_____________________________________________________________________
The concept of forensic identification is based on the evaluation of information. With fingerprint identification, information is analyzed and compared to available exemplar and other sources to determine if the impressions in question originated from one and the same source. Fingerprint identification information sources are generally divided into three levels. Level one is macro detail such as ridge flow and pattern type. Level two is the Galton characteristics, or points of identification, such as bifurcations and ending ridges. Level three information is contained in the structure of the ridges themselves. Of course, multiple types of forensic information can be found in a fingerprint. Not only can all three levels be present, but other information can also be available such as DNA, chemical information, as well as information contained in the distortion of the print itself. These additional sources of information are not considered as formal levels of comparison, yet they may offer additional means to allow for individualization and/or evidence correlation.
For many decades, print examiners have been asked during testimony; “what the minimum point requirement is for print identification?” At one time this was a acceptable question since many countries required a minimum number of Galton characteristics before an identification was legally accepted. However, research into the statistical study of fingerprint identification has shown that there is no statistical foundation for a minimum point requirement. Level three detail, when present in an impression, is valid for comparison since it is also permanent and unique. Accordingly, the question of how much non-Galton detail information is needed is pointless since the statistical nature of randomness is in itself unique. Likewise, the information is too complex to be effectively quantized for statistical model comparison. The detail being analyzed in a fingerprint comparison is potentially the most multifarious of any forensic science.
An understanding of the information being analyzed and how it is being analyzed is the key to deducing its value for identification purposes. Information is always found in related groups of varying content. This can be thought of as a principle of minimum information. A single bit of information is not possible due to the fact that at the very least it will have a relationship with one other bit of information such as its opposite value. The very fact that an identifying characteristic is present is information. Its minimum opposite value would be the fact that we know it is not a missing. With this reasoning the lack of information also has value. Additional information allows further information correlation at relative quality and quantity values. Information can be evidence. Its proper evaluation and correlation is imperative.
Thus, the discovery and documentation of a single characteristic generates more information than the simple fact of that characteristic’s existence. With comparative analysis, it generates information on its relative position, size, shape. etc. Uniqueness, such as that of a Galton characteristic, is simply a large grouping of information. Hence, uniqueness can be defined as; sufficient information that allows for a relative distinction or possibly an individualization.
With fingerprint identification a variable threshold for individualization can exist based on quantitative and qualitative information values. With threshold comparisons this grouping aspect of discovered information is more noticeable as the examiner focuses their attention on the limited details. Threshold fingerprint identifications are based on the evaluation of groups of related information of varying quality. Accordingly, there is never a single bit of information that would make the difference between a conclusion of individualization or non-individualization.
Most statistic models that have illustrated the concept of fingerprint identification have been limited to specific levels of detail such as Galton points. The exclusion of third level detail omits the very foundation of forensic identification. Its inclusion would further strengthen the statistical model. However, if the complexity of statistical models involving third level detail could be overcome we would still encounter the minimum information principle and its relative nature.
Following is a list of relationships of level two friction ridge characteristics. This list is expanded to 75 but this is not the limit. Palm print impressions could contain many more characteristics. Each relationship is counted only once. This is an idealized list as normal friction skin characteristics would not necessarily have direct line of sight with every single characteristics to every other characteristics, yet it does illustrate the linear increase in possible relationships. This also follows Metcalfe's Law of network nodes where ½ N(N-1). The network count is divided by 2 due to the fact that a node connection between two specific minutiae is only counted once, not twice. With forensic comparison, level one and three information would be added to this level two information for a more comprehensive analysis.
Minutiae
|
Unique Relationships
|
Ratio
|
%
|
1
|
0
|
||
2
|
1
|
||
3
|
3
|
1:1
|
|
4
|
6
|
1:1.5
|
0.666
|
5
|
11
|
1:2.2
|
0.4545
|
6
|
15
|
1:2.5
|
0.4
|
7
|
21
|
1:3
|
0.3333
|
8
|
28
|
1:3.625
|
0.2857
|
9
|
36
|
1:4
|
0.25
|
10
|
45
|
1:4.5
|
0.2222
|
11
|
55
|
1:5
|
0.2
|
12
|
66
|
1:5.5
|
0.1818
|
13
|
78
|
1:6
|
0.1666
|
14
|
91
|
1:6.5
|
0.1538
|
15
|
105
|
1:7
|
0.1428
|
16
|
120
|
1.7.5
|
0.1333
|
17
|
136
|
1:8
|
0.125
|
18
|
153
|
1:8.5
|
0.1176
|
19
|
171
|
1:9
|
0.1111
|
20
|
190
|
1:9.5
|
0.1052
|
21
|
210
|
1:10
|
0.1
|
22
|
231
|
1:10.5
|
0.0952
|
23
|
253
|
1:11
|
0.0909
|
24
|
276
|
1:11.5
|
0.0869
|
25
|
300
|
1:12
|
0.0833
|
26
|
325
|
1:12.5
|
0.08
|
27
|
351
|
1:13
|
0.0769
|
28
|
378
|
1:13.5
|
0.074
|
29
|
406
|
1:14
|
0.0714
|
30
|
435
|
1:14.5
|
0.0689
|
31
|
465
|
1:15
|
0.0666
|
32
|
496
|
1:15.5
|
0.0645
|
33
|
528
|
1:16
|
0.0625
|
34
|
561
|
1:16.5
|
0.0606
|
35
|
595
|
1:17
|
0.0588
|
36
|
630
|
1:17.5
|
0.0571
|
37
|
666
|
1:18
|
0.0555
|
38
|
703
|
1:18.5
|
0.054
|
39
|
741
|
1:19
|
0.0526
|
40
|
780
|
1:19.5
|
0.0512
|
41
|
820
|
1:20
|
0.05
|
42
|
861
|
1:20.5
|
0.0487
|
43
|
903
|
1:21
|
0.0476
|
44
|
946
|
1:21.5
|
0.0465
|
45
|
990
|
1:22
|
0.0454
|
46
|
1035
|
1:22.5
|
0.0444
|
47
|
1081
|
1:23
|
0.0434
|
48
|
1128
|
1:23.5
|
0.0425
|
49
|
1176
|
1:24
|
0.0416
|
50
|
1225
|
1:24.5
|
0.0408
|
51
|
1275
|
1:25
|
0.04
|
52
|
1326
|
1:25.5
|
0.0392
|
53
|
1378
|
1:26
|
0.0384
|
54
|
1431
|
1:26.5
|
0.0377
|
55
|
1485
|
1:27
|
0.037
|
56
|
1540
|
1:27.5
|
0.0363
|
57
|
1596
|
1:28
|
0.0357
|
58
|
1653
|
1:28.5
|
0.035
|
59
|
1711
|
1:29
|
0.0344
|
60
|
1770
|
1:29.5
|
0.0338
|
61
|
1830
|
1:30
|
0.0333
|
62
|
1891
|
1:30.5
|
0.0327
|
63
|
1953
|
1:31
|
0.0322
|
64
|
2017
|
1:31.5
|
0.0317
|
65
|
2080
|
1:32
|
0.03125
|
66
|
2145
|
1:32.5
|
0.0307
|
67
|
2211
|
1:33
|
0.0303
|
68
|
2278
|
1:33.5
|
0.0298
|
69
|
2346
|
1:34
|
0.0294
|
70
|
2415
|
1:34.5
|
0.0289
|
71
|
2485
|
1:35
|
0.0285
|
72
|
2556
|
1:35.5
|
0.0281
|
73
|
2628
|
1:36
|
0.0277
|
74
|
2701
|
1:36.5
|
0.0273
|
75
|
2775
|
1:37
|
0.027
|
Craig A. Coppock Updated: 27MAR2016
No comments:
Post a Comment