AUTOMATIC WRITER RECOGNITION (FROM CARNAHAN CONFERENCE ON CRIME COUNTERMEASURES - PROCEEDINGS, MAY 16-18, 1979, BY JOHN S JACKSON - SEE NCJ-62284)

• NCJ Number: 62292
• Author(s): W KUCKUCK; B RIEGER; K STEINKE
• Date Published: 1979
• Length: 8 pages
• Corporate Author: Ores Publications
• Publication Format: Document
• Language: English
• Country: United States of America
• Note: *This document is currently unavailable from NCJRS.

Annotation

THE AUTOMATIC IDENTIFICATION OF PERSONS BASED ON HANDWRITING WAS THE OBJECTIVE OF A RESEARCH PROJECT CONDUCTED BY A FORENSIC SCIENCE GROUP IN THE FEDERAL REPUBLIC OF GERMANY.

Abstract

THE HANDWRITING ANALYSIS, WITH PRIMARY EMPHASIS ON WRITER RECOGNITION, WAS CONCERNED WITH WHETHER TWO SAMPLES OF HANDWRITING HAD BEEN ORIGINATED BY THE SAME WRITER (VERIFICATION) AND WHETHER A PARTICULAR SAMPLE OF HANDWRITING WAS IDENTIFIED BY A KNOWN WRITER (IDENTIFICATION). TO ANSWER THE LATTER QUESTION, THE FEDERAL CRIMINAL POLICE GATHERED APPROXIMATELY 80,000 HANDWRITING SAMPLES OF OFFENDERS SPECIALIZING IN CHECK OR HOTEL FRAUD, IN WRITING THREATENING LETTERS, AND IN SIMILAR CRIMES. AN IMAGE PROCESSING SYSTEM LINKED WITH TWO PROCESS CONTROL COMPUTERS HELPS CLASSIFY PATTERNS OF HANDWRITING IMAGES. THE SYSTEM HANDLES ALL PARTS OF THE RESEARCH PROJECT, I.E., IMAGE READING, IMAGE PROCESSING, FEATURE EXTRACTION, FEATURE COMPRESSION, AND CLASSIFICATION. INITIAL CLASSIFICATION EXPERIMENTS, BASED ON A 'HOLD ONE OUT' ALGORITHM ADMITTING NO REJECTIONS, GAVE RECOGNITION RATES OF UP TO 96 PERCENT FOR 720 HANDWRITING IMAGES FROM 20 WRITERS. SUPPORTING DATA, ILLUSTRATIONS, AND HANDWRITING SAMPLES ARE PROVIDED. REFERENCES ARE CITED. (DEP)