INDIVIDUAL IDENTIFICATION METHOD/DEVICE AND RECORDING MEDIUM

摘要

PROBLEM TO BE SOLVED: To improve an individual identification rate without dropping an other people removal rate in terms of probability by calculating an individual threshold based on a feature data score distribution extracted from data showing the physical feature of the individual of a reference source and a reference score distribution between feature point data extracted from data showing the physical feature of the individual of the reference source and non-pair feature point data. SOLUTION: Digitized feature point data is supplied to a registered feature point selection means 22 and feature point data which are adopted for plural times are referred to. Data whose reference score is the highest is preserved in a feature point data storage part 31 as registered feature point data. It is referred to feature point data which is previously registered. The maximum value of the reference score with non-pair feature point data is calculated. The average of the reference score and a standard deviation are calculated and are transmitted to an individual threshold calculation means 23 with feature point data preserved in the feature point data storage part 31. The individual threshold calculation means 28 calculates the individual threshold by using the maximum value of the reference score with non-pair feature point data.