Transductive cartoon retrieval by multiple hypergraph learning

Publication Type:
Conference Proceeding
Citation:
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2012, 7665 LNCS (PART 3), pp. 269 - 276
Issue Date:
2012-11-19
Metrics:
Full metadata record
Files in This Item:
Filename Description Size
Thumbnail2012004180OK.pdf739.17 kB
Adobe PDF
Cartoon characters retrieval frequently suffers from the distance estimation problem. In this paper, a multiple hypergraph fusion based approach is presented to solve this problem. We build multiple hypergraphs on cartoon characters based on their features. In these hypergraphs, each vertex is a character, and an edge links to multiple vertices. In this way, the distance estimation between characters is avoided and the high-order relationship among characters can be explored. The experiments of retrieval are conducted on cartoon datasets, and the results demonstrate that the proposed approach can achieve better performance than state-of-the-arts methods. © 2012 Springer-Verlag.
Please use this identifier to cite or link to this item: