Personalized tag recommendation system using deep learning

Abstract

The development of smart devices and digital cameras, the amount of multimedia information posted on social multimedia platforms is rapidly increasing. User tags are either recommending images to similar interested users or deteriorate image recommendations and searches due to easy-to-search but inaccurate tags or spam tags. Many tag-recommend studies are being conducted to address. This paper proposes a tag recommendation technique using deep learning to recommend personalized tags for input images. Existing personalized tag recommendation studies recommend personalized tags through profiling methods, but profiling contained unnecessary information. To solve this problem, this paper only utilizes the user's existing tags to find categories of interest to users and recommend tags corresponding to those categories. The experimental results show the superiority of the proposed system and the personalization custom tag results.