Efficient Symbolic Signatures for Classifying Craniosynostosis Skull Deformities

Craniosynostosis is a serious and common pediatric disease caused by the premature fusion of the sutures of the skull. Early fusion results in severe deformities in skull shape due to the restriction of bone growth perpendicular to the fused suture and compensatory growth in unfused skull plates. Calvarial (skull) abnormalities are frequently associated with severe impaired central nervous system functions due to brain abnormalities, increased intra-cranial pressure and abnormal build-up of cerebrospinal fluid. In this work, we develop a novel approach to efficiently classify skull deformities caused by metopic and sagittal synostoses using our newly introduced symbolic shape descriptors. We demonstrate the efficacy of our methodology in a series of large-scale classification experiments that compare the performance of our symbolic-signature-based approach to those of traditional numeric descriptors that are frequently used in clinical research. We also demonstrate an application of our symbolic descriptors in shape-based retrieval of skull morphologies.