secondary local origins located at the respective primary proximate minutiae selected for said each primary local origin in a prescribed number of sectors which have a common vertex at said each of the secondary local origins and into which the secondary local coordinate system having said each of the secondary local origins is divided;

said secondary data comprising ridge counts between said each primary local origin and the respective secondary proximate minutiae which are selected for the secondary local origins located at the respective primary proximate minutiae selected for said each primary local origin.

7. A fingerprint matching device as claimed in claim 4, wherein:

said minutia list producing means comprises;

first minutia list producing means (71,

     72) for producing a first minutia list which shows the original position and direction data given by the respective preliminarily matched coordinate systems and the relation data of the respective minutiae; and

second minutia list producing means (81) coupled to said coordinate transforming means and said pair candidate list memory for producing a second minutia list which shows said transformed position and direction data, the original position and direction data given by said different one of the preliminarily matched coordinate systems for the minutiae of said pair candidates, and the relation data for the minutiae of said pair candidates;

said selecting means being coupled to said first minutia list producing means;

said optimum amount deciding means being coupled to said first minutia list producing means and said pair candidate list memory means;

said coordinate transforming means being coupled to said first minutia list producing means, said pair candidate list memory means, and said optimum amount deciding means;

said pair list memory means being coupled to said second minutia list producing means, said pair candidate list memory means, and said fingerprint area memory means, the transformed position and direction data used in giving said precise local similarity being those given by said second minutia list producing means.

8. A fingerprint matching device as claimed in claim 4, wherein said relation data comprise a ridge count between each minutia of said search and said file fingerprint and a minutia which is proximate to said each minutia in said search and said file fingerprints.

9. A fingerprint matching device as claimed in claim 4, wherein:

said pair candidate list memory means is for memorizing the pair candidate list which furthermore shows a local similarity measure for each pair candidate;

said optimum amount deciding means uses the local similarity measures for the respective pair candidates on deciding said optimum amount;

said pair list memory means determines said precise local similarity by using the local similarity measures selected for the respective precise pairs.

10. A fingerprint matching device as claimed in claim 9, the pair candidate consisting of a first and a second minutia, wherein said optimum amount deciding means comprises:

a two-dimensional memory (211) supplied with the local similarity measures for the respective pair candidates from said pair candidate list memory means;

means (212-214) for accumulating the local similarity measures in said two-dimensional memory according to coordinate increment sets and angular increments, each coordinate increment set representing differences between the original position data of the first minutia of each pair candidate and the transformed position data into which the original position data of the second minutia of said each pair candidate are transformed when each angular increment is given to the preliminarily matched coordinate system selected for the fingerprint including the second minutia of said each pair candidate; and

means (75) for giving said optimum amounts by the coordinate increment set and the angular increment for which the local similarity measures are most accumulated in said two-dmensional memory.