Visualization approach based on multimedia clustering

Multimedia is one kind of the most important resources on the current Web. To search and browse the multi-media content efficiently, it is necessary to improve the multimedia retrieval and classification approaches. For a multimedia snippet, the annotated tags and introduction texts around the multimedia can reflect the content and characteristics of the multimedia. This paper proposes an efficient approach to cluster and visualize the multimedia by constructing semantic field with a set of multimedia tags. A semantic field is a circle region with several gravitation sources (tags)to locate a multimedia by calculating the positions in the field. We have developed an experiment on audio clustering for about 1000 multimedia resources including images, audios, and videos using the semantic field approach, and the result shows the approach is efficient.

[1]  Marcel Worring,et al.  Content-Based Image Retrieval at the End of the Early Years , 2000, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  Alexander Hauptmann,et al.  Meta-Classification of Multimedia Classifiers , 2002, KDMCD.

[3]  Takeo Kanade,et al.  Human Face Detection in Visual Scenes , 1995, NIPS.

[4]  Huimin Zhao,et al.  Semantic matching across heterogeneous data sources , 2007, Commun. ACM.

[5]  Takeo Kanade,et al.  Neural Network-Based Face Detection , 1998, IEEE Trans. Pattern Anal. Mach. Intell..

[6]  Sharath Pankanti,et al.  An identity-authentication system using fingerprints , 1997, Proc. IEEE.

[7]  P. L. Venetianer,et al.  An Iris Biometric System for Public and Personal Use , 2000, Computer.

[8]  Herbert Gish,et al.  GMM sample statistic log-likelihoods for text-independent speaker recognition , 1997, EUROSPEECH.

[9]  Takeo Kanade,et al.  Rotation Invariant Neural Network-Based Face Detection , 1998, Proceedings. 1998 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.98CB36231).