TEST RESULTS - ADVANCED DEVELOPMENT MODELS OF BISS (BASE AND INSTALLATION SECURITY SYSTEM) IDENTITY VERIFICATION EQUIPMENT, V 4 - AUTOMATIC FINGERPRINT IDENTIFICATION

• NCJ Number: 57685
• Author(s): P BENSON
• Date Published: 1978
• Length: 98 pages
• Contract Number(s): 19628-77-C-0001
• Sponsoring Agency: US Dept of the Air Force
• Corporate Author: Mitre Corporation
• Sale Source: National Technical Information Service
• Publication Type: Statistics
• Language: English
• Country: United States of America

Annotation

RESULTS OF TESTING THE DEVELOPMENT MODEL FOR FINGERPRINT IDENTITY VERIFICATION TO DETERMINE TWO TYPES OF ERROR RATES ARE PRESENTED.

Abstract

THREE AUTOMATED IDENTITY VERIFICATION SYSTEMS FOR USE IN ENTRY CONTROL ARE EVALUATED, INCLUDING THE AUTOMATIC SPEAKER VERIFICATION (ASV) SYSTEM, THE AUTOMATIC HANDWRITING VERIFICATION (AHV) SYSTEM, AND THE AUTOMATIC FINGERPRINT VERIFICATION (AFV) SYSTEM. ALL THREE WERE ACQUIRED UNDER THE BASE AND INSTALLATION SECURITY SYSTEM (BISS) PROGRAM, AND WERE TESTED BY THE MITRE CORPORATION. THE PURPOSE OF THE PROGRAM IS TO DETERMINE THE TYPE I ERROR RATE (FALSE REJECTION OF AUTHORIZED PERSONNEL), AND RANDOM TYPE II ERROR RATE (FALSE ADMITTANCE OF UNAUTHORIZED PERSONNEL. (RESULTS OF ASV AND AHV TESTING ARE PRESENTED IN VOLUMES II AND III OF THIS DOCUMENT.) ANALYSIS OF THE TEST RESULTS FOR AFV REVEAL THAT THE TYPE I ERROR RATE DOES NOT MEET THE BISS REQUIREMENT OF ONE PERCENT. IN ADDITION, THE TYPE I RATE IS SIGNIFICANTLY HIGHER IN THE MORNING THAN IN THE AFTERNOON, IS DEPENDENT ON SCAN NUMBER, AND IS NOT EVENLY DISTRIBUTED AMONG THE USER POPULATION. THE AFV TYPE II ERROR RATE DOES NOT MEET THE BISS REQUIREMENT OF TWO PERCENT, AND DOES NOT SHOW AN INCREASE WITH TIME SINCE ENROLLMENT. TYPE II ERROR RATES WILL BE HIGHER FOR AN IMPOSTER WHO KNOWS THE PRIMARY AND ALTERNATE FINGER OF HIS VICTIM, AND IS INDEPENDENT OF SCAN NUMBER. AS WITH TYPE I, TYPE II ERROR RATES ARE NOT EVENLY DISTRIBUTED AMONG THE USER POPULATION. AVERAGE VERIFICATION TIME IS APPROXIMATELY 8.7 SECONDS. FINALLY, THE PERSONNEL STATISTICS OF AGE, HEIGHT, BIRTH PLACE, AND AREA OF PRIMARY EDUCATION DO NOT EXHIBIT ANY CONSISTENT INFLUENCE ON THE ERROR RATES. STATISTICAL TABULAR DATA, GRAPHS, CHARTS, DIAGRAMS, APPENDICES, AND A BRIEF REFERENCE LIST ARE PROVIDED. FOR RELATED DOCUMENTS, SEE NCJ # 57683, 57684, AND 57686. (LWM)