Table 3 Cross validated syndrome classification results using different facial representations and classifier models.

From: Comparing 2D and 3D representations for face-based genetic syndrome diagnosis

Facial Representation

Model

Mean Top-1 Sensitivity (%)

Mean Top-3 Sensitivity (%)

3D Surface

PCA + MLP

40.7 (0.9)

61.6 (2.1)

2D Color

ArcFace + MLP

34.2 (0.9)

53.0 (1.4)

2D Color

CNN

32.6 (1.1)

53.4 (1.7)

2D Colorless

CNN

26.3 (3.2)

46.6 (3.7)

Depth

CNN

24.8 (2.9)

42.1 (4.7)

  1. Top-1 and top-3 syndrome sensitivity scores were averaged across the 43 syndrome classes. Standard deviations across cross validation folds are shown in parentheses.
Back to article page