Method of registering original information for personal identity authentication and personal identity authentication device
Patent Information
- Application Number
- MYPI1999000555
- Authority / Receiving Office
- MY · MY
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 1998-02-17
- Filing Date
- 1999-02-18
- Publication Date
- 2000-08-18
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Current fingerprint registration systems face challenges in efficiently comparing and selecting a reliable reference feature point data from multiple fingerprint images, leading to suboptimal matching and security vulnerabilities.
A CPU-based method that extracts feature points from multiple fingerprint images, compares them to determine similarity levels, and selects the maximum sum data as reference feature point data for registration, enhancing the robustness of fingerprint recognition.
This approach improves the accuracy and security of fingerprint registration by identifying a robust reference feature point data, thereby enhancing the reliability of fingerprint matching and reducing false positives.
Abstract
Description
On registration of a fingerprint image, a fingerprint is taken as the fingerprint image in number of N by a fingerprint image input section to produce first through N-th fingerprint images, where N represents a positive integer which is not less than three. A CPU (120) extracts (132) a feature point from each of the first through the N-th fingerprint images to produce first through N-th feature point data. The CPU compares (135) an n-th feature data with each of remaining feature data to produce levels of similarities as coincidence degrees in each of the first through the N-th feature point data, where n is variable between 1 and N, both of inclusive. The CPU selects (138) a maximum one of the first through the N-th sum data to register, as a reference feature point data therein, one of the first through the N-th feature data that has the maximum sum data.
Claims
1. A method of registering original information for personal identity authentication, said original information being used as a reference for verification of an authentication subject, comprising: 5 a first step of extracting (132) a feature point in number of N from said authentication subject to produce first through N-th feature point data, where N represents a positive integer which is not less than three; a second step of comparing (135) an n-th feature data with each of remaining feature data to produce levels of similarities as coincidence 10 degrees in each of said first through said N-th feature point data, where n is variable between 1 and N, both of inclusive; a third step of calculating (137) a sum degree of said coincidence degrees in each of said first through said N-th feature point data to produce first through N-th sum data; and 15 a fourth step of selecting (138) a maximum one of said first through said N-th sum data to register, as said original information, one of said first through said N-th feature data that has said maximum sum data.
2. A method as claimed in Claim 1, wherein said first step comprises: a fifth step of extracting (132, S110) a feature point in number of N 20 from said authentication subject to produce first through N-th feature points; and a sixth step of converting (133, S111) said first through said N-th feature points into first through N-th numerical values to produce said first through N-th numerical values as said first through said N-th feature point a seventh step of comparing (135, S113) said n-th feature data with each of said remaining feature data to produce levels of similarities in each of said first through said N-th feature point data; and an eight step of converting (135, S113) said levels into numerical 5 levels in each of said first through said N-th feature point data to produce said numerical levels as said coincidence degrees.
4. A method as claimed any one of Claims 1 to 3, wherein said authentication subject is a fingerprint.
5. A method as claimed any one of Claims 1 to 3, wherein said 10 authentication subject is a palm print.
6. A method as claimed any one of Claims 1 to 3, wherein said authentication subject is a voice print.
7. A method as claimed any one of Claims 1 to 3, wherein said authentication subject is a structural pattern of capillary vessels of retinas.
158. A method as claimed any one of Claims 1 to 3, wherein said authentication subject is a pattern of irises.
9. A method as claimed any one of Claims 1 to 3, wherein said authentication subject is a facial shape.
10. A personal identity authentication device for authenticating an 20 authentication subject, comprising: first means (132) for extracting a feature point from said authentication subject to produce extracted feature point data on authentication, said first means extracting a feature point in number of N from said authentication subject to produce first through N-th feature point 25 data on registration, where N represents a positive integer which is not less than three; second means (134) for registering a reference feature point data therein; third means (135) for comparing said extracted feature point data with said reference feature point data to judge whether or not said authentication subject is true on authentication, said third means comparing an n-th feature data with each of remaining feature data to produce levels of similarities as coincidence degrees in each of said first through said N-th feature point data on registration, where n is variable between 1 and N, both of inclusive; means (137) for calculating a sum degree of said coincidence degrees in each of said first through said N-th feature point data to produce first through N-th sum data; and fourth means (138) for selecting a maximum one of said first through said N-th sum data to register, as said reference feature point data in said second means, one of said first through said N-th feature data that has said maximum sum data.
11. A personal identity authentication device as claimed in Claim 10, wherein said second means is an IC card.
12. A recording medium memorizing a control program which is for registering original information for personal identity authentication, said original information being used as a reference for verification of an authentication subject, wherein said control program comprises: a first procedure of extracting (132) a feature point in number of N from said authentication subject to produce first through N-th feature point data, where N represents a positive integer which is not less than three; a second procedure of comparing (135) an n-th feature data with each of remaining feature data to produce levels of similarities as coincidence degrees in each of said first through said N-th feature point data, where n is variable between 1 and N, both of inclusive; a third procedure of calculating (137) a sum degree of said coincidence degrees in each of said first through said N-th feature point data to produce first through N-th sum data; and a fourth procedure of selecting (138) a maximum one of said first through said N-th sum data to register, as said original information, one of said first through said N-th feature data that has said maximum sum data. l / Ί
