Asymmetry Level in Cleft Lip Children Using Dendrite Morphological Neural Network

Approximately 3% of live newborn children suffer from cleft lip syndrome. Technology can be used in the medical area to help treatment for patients with Congenital Facial Abnormalities in cases with oral fissures. Facial dysmorphism classification level is relevant to physicians, since there are no tools to determine its degree. In this paper, a mobile application is proposed to process and analyze images, implementing the DLIB algorithm to map different face areas in healthy pediatric children, and those presenting cleft lip syndrome. Also, two repositories are created: 1. Contains all extracted facial features, 2. Stores training patterns to classify the severity level that the case presents. Finally, the Dendrite Morphological Neural Network (DMNN) algorithm was selected according to different aspects such as: one of the highest performance methods compared with MLP, SVM, IBK, RBF, easy implementation used in mobile applications with the most efficient landmarks mapping compared with OpenCV and Face detector.