IJRCS – Volume 2 Issue 3 Paper 1


Author’s Name : D Dhayalan | R Keerthana

Volume 02 Issue 03  Year 2015  ISSN No:  2349-3828  Page no: 1-5


Abstract – Multimedia re-ranking, as an efficient way to improve the search results of web-based image search, has been adopted by recent commercial search engines such as yahoo and Google. Given a query keywords, a pool of images are first retrieved based on text based information. A major challenge is that the similarities of visual features do not correlated with images semantic meanings which interpret users search intention. Currently people proposed to match images in a semantic space which is used attributes or reference classes closely related to the semantic meanings of images as basic. A universal visual semantic space to highly diverse images from the web is difficulty and effectiveness. In this paper, we propose a Bayesian re-ranking framework which automatically offline learns different semantic signatures.  At the online stage, multimedia images are re-ranked by comparing their semantic signature obtained from the semantic spaces specified by the query keywords. The proposed Query-specific semantic signatures significantly improve both the accurate and efficient of multimedia re-ranking.

Keywords –Image search, Image re-ranking, Semantic signature, and Semantic spaces, image retrieving