Development of an Image Classifier for Clustering News Resources

The growth of information volumes on the Internet emphasizes the need for effective clustering of news content, particularly the images accompanying texts. The article presents an image classifier with low sensitivity to the expansion of databases, which enables improved efficiency in searching for identical images within datasets growing daily by up to 12,000 items. The study employs methods of mathematical modeling, image processing, and content-based retrieval. Experimental results confirm the speed and cost-effectiveness of the proposed approach. Bibliogr.: 28 titles.