Capturing User Intention for Efficient Image Retrieval
Largely used web-search engines mostly rely on the textual based image search. This provides us with random and noisy results as outcome. Even the image based search does not give satisfactory results. With even the combinations of both the textual and image based search, the user’s intention is not achieved. In this paper, we present a novel internet image search approach. It only requires the user to click on one query image with minimum effort and images from a pool retrieved by text-based search are re-ranked based on both visual and textual content. Our contribution for achieving user’s intention from one-click is in four steps; 1) The query image is categorized according to the adaptive weight categories and is re-ranked to get better results. 2) With the help of selected image by the user, the query keyword is expanded to capture user’s intention. 3) As per the expanded keyword, the image pool is expanded to get more relevant images. 4) Expanded keywords are also used to expand the query image to multiple positive visual examples from which new query specific visual and textual similarity metrics are learned to further improve content-based image re-ranking. All the above steps are automatic. No extra efforts are needed. It is extremely simple for the user to achieve their intention and to use this interface. Experimental evaluation shows that our approach significantly improves the precision of top-ranked images and also the user experience.
Keywords - Query image, Search Engines, Image Retrieval, Algorithms, Re-Ranking, User Intention, Image Search.