A medical decision support system based on Gabor fisher classifier for evaluation of Down syndrome affecteds

Today, one of the most frequently encountered dysmorphic diseases, Down syndrome, includes different findings belonging to the face. Early diagnosis of Down syndrome is vital for individual's further life. The databases, which were benefited by making comparisons on pre-diagnosis of Down syndrome, are not objective as doctor's experiences come into prominence. To speed up pre-diagnosis and eliminate subjectivity, computer-assisted objective methods are needed. In this research, belonging to children's pictures with Down syndrome and normal operation of two separate groups are tried to distinguish on database groups. As feature, filter responses obtained as a result of Gabor Wavelet Transform (GWT) are used. Dimension reduction is provided with Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) on the reaching higher dimension feature vectors. Classification process was carried out with kth Nearest Neighbor (kNN) and Support Vector Machines (SVM) which has been successfully used in pattern recognition applications.