Turner Syndrome Prognosis with Facial Features Extraction and Selection Schemes

Abstract

Turner syndrome adheres serious health-related complications with a tendency to affect various organs during different stages of life which includes hypertension, infertility, and retarded growth. The proper diagnosis of TS requires an expensive test named karyotype test which is not easily available in remote health care units in the countryside. Therefore, we proposed to use facial images to detect TS to pursue a higher accuracy of recognition. The proposed scheme achieved the accuracy of 91.3% with mixed feature extraction schemes using thirty principle components selected with criteria that retained 95% of the information from the turner dataset. Moreover, this research is the first that uses facial features to accurately diagnose TS patients and has the capability to help doctors to establish a cost-effective TS prognosis process in remote health care units that lack required health care facilities.