

CALIFORNIA IAI STATISTICAL COMMITTEE REPORT 2001 Summation reported in the July 2002 CSDIAI Identification Digest. CALIFORNIADOJ CENTRAL SITE LATENT DATABASE FOR THE YEAR OF 2000: DATABASE FINGERS  42,391,980 Latents
searched  125,732 Average
candidate list  10 This report is more defining than the simple reported numbers. The California database is comprised of NEC minutiae extraction based upon an algorithm that emulates the way the human makes a fingerprint comparison. The algorithm not only plots the type and directionality of level two minutia on an xy coordinate. It also counts the number of intervening ridges between minutiae. ANALYSIS: The historical statistical breakdown of fingerprint pattern types applied to the total number of fingers in this latent searchable fingerprint database would be as follows: FINGERS  42,391,980 Avg.
60% loop
25,435,188 Avg.
35% whorl 14,837,193 Avg. 5% arch 2,119,599 The historical statistical breakdown of fingerprint pattern types applied to the total latent searches would be as follows: LATENTS
SEARCHED
125,732 Avg.
60% loop
75,439 Avg.
35% whorl
44,006 Avg. 5% arch 6,286 The total number
of fingers
searched would be somewhere between only searching single latent pattern calls
to total pattern reference. This would equate as follows: No reference 2,585,054,462,004
(75,439 x 25,435,188) + (44,006 x 14,837,193) + (6,286 x 2,119,599) (125,732 x 42,391,980) Between 2.5 and 5.3 trillion individual latent searches were made by California DOJ in the year 2000. The computerized searches utilized an algorithm that emulates how a human would make a level two comparison of minutia type, flow, and direction in an xy coordinate related to the pattern core, counting the number of intervening ridges between the minutiae. The 2.55.3 trillion searches provided list of candidates of the ten fingerprints with the closest agreement, to a human for Analysis, Comparison, and Evaluation. For the 125,732 latents searched by California DOJ, a human latent fingerprint analyst examined the top 1,257,320 candidates. These examinations found no two different fingerprints that contained twelve or more level two Galton detail of the same type, direction, ridge flow, xy axis location, and ridge count spatial relationship. No where near twelve level two Galton detail were found. The past limited statistical models, as well as Dr. Edmond Locard's tripartite rule, have been validated by the California DOJ latent database, supporting quantifiable thresholds for friction ridge individualization.
