NCJ Number
57719
Date Published
1976
Length
8 pages
Annotation
THIS REPORT REINTRODUCES THE CONCEPT OF DESCRIBING A FINGERPRINT AS AN N-DIMENSIONAL VECTOR, THUS ALLOWING ONE TO THINK OF A FINGERPRINT AS A POINT IN HYPERSPACE.
Abstract
RIDGE SLOPE MEASUREMENT OR OTHER DESCRIPTIVE MEASUREMENT ON A FINGERPRINT ARE COMBINED TOGETHER TO FORM AN N-DIMENSIONAL VECTOR, THUS ALLOWING THE FINGERPRINT TO BE CONSIDERED AS A POINT IN N-DIMENSIONAL SPACE. THE VARIATION OF A FINGERPRINT AFTER IT HAS BEEN INK ROLLED AND PROCESSED A NUMBER OF TIMES IS ASSUMED TO BE GAUSSIAN OR A SIMILAR VARIATION. IF THIS VARIATION IS KNOWN, THE PROBABILITY OF MISCLASSIFYING A FINGERPRINT CAN BE DETERMINED BY KNOWING THE DISTANCE RELATIONSHIP BETWEEN IT AND STANDARD PRINTS. EACH STANDARD PRINT REPRESENTS A FINGEPRINT CLASS. THE REPORT SHOWS THAT THE PROBABILITY CALCULATIONS ARE INDEPENDENT OF THE DIMENSIONALITY OF SPACE. BASED ON PREVIOUS RESEARCH, THE REPORT SHOWS THAT THE PROBABILITY OF MISCLASSIFYING A FINGERPRINT IS A LINEAR FUNCTION OF THE PERCENTAGE OF THE PRINT MISSING. USING THESE PROBABILITY CONSIDERATIONS, THE PROBABILITY OF MISCLASSIFYING A PRINT, THUS THE PROBABILITY OF MISCLASSIFYING A CARD, CAN BE DETERMINED. FILE ORGANIZATION AND SEARCHING STRATEGIES ARE DEVELOPED BASED ON THE PROBABILITY CONSIDERATION. THIS PROCESS ALLOWS ONE TO DETERMINE THE DEPTH OF SEARCH WHICH MUST BE PERFORMED ON A FILE EITHER BEFORE ONE HAS IDENTIFIED THE UNKNOWN CARD OR IS ASSURED THAT THE CARD CANNOT BE IDENTIFIED IN THE FILE. SEVERAL EXPERIMENTS ARE SUGGESTED TO PROVE THE VALIDITY OF THIS CONCEPT. FIGURES AND REFERENCES ARE INCLUDED. SEE ALSO NCJ 38864, 52569, 57707-57718, AND 57720-57726. (AUTHOR ABSTRACT MODIFIED--PRG)