Web Image Search.Image search re – Ranking is an effective approach to refine the text-based image search result. Most existing re – Ranking approaches are based on low-level visual features. In this paper, we propose to exploit semantic attributes for image search re ranking.Its basic principle is that visually similar images should have similar ranking scores.

 

In this paper, we propose a visual-attribute joint hypergraph learning approach to simultaneously explore two information sources.A hypergraph is constructed to model the relationship of all images. We conduct experiments on more than 1,000 queries in MSRA-MMV2.0 data set. The experimental results demonstrate the effectiveness of our approach.

With the dramatic increase of online images, image retrieval has attracted significant attention in both academia and industry. The local patterns or salient features discovered using graph analysis are very powerful to improve the effectiveness of rank lists.

 Thus, attributes are expected to narrow down the semantic gap between low-level visual features and high-level semantic meanings.

The high level semantic concepts which are crucial to capture property of images could deliver more clear semantic messages between various nodes in the graph.

Furthermore, attribute-based image representation has also shown great promises for discriminative and descriptive ability due to intuitive interpretation and cross-category generalization property.

Thus, in this paper, we propose to exploit stronger semantic relationship in the graph for image search reranking.