CONCEPTS FOR A NEXT GENERATION AUTOMATED FINGERPRINT SYSTEM (FROM CARNAHAN CONFERENCE ON CRIME COUNTERMEASURES - PROCEEDINGS, 1978, BY JOHN S JACKSON - SEE NCJ-61021)

• NCJ Number: 61034
• Author(s): J A RATKOVIC; F W BLACKWELL; H H BAILEY
• Date Published: 1978
• Length: 5 pages
• Corporate Author: Ores Publications
• Publication Format: Document
• Publication Type: Statistics
• Language: English
• Country: United States of America

Annotation

A TECHNIQUE DEVELOPED BY THE RAND CORPORATION TO ESTIMATE THE PROBABILITY OF A MATCH BETWEEN TWO DIGITAL IMAGES OF TERRAIN IN NEAR REAL-TIME AND TO FACILITATE AN AUTOMATED FINGERPRINT SYSTEM IS NOTED.

Abstract

PATTERN RECOGNITION, A TECHNICAL TERM FOR THE COMPARISON OF FINGERPRINTS ON THE BASIS OF SIMILARITIES BETWEEN EXTRACTED FEATURES, IS EMPLOYED IN MOST SEMIAUTOMATED FINGERPRINT-MATCHING SYSTEMS. CORRELATION INVOLVES THE DIRECT COMPARISON OF FINGERPRINTS BY SUPERIMPOSING AND MEASURING THE DEGREE OF CONGRUENCE. SUBAREA CORRELATION IS THE DIVIDING OF A FINGERPRINT INTO A NUMBER OF SMALL PARTS, CORRELATING EACH INDIVIDUALLY AND INTEGRATING THE RESULTS SO OBTAINED. DUE TO SHORTCOMINGS IN THESE TECHNIQUES, THE FINGERPRINT MATCHING PROCESS IS SUGGESTED FOR IMPLEMENTATION IN AN AUTOMATED FINGERPRINT SYSTEM. THE PROCESS USES CORRELATION TO GENERATE CERTAIN SCENE STATISTICS AND THEN WORKS BACKWARD FROM THESE DATA THROUGH THEORETICAL RELATIONSHIPS TO ESTIMATE OTHER SCENE CHARACTERISTICS FROM WHICH ESTIMATES OF A MATCH BETWEEN TWO DIGITAL IMAGES OF TERRAIN ARE OBTAINED. THE MATCHING TECHNIQUE HAS BEEN TESTED USING DIGITIZED IMAGERY OF TERRAIN OBTAINED BY REMOTE SENSORS. IT HAS BEEN ABLE TO DISTINGUISH THE CASE OF TWO IMAGES WHICH ARE NEARLY IDENTICAL FROM THE CASE IN WHICH IMAGES DO NOT MATCH AT ALL. THE ANALOGY BETWEEN THE USE OF THE TECHNIQUE IN IMAGE PROCESSING AND ITS USE IN FINGERPRINT APPLICATIONS IS CLOSE ENOUGH TO WARRANT AN EXAMINATION OF ITS EFFECTIVENESS. SPECIFICALLY, THE TECHNIQUE SHOULD PROVE USEFUL IN DEFINING THE SITUATION WHERE THE FINGERPRINT IN QUESTION IS NOT IN THE MASTER FILE AND IN DECREASING OR EVEN ELIMINATING MANUAL SEARCH TIME INVOLVED IN FINDING A MATCH AFTER AN AUTOMATED PRINT CLASSIFICATION PROCESS HAS REDUCED THE FINGERPRINT FILE DOWN TO A MANAGEABLE SIZE. A TOTAL AUTOMATED SYSTEM DESIGN, INCLUDING SUBAREA CORRELATION TO ACCOMMODATE ERROR SOURCES, IS DISCUSSED. ILLUSTRATIONS AND REFERENCES ARE PROVIDED. (AUTHOR'S ABSTRACT MODIFIED--DEP)