Personalized Image Recommendation

Authors:
Yuli Gao;Hangzai Luo;Jianping Fan
Affiliations:
CS Department, UNC-Charlotte, USA;CS Department, UNC-Charlotte, USA;CS Department, UNC-Charlotte, USA
Venue:
MMM '09 Proceedings of the 15th International Multimedia Modeling Conference on Advances in Multimedia Modeling
Year:
2009

Citing 2
Cited 0

Exploratory search: from finding to understanding

Communications of the ACM - Supporting exploratory search
Integrating Concept Ontology and Multitask Learning to Achieve More Effective Classifier Training for Multilevel Image Annotation

IEEE Transactions on Image Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

We have developed a novel system to support personalized image recommendation via exploratory search from large-scale collections of weakly-tagged Flickr images. First, topic network is automatically generated to index and summarize large-scale collections of Flickr images at a semantic level. Hyperbolic visualization is used to allow users to navigate and explore the topic network interactively, so that they can gain insights of large-scale Flickr image collections at the first glance, build up their mental query models quickly and specify their queries more precisely by selecting the visible image topics directly. Second, the most representative images are automatically recommended according to their representativeness for a given topic and they are visualized according to their inherent visual similarity contexts, so that users can assess the diverse visual similarity contexts between the images interactively and evaluate the relevance between the recommended images and their real query intentions effectively. Our experiments on large-scale weakly-tagged Flickr image collections have obtained very positive results.