Abstract
The functional categorization of Web images using data collected from news Web sites was discussed. The effectiveness of the image features in separating graphic images and photographic images was tested. Experimentation was carried out with both linear and the Radial Basis Function (RBF) kernels for the Support Vector Machines (SVM). The results indicate that the RBF kernels performed better than the linear kernel for both the intermediate frequency domain classifier and the final image classifiers. Image-reconition techniques to interpret images more effectively in the icon/logo category were also explored.