Learning to tag
- 20 April 2009
- conference paper
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 361-370
- https://doi.org/10.1145/1526709.1526758
Abstract
Social tagging provides valuable and crucial information for large-scale web image retrieval. It is ontology-free and easy to obtain; however, irrelevant tags frequently appear, and users typically will not tag all semantic objects in the image, which is also called semantic loss. To avoid noises and compensate for the semantic loss, tag recommendation is proposed in literature. However, current recommendation simply ranks the related tags based on the single modality of tag co-occurrence on the whole dataset, which ignores other modalities, such as visual correlation. This paper proposes a multi-modality recommendation based on both tag and visual correlation, and formulates the tag recommendation as a learning problem. Each modality is used to generate a ranking feature, and Rankboost algorithm is applied to learn an optimal combination of these ranking features from different modalities. Experiments on Flickr data demonstrate the effectiveness of this learning-based multi-modality recommendation strategy.Keywords
This publication has 22 references indexed in Scilit:
- Supporting OLAP operations over imperfectly integrated taxonomiesPublished by Association for Computing Machinery (ACM) ,2008
- Flickr tag recommendation based on collective knowledgePublished by Association for Computing Machinery (ACM) ,2008
- Query expansion using probabilistic local feedback with application to multimedia retrievalPublished by Association for Computing Machinery (ACM) ,2007
- Semantics, content, and structure of many for the creation of personal photo albumsPublished by Association for Computing Machinery (ACM) ,2007
- Bipartite graph reinforcement model for web image annotationPublished by Association for Computing Machinery (ACM) ,2007
- Correlative multi-label video annotationPublished by Association for Computing Machinery (ACM) ,2007
- Dual cross-media relevance model for image annotationPublished by Association for Computing Machinery (ACM) ,2007
- Why we tagPublished by Association for Computing Machinery (ACM) ,2007
- Random sampling from a search engine's indexPublished by Association for Computing Machinery (ACM) ,2006
- Incremental formalization of document annotations through ontology-based paraphrasingPublished by Association for Computing Machinery (ACM) ,2004