Apparatus for automatic fingerprint classification

• US 5,497,429 A
• Filed: 09/29/1994
• Issued: 03/05/1996
• Est. Priority Date: 10/01/1993
• Status: Expired due to Term

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Shibuya, Tatsuo
• Primary Examiner(s): Mancuso, Joseph
• Assistant Examiner(s): PARK, BAEIL PET
• Application Number:
  US08/313,562
• Time in Patent Office: 523 Days
• Field of Search: 382/2, 382/3, 382/4, 382/5, 382/30, 382/34, 382/22, 382/60, 382/124, 382/125, 382/127, 382/126, 382/199, 382/313, 382/316, 382/209, 382/218
• US Class Current: 382/125
• CPC Class Codes: G06V 40/1347 Preprocessing; Feature extraction
• CPC Class Codes: G07C 9/37 using biometric data, e.g. fingerprints, iris scans or voice recognition

Abstract

In fingerprint pattern classification, radial scanning vectors are used. A space between two adjacent ridge lines including a core point is traced by radial scanning. Radial distribution of number of cross points from the core point is a clue to a fingerprint classification.

Claims

1. An apparatus for automatic fingerprint classification of a fingerprint image having a finger joint image, a fingertip image, ridge lines and a core point, said core point being used as an origin for original x-y rectangular coordinates, an x-axis of said original x-y rectangular coordinates being parallel to said finger joint image and said fingertip image being in a y>

0 direction, said apparatus comprising;

an initialization unit for setting current temporary x-y rectangular coordinates with said original x-y rectangular coordinates;

means for tracing space between two ridge lines starting from said core point in an area where x>

0 and in an area where x<

0 with respect to said original x-y rectangular coordinates, said means for tracing space including,

a scanning unit for distance measurement, wherein said scanning unit for distance measurement scans said fingerprint image by vectors having a same magnitude, starting from an origin of said current temporary x-y rectangular coordinates and diverging in different directions, for obtaining measured distances from said origin of said current temporary x-y rectangular coordinates to first cross points between said ridge lines in said fingerprint image and said vectors,

a next coordinate determination unit, wherein said next coordinate determination unit selects a nearest vector to one of a -x axis and a +x axis of said current temporary x-y rectangular coordinates among said vectors with said measured distances larger than a predetermined threshold distance, sets said nearest vector as one of a -x axis and a +x axis of next temporary x-y rectangular coordinates, and decides a middle point between said origin of said current temporary x-y rectangular coordinates and said first cross point of said one of said -x axis and said +x axis of said next temporary x-y rectangular coordinates as an origin of said next temporary x-y rectangular coordinates,

a tracing vector memory for storing position coordinates of all origins of said current and next temporary x-y rectangular coordinates with respect to said original x-y rectangular coordinates,

a judgment unit for deciding a termination condition to go back to said scanning unit, and

an update unit for setting said current temporary x-y rectangular coordinates with said next temporary x-y rectangular coordinates; and

a classifier unit for classifying said fingerprint image from data stored in said tracing vector memory, wherein loci of said origins of said next temporary x-y rectangular coordinates stored in said tracing vector memory in -x and +x areas with respect to said original x-y rectangular coordinates are examined and classified as one of a straight line and a semicircle,

a fingerprint pattern with a semicircle in both said -x and +x area with respect to said original x-y rectangular coordinates is classified as a whorl,

a fingerprint pattern with a semicircle in said +x area with respect to said original x-y rectangular coordinates and a straight line in said -x area with respect to said original x-y rectangular coordinates is classified as a left loop,

a fingerprint pattern with a semicircle in said -x area with respect to said original x-y rectangular coordinates and a straight line in said +x area with respect to said original x-y rectangular coordinates is classified as a right loop, and

a fingerprint pattern with straight line in both said -x and +x area with respect to said original x-y rectangular coordinates is classified as an arch.

2. An apparatus for automatic fingerprint classification, comprising:

means for converting a fingerprint image to a set of binary data and storing said set of binary data in a fingerprint image memory, said fingerprint image including a finger joint image, a fingertip image, and ridge lines;

a core point determination unit for detecting a core point of a fingerprint image stored in said fingerprint image memory, and for choosing original x-y rectangular coordinates, wherein an origin of said original x-y rectangular coordinates is at said core point, an x axis of said original x-y rectangular coordinates being parallel to said finger joint image, and said fingertip image being in a y>

0 direction;

an initialization unit for setting current temporary x-y rectangular coordinates with said original x-y rectangular coordinates;

means for tracing space between two ridge lines starting from said core point in an area where x>

0 and in an area where x<

0 with respect to said original x-y rectangular coordinates, said means for tracing space including,

a scanning unit for distance measurement, wherein said scanning unit for distance measurement scans said fingerprint image by vectors having a same magnitude, starting from an origin of said current temporary x-y rectangular coordinates and diverging in different directions, for obtaining measured distances from said origin of said current temporary x-y rectangular coordinates to first cross points between said ridge lines in said fingerprint image and said vectors,

a next coordinate determination unit, wherein said next coordinate determination unit selects a nearest vector to one of a -x axis and a +x axis of said current temporary x-y rectangular coordinates among said vectors with said measured distances larger than a predetermined threshold distance, sets said nearest vector as one of a -x axis and a +x axis of next temporary x-y rectangular coordinates, and decides a middle point between said origin of said current temporary x-y rectangular coordinates and said first cross point of said one of said -x axis and said +x axis of said next temporary x-y rectangular coordinates as an origin of said next temporary x-y rectangular coordinates,

a tracing vector memory for storing position coordinates of all origins of said current and next temporary x-y rectangular coordinates with respect to said original x-y rectangular coordinates,

a judgment unit for deciding a termination condition to go back to said scanning unit, and

an update unit for setting said current temporary x-y rectangular coordinates with said next temporary x-y rectangular coordinates; and

a classifier unit for classifying said fingerprint image from data stored in said tracing vector memory, wherein loci of said origins of said next temporary x-y rectangular coordinates stored in said tracing vector memory in -x and +x areas with respect to said original x-y rectangular coordinates are examined and classified as one of a straight line and a semicircle,

a fingerprint pattern with a semicircle in both said -x and +x area with respect to said original x-y rectangular coordinates is classified as a whorl,

a fingerprint pattern with a semicircle in said +x area with respect to said original x-y rectangular coordinates and a straight line in said -x area with respect to said original x-y rectangular coordinates is classified as a left loop,

a fingerprint pattern with a semicircle in said -x area with respect to said original x-y rectangular coordinates and a straight line in said +x area with respect to said original x-y rectangular coordinates is classified as a right loop, and

a fingerprint pattern with straight line in both said -x and +x area with respect to said original x-y rectangular coordinates is classified as an arch.