Image cataloging using Bayes, Function, Lazy, Rule, Tree classifier families with row mean of Fourier transformed image content

Increased use of social Networking sites has resulted into vast image data being uploaded on internet. Organizing this data appropriately for easy retrieval is a tedious task. If the data is organized based on the predefined categories, it will help to speed up the retrieval of data. In this paper an image classification technique has been proposed using Fourier transform to classify the images into predefined classes. To classify the images row mean of Fourier transformed image content are used as features using the wand dataset, the experimentation is carried on different sizes of feature vectors which are formed by taking fractional coefficients of Fourier transformed contents of image. Classification algorithm from different families such as Bayes (Naive Bayes and Bayes Net), Function (RBFNetwork and Simple Logistic), Lazy (IB1 and Kstar), Rule (Decision and Part) and Tree (BFTree, J48 Random Tree and Random Forest) are used for classification. The classification accuracy is used to compare performances of used classifiers with Fourier transform. Simple Logistic classifier with fractional coefficients of Fourier transformed images have given superior classification accuracy among the experimented variants of proposed classification technique.