Page 193 - Automated Fingerprint Identification Systems (AFIS)
P. 193

178  AUTOMATED FINGERPRINT IDENTIFICATION SYSTEMS



         Figure 8.3
                                                          Remaining
         Prints Remaining After                           latent prints
         Elimination Prints                               examined in
                                                            HQ
                                     Agency A                                  Agency B

                                    Of 60 prints,        Of value             Of 100 prints,
                                    50 are “of value”    determination        70 are “of value”
                                    5 cases                                   7 cases


                                  60 lifts examined at headquarters      100 lifts examined at headquarters
                                  10 have no value                       30 have no value
                                  50 lifts remain                        70 lifts remain


                                 Time for a statistical recap. Both agencies went to ten crime scenes and
                              looked at 100 latent images. Agency A considered 20 from two crime scenes to
                              be of no value and did not process them any further. The number of latent
                              images retrieved from the eight remaining crime scenes is 80. This is the base
                              for any further comparisons. Agency B did not make any quality determination
                              at the ten crime scenes and retained all 100 latent images. Their base for further
                              comparisons is 100 latents from ten crime scenes. Agency A has 20% fewer
                              latent images to search than Agency B, and their images are of reasonably good
                              quality. The quality of each of the 100 images for Agency B is unknown. The
                              denominator for Agency A is 80, and the denominator for Agency B is 100.
                              With the use of the elimination print process, however, Agency A reduced this
                              denominator to 60. Agency A could also claim a 25% identification rate (20
                              elimination prints identified from the group of 80, i.e., 20 divided by 80).
                              Agency B has an identification rate of 0%.
                                 Back at headquarters, both teams used photographic technology for a better
                              view of the latent images. The technicians from Agency A removed an addi-
                              tional ten images that were of no value. The Agency B technicians removed 30
                              images that were not of value. The denominator of Agency A is now 50, and
                              the denominator of Agency B is now 70.
                                 Next, Agency A searches the remaining 50 latent images on the AFIS systems,
                              while Agency B searches their 70 prints. As shown in Figure 8.4, each agency
                              makes 25 identifications that close three cases.
                                 What is the resulting ident or hit rate for each agency? There is no simple
                              answer. The hit rate for Agency A may be

                              • 56% of latents (20 elimination print idents plus 25 AFIS idents from 80 latent
                                images)
                              • 62% of cases (45 idents in five of the eight cases)
   188   189   190   191   192   193   194   195   196   197   198