A system for recommending tags of images using co-occurrence of tags and similar images

Abstract

In this paper, we propose a system for recommending tags of images. The proposed method presents to its user various tags with high precision by taking into account both the co-occurrence of tags and tags of similar images. Additional search with the user feedback enables us to find some new tags relevant to the image of interest. In order to test the proposed method, we compare it with a method based only on the co-occurrence of tags, as well as a method based only on the similarity of images. From the experimental results using 16002 images, we have confirmed that the proposed method is capable of finding more tags as compared to the methods for comparison. We have also confirmed that the additional search is effective to find additional tags. © 2009 The Institute of Electrical Engineers of Japan.