¡¡Chinese Journal of Computers   Full Text
  TitleImage Annotation Based on Graph Learning
  AuthorsLU Han-Qing LIU Jing
  Address(Institute of Automation, Chinese Academy of Sciences, Beijing 100190)
  Year2008
  IssueNo.9(1629¡ª1639)
  Abstract &
  Background 		Abstract Image annotation is an important and challenging task in image retrieval. This paper discusses the annotation process theoretically by reviewing some related work, and proposes a unified annotation framework via graph learning. The framework includes two sub-processes, i.e., basic image annotation and annotation refinement. In the basic annotation process, the image-based graph learning is utilized to obtain the candidate annotations. In the annotation refinement process, the word-based graph learning is used to refine those candidate annotations from the prior process. This paper also proposes some improvements on sub-problems involved in the framework and expect their combination to enhance the overall performance. Finally, experiments conducted on the Corel dataset and Web image dataset demonstrate the effectiveness of the unified framework and the proposed improvements.
Keywords image annotation£» graph learning£» image similarity; word correlation
Background Image annotation has been an active research topic in recent years due to its potential impact on both image understanding and Web image search. All kinds of annotation approaches are proposed. As they seem to be different from each other, it is not easy to answer such questions as which models are better, what the connections among them are, and how they should be utilized. In this paper, the authors conduct a formal study on these issues and find that previous research work can be explained in a unified framework. The framework offers some potential guidance on the study of image annotation, and some improvements under the framework can achieve positive effect.
The research was supported by the National Natural Science Foundation of China (60723005) and the National High Technology Research and Development Program (863 Program) of China Project (2006AA01Z315).