WEB Image Classification using Classifier Combination

This paper presents a novel method for the classification of images that combines information extracted from the images and contextual information. The main hypothesis is that contextual information related to an image can contribute in the image classification process. Web pages containing images and text were collected and stored in an organized and structured fashion to build a database. First, independent classifiers were designed to deal with images and text. From the images were extracted several features like color, shape and texture. These features combined form feature vectors which are used together with a neural network classifier. On the other hand, contextual information is processed and used together with a Naive Bayes classifier. At the end, the outputs of both classifiers are combined through different rules. Experimental results on a database of more than 5,000 images have shown that the combination of classifiers provides a meaningful improvement (about 16%) in the correct image classification rate relative to the results provided by the neural network based image classifier which does not use contextual information.