Page 258 - Automated Fingerprint Identification Systems (AFIS)

CHAPTER 10

CASE STUDY—DIAMONDS IN THE
ROUGH: INCREASING THE NUMBER
OF LA TENT PRINT IDENTIFICA TIONS

10.1 INTRODUCTION
The following was presented by the author at the 2002 International Associa-
tion for Identification (IAI) Educational Conference. This chapter describes
how the New York State Division of Criminal Justice Services (DCJS) doubled
the number of latent print identifications through a combination of technical
improvements and human intervention. Better matchers, coders, and databases
contributed to this success, but the greatest single cause was the management
decision to make it happen.
The New York State DCJS initiated a Statewide Automated Fingerprint Iden-
tification System (SAFIS) in 1989. DCJS is the state identification agency and
houses all the records for anyone fingerprinted under New York State law. This
encompasses all municipal governments, including the city of New York. DCJS
completes over 1 million transactions per year, and all SAFIS transactions inter-
face with a Computerized Criminal History (CCH) file. The sheer volume of
transactions requires speed, reliability, and accuracy in responses. Most crimi-
nal inquiries can be answered in approximately 30 minutes.
There are approximately 5.5 million records in the SAFIS tenprint (two
index finger) database, which is used for criminal and applicant identifications.
The 11 million image records (5.5 million records ¥ two fingers) are a mix of
inked tenprint impressions and livescan records. The percentage of tenprint
records on the database that are from inked impressions is diminishing as live-
scan is becoming the predominant method for image capture and transmission.
Nearly all of the images from New York City are taken via livescan.
A latent cognizant subset of this database, used to search latent prints, con-
tains all ten images of approximately 2.5 million records, or 25 million images.
A latent print can be searched with parameters such as geographic area or crime
type that can narrow the area or the database that is searched. Latent print
examiners can also search the entire database in a “cold search,” in which no
parameter is selected.