Skin pattern and fingerprint classification system

• US 5,901,239 A
• Filed: 08/01/1996
• Issued: 05/04/1999
• Est. Priority Date: 08/02/1995
• Status: Expired due to Term

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Kamei, Toshio
• Primary Examiner(s): Couso, Jose L.
• Assistant Examiner(s): Dang, Duy M.
• Application Number:
  US08/690,744
• Time in Patent Office: 1,006 Days
• Field of Search: 382/115, 382/124, 382/125, 382/156, 382/157, 382/203, 382/204, 382/260, 382/266, 382/286, 395/23, 356/71
• US Class Current: 382/125
• CPC Class Codes: G06V 40/1347 Preprocessing; Feature extraction

Abstract

A skin pattern classification system extracts ridge line data and valley line data from image data of a skin pattern. Minutiae are extracted from the ridge line data and the valley line data. Subsequently, minutia correspondence information is obtained by detecting mutual correspondence among the minutiae from the ridge line data, valley line data and the minutiae. Skin patterns are obtained by finding and tracing characteristic lines referring to the minutia correspondence information, the ridge line data and the valley line data, and the characteristic lines and the minutia correspondence information serve to provide pattern classification.

Claims

1. A skin pattern tracing system comprising:

image data input means wherein image data of a skin pattern are input, said skin pattern comprising a plurality of ridges and valleys;

ridge line extracting means for extracting ridge line data corresponding to said ridges of said skin pattern from said image data;

valley line extracting means for extracting valley line data corresponding to said valleys of said skin pattern from said image data;

ridge minutia extracting means for extracting ridge minutiae from said ridge line data;

valley minutia extracting means for extracting valley minutiae from said valley line data;

minutia correspondence extracting means for extracting minutia correspondence information by detecting dual correspondence among said ridge minutiae and said valley minutiae, from said ridge line data, said valley line data, said ridge minutiae, and said valley minutiae, said dual correspondence defined to have a distance between minutiae less than a predetermined value and a line segment between said minutiae not crossing any valley or ridge; and

means for extracting features of characteristic lines of said skin pattern by finding and tracing said characteristic lines referring to said minutia correspondence information, said ridge line data and said valley line data.

2. A skin pattern classification system comprising a skin pattern tracing system recited in claim 1 and comprising a pattern classification means for classifying said skin pattern according to said features of characteristic lines and said minutia correspondence information.

3. A skin pattern classification system recited in claim 2, wherein said pattern classification means comprises:

means for pre-classifying said skin pattern into one of a plurality of detailed categories, said detailed categories defined according to said features of characteristic lines and said minutia correspondence information;

a probability table to be referred to with each of said detailed categories for obtaining classification probabilities, each of said classification probabilities corresponding to each of a plurality of output categories, said classification probabilities indicating a probability of a skin pattern that is pre-classified into said each of said detailed categories that is to be classified into said each of said output categories; and

means for classifying a skin pattern, wherein said skin pattern is pre-classified into each of said detailed categories, and wherein said skin pattern classifying means classifies said skin pattern into one of said output categories referring to said probability table.

4. A fingerprint tracing system comprising:

image data input means wherein image data of a fingerprint are input, said fingerprint comprising a pattern of a plurality of ridges and valleys;

ridge line extracting means for extracting ridge line data corresponding to said ridges of said pattern of said fingerprint from said image data;

valley line extracting means for extracting valley line data corresponding to said valleys of said pattern from said image data;

ridge minutia extracting means for extracting ridge minutiae from said ridge line data;

valley minutia extracting means for extracting valley minutiae from said valley line data;

minutia correspondence extracting means for extracting minutia correspondence information by detecting dual correspondence among said ridge minutiae and said valley minutiae, from said ridge line data, said valley line data, said ridge minutiae, and said valley minutiae, said dual correspondence defined to have a distance between minutiae less than a predetermined value and a line segment between said minutiae not crossing any valley or ridge; and

means for extracting features of characteristic lines of said fingerprint by finding and tracing said characteristic lines referring to said minutia correspondence information, said ridge line data and said valley line data.

5. A fingerprint classification system comprising a fingerprint tracing system recited in claim 4 and comprising a pattern classification means for classifying said fingerprint according to said features of characteristic lines and said minutia correspondence information.

6. A fingerprint classification system recited in claim 5, wherein said pattern classification means comprises:

means for pre-classifying said fingerprint into one of a plurality of detailed categories, said detailed categories defined according to said features of characteristic lines and said minutia correspondence information;

a probability table to be referred to with each of said detailed categories for obtaining classification probabilities, each of said classification probabilities corresponding to each of a plurality of output categories, said classification probabilities indicating a probability of a fingerprint that is pre-classified into said each of said detailed categories that is to be classified into said each of said output categories; and

means for classifying a fingerprint, wherein said fingerprint is pre-classified into each of said detailed categories wherein said fingerprint classifying means classifies a fingerprint into one of said output categories referring to said probability table.

7. A skin pattern tracing method comprising the steps of:

inputting image data of a skin pattern, said skin pattern comprising a plurality of ridges and valleys;

extracting ridge line data corresponding to said ridges of said skin pattern from said image data;

extracting valley line data corresponding to said valleys of said skin pattern from said image data;

extracting ridge minutia from said ridge line data;

extracting valley minutiae from said valley line data;

extracting minutia correspondence information by detecting dual correspondence among said ridge minutiae and said valley minutiae, from said ridge line data, said valley line data, said ridge minutiae and said valley minutiae, said dual correspondence defined to have a distance between minutiae less than a predetermined value and a line segment between said minutiae not crossing any valley or ridge; and

extracting features of characteristic lines of said skin pattern by finding and tracing said characteristic lines referring to said minutia correspondence information, said ridge line data and said valley line data.

8. The method as recited in claim 7 further comprising the step of classifying said skin pattern according to said features of characteristic lines and said minutia correspondence information.

9. A method as recited in claim 8 wherein said step of classifying said skin pattern further comprises the steps of:

pre-classifying said skin pattern into one of a plurality of detailed categories, said detailed categories defined according to said features of characteristic lines and said minutia correspondence information;

obtaining classification probabilities from a probability table for each detailed category, each of said classification probabilities corresponding to each of a plurality of output categories, said classification probabilities indicating a probability of a skin pattern that is pre-classified into said each of said detailed categories that is to be classified into said each of said output categories; and

classifying a skin pattern, wherein said skin pattern is pre-classified into each of said detailed categories, by classifying said skin pattern into one of said output categories referring to said probability table.

10. A fingerprint tracing method comprising the steps of:

inputting image data of a fingerprint, said fingerprint comprising a pattern of a plurality of ridges and valleys;

extracting ridge line data corresponding to said ridges of said pattern of said fingerprint from said image data;

extracting valley line data corresponding to said valleys of said pattern from said image data;

extracting ridge minutiae from said ridge line data;

extracting valley minutiae from said valley line data;

extracting minutia correspondence information by detecting dual correspondence among said ridge minutiae and said valley minutiae, from said ridge line data, said valley line data, said ridge minutiae, and said valley minutiae, said dual correspondence defined to have a distance between minutiae less than a predetermined value and a line segment between said minutiae not crossing any valley or ridge; and

extracting features of characteristic lines of said fingerprint by finding and tracing said characteristic lines referring to said minutia correspondence information, said ridge line data and said valley line data.

11. A method as recited in claim 10 further comprising the step of classifying said fingerprint according to said features of characteristic lines and said minutia correspondence information.

12. The method as recited in claim 11 wherein said classifying step further comprises the steps of:

pre-classifying said fingerprint into one of a plurality of detailed categories, said detailed categories defined according to said features of characteristic lines and said minutia correspondence information;

obtaining classification probabilities from a probability table for each detailed category, each of said classification probabilities corresponding to each of a plurality of output categories, said classification probabilities indicating a probability of a fingerprint that is pre-classified into said each of said detailed categories that is to be classified into said each of said output categories; and

classifying a fingerprint, wherein said fingerprint is pre-classified into each of said detailed categories, wherein said classifying step classifies said fingerprint into one of said output categories referring to said probability table.