Abstract:Large-scale user-contributed images with texts are rapidly increasing on the social media websites, such as Sina microblog. However, the noise and incomplete correspondence between the images and the texts give rise to the difficulty in precise image retrieval and ranking. In this paper, a hypergraph-based learning framework is proposed for image ranking, which simultaneously utilizes visual feature, textual content and social link information to estimate the relevance between images. Representing each image as a vertex in the hypergraph, complex relationship between images can be reflected exactly. Then updating the weight of hyperedges throughout the hypergraph learning process, the effect of different edges can be adaptively modulated in the constructed hypergraph. Furthermore, the popularity degree of the image is employed to re-rank the retrieval results. Comparative experiments on a large-scale Sina microblog data-set demonstrate the effectiveness of the proposed approach.

A Unified Framework of Image Latent Feature Learning on Sina Microblog