Re-Ranking Web Images

Image Search engines mostly use s and they rely on text for searching images. The search based on only s fire by the users is not efficient and may results in imprecise output. The major bottleneck is likely mismatch between the image content and the given text. Image search reranking attempts to resolve this bottleneck by relying on both the text information and visual information during the image search process. This paper reviews the trends in image search reranking, discusses the problems found along the way, and identifies the ideas that should guide future activities in this research. According to the paper by Xiaogang Wang, Shi Qiu, Ke Liu, and Xiaoou Tang, images are re-ranked by comparing their semantic signatures and the query [1]. The query-specific semantic signatures, in the reviewed paper, improve both the accuracy and efficiency of the re-ranking process. In literature, experimental result shows that by using the re-ranking technique they can improve the web image search. In this reviewed paper, different types of web image search re-ranking techniques are explained.