Intelligent Web Mining for Semantically Adequate Images

Web mining is good established technology when we must search data in the form of plain text. For this common purpose many well designed algorithms an technologies (like ontologies) has been developed and adopted. Other situation is when the most important information, which must be searched and retrieved from multimedial databases, is presented in the form of digital image. Meanwhile we observe more and more important information which is obtained, collected, stored and retrieved in form of the many kind of images. Moreover the information is represented not only by the form of objects presented on the image, but also can be hidden in complex form in relations between objects. Therefore for achieve successful web mining for semantically adequate images we must have tools suitable for automatic understanding of image merit content. In the paper we describe the new method of automatic understanding of the images, proved in the medical images practice and ready for use for other kind of images.