A Content Based Model for Image Categorization

Recent years have seen a rapid increase in the size of digital image collections. Any advance in our ability to organize unlabelled images according to their semantic content is a very useful step in managing these collections. In this paper we present a content based model for image categorization. In this model images in the database are grouped into classes of images with similar visual patterns. The proposed model uses two visual features, color and spatial location information of segmented image regions. Also, the proposed model is designed for using in applications that need image classification based on the general structures of images such as scene classification. Experiments are conducted and results are presented on a set of 1000 images.