3D Head Shape Analysis of Suspected Zika Infected Infants

Abstract

The babies infected from Zika before they are born are at risk for problems with brain development and microcephaly. 3D head images of 43 Zika cases and 43 controls were collected aiming to extract shape characteristics for diagnosis purposes. Principal component analysis (PCA) has been applied on the vaults and faces of the collected 3D images and the scores on the second principal components of the vaults and faces showed significant differences between the control and Zika groups. The shape variations from -2s to 2s illustrated the typical characteristics of microcephaly of the Zika babies. Canonical correlation analysis (CCA) showed a significant correlation in the first CCA variates of face and vault which indicated the potential of 3D facial imaging for Zika surveillance. Further head circumferences and distances from ear to ear were measured from the 3D images and preliminary results showed the adding ear to ear distances for classifying control and Zika children strengthened the abilities of tested classification models.