Figure S7. 2D and 3D radar distributions of expression-feature
mutual information (MI). The MI quantitatively measures the mutual
dependency between the distribution of 3D and 2D facial features
(dpi-pj ) and the distribution of expression
classes (0 or 1), classified through the MLP. The MI values from the 3D
facial features exhibit a distinct and high mutual dependence for each
facial expression, whereas those from the 2D features have an overlapped
and low mutual dependence between other facial expressions except
sadness. The four dominant mutual information values of both 3D and 2D
facial features are normalized. More specific, the top 10 dominant
features of each expression are as follows:
Happiness - dp0-p6 ,dp9-p10 , dp6-p10 ,dp5-p10 , dp7-p10 ,dp0-p7 , dp2-p12 ,dp2-p7 , dp9-p11 ,