Destekçi vektör makinesi kullanarak resim sınıflandırma

Image categorization has become more and more important in the last decade with the development of Internet, digital cameras becoming widespread and the growth in the size of image databases. Image categorization task consists of categorizing real-world natural scenes based on different features. The objective is to regroup images into semantically meaningful categories.Computer vision researchers have been working to design computational systems that are capable of automatic scene categorization. A computational system that can perfectly mimic the human visual system and perception in order to categorize images is still missing.In this work, the categorization task is accomplished using Support Vector Machines (SVM) that has been applied to many real-world problems producing state-of-the-art results. These include text categorization, biological data mining and handwritten character recognition. In other words SVM is a very effective method for general purpose pattern recognition and classification.For an effective use of a classification algorithm, the data that is the subject to the classification has to be represented in a suitable way. We insisted on image representation using local, global and intermediate representations in order to obtain good results and take full advantage of SVM.