Modified support vector machines for MR brain images recognition
Anis Ladgham, Ghada Torkhani, Anis Sakly, Abdellatif Mtibaa · 2013
Support vector machine (SVM) is a popular method of learning classification with lots of applications. In this work, we extend SVM to recognize the appearance of tumors in MR brain image. Parameterization of the kernel in SVM learning procedure, along selecting features, influences the accuracy of the recognition and increases the computational effect. For this, a Shuffled Frog Leaping Algorithm (SFLA) based approach for feature selection of the SVM, termed SFLA-SVM, is developed. To demonstrate the quality of our technique, we give some experiments on MR brain images.