Abstract
Computer-aided design (CAD) systems are known to be used in manufacturing, modern engineering design and modeling. New applications to this technology have been created in other fields, such as biomedicine and health informatics, due to the remarkable improvements achieved in recent years. This paper proposes a new biomedical computer-aided design (Bio-CAD) system based on an optimized kernel principal component analysis (OKPCA) for brain tumor diagnosis entitled CAD-OKPCA. The concept of this method consists of reducing the complexity involved in the medical images by selecting only the relevant features using the OKPCA, while maintaining good classification rates. Three databases have been used to validate the proposed CAD-OKPCA method and the results were satisfactory.