TAD-RR: Tag-Annotation Based Demand Re-ranking Approach for Personalized Image Retrieval

Every user has an individual background and a particular aim while searching the information or images on the Web. The goal of Web search personalization is to retrievce the search results to a particular user depends on that user´s preferences. These data allows the users not only to facilitate and share the data, but to afford information and improve media retrieval and management. One such examples is personalized image retrieval where user´s explicit information is considered to produce exact images using re-ranking results. In this paper, the Tag-Annotation based approach by Re-ranking (TAD-RR) is proposed to generate personalized images. The basic assertion is to collect the images from the user and store it in a database with appropriate tag and annotation. Thereby, semantic search is developed to compare the query with dictionary to find out the relevant tag name. The images are then found from the database by applying TAD-RR procedures. Finally, page refreshing is handled at the time of image downloading. If a particular image is downloaded, the demand will be increased for that image in the server and the page gets reloaded based on the retrieved image annotation by ranking similarly. The experimental results show that the proposed approach accurately retrieves the personalized images.