Transferring 3D facial expressions from adults to children

Diagnosis of craniofacial conditions is shifting towards pre- and peri-natal stages, since early assessment has shown to be crucial for the effective treatment functional developmental aspects children. 3D Morphable Models are a valuable tool such evaluation. However, limited data availability on newborn geometry, highly variable imaging environments, challenge construction baby face models. Our hypothesis that constructing bi-linear model allows identity expression decoupling, enables improve brain function assessments. Thus, given adult infants facial configurations very similar expressions in babies difficult scanned controlled manner, we propose transferring from available FaceWarehouse (FW) database scans, construct baby-specific model. First, defined spatial mapping between BabyFM FW. Then, an automatic neutralization remove scans. Finally, apply transfer obtain complete tensor. We test performance generalization resulting with set. Results show obtained allow us successfully realistically manipulate while keeping them decoupled variations.