Table 3 Probe/perpetrator similarity Euclidean values to the lineup images in each video
From: Application of artificial intelligence to eyewitness identification
| Â | Clear/White | Unclear/White | Clear/Black | Unclear/Black | Clear/Hispanic | Video 6 unclear/Hispanic |
|---|---|---|---|---|---|---|
Probe versus guilty suspect | 0.8759 | 1.0740 | 0.8387 | 1.0122 | 1.2415 | 1.1231 |
Probe versus image 2 | 1.2342 | 1.2859 | 1.0387 | 1.2232 | 1.4467 | 1.2652* |
Probe versus image 3 | 1.3474 | 1.3318 | 1.1547 | 1.2503* | 1.1473 | 1.2791 |
Probe versus image 4 | 1.2398* | 1.1821* | 1.2915 | 1.4450 | 1.3645* | 1.4459 |
Probe versus image 5 | 1.2717 | 1.3981 | 1.1154* | 1.2667 | 1.1586 | 1.4535 |
Probe versus image 6 | 1.2998 | 1.2208 | 1.1932 | 1.4921 | 1.4552 | 1.2728 |
Probe versus image 7 | 1.2906 | 1.2254 | 1.1042 | 1.3032 | 1.2347 | 1.2026 |