Automatic indexing of comic page images for query by example based focused content retrieval

Muhammad Muzzamil Luqman (1) , Hoang Nam Ho (1) , Jean-Christophe Burie (1) , Jean-Marc Ogier (1)

1 L3I - Laboratoire Informatique, Image et Interaction - EA 2118

Automatic indexing of comic-page-image reposito- ries has evolved as an interesting research problem for graphics recognition research community. In this paper we present a sys- tem for automatically indexing the comic-page-images in order to achieve query by example (QBE) based focused content retrieval. In our system we represent the comic-page-images by attributed graphs and translate the problem of automatic indexing / QBE based focused content retrieval as a subgraph spotting problem. Our system uses an explicit graph embedding technique to embed the comic-page-image graphs into numeric feature vectors and then employs state-of-the-art machine learning tools for automatic indexing / QBE based focused content retrieval. Experimental results are presented for automatic indexing and QBE based focused content retrieval in a comic-page-image repository.

Domaines: Traitement des images [eess.IV]

Muhammad Muzzamil Luqman, Hoang Nam Ho, Jean-Christophe Burie, Jean-Marc Ogier. Automatic indexing of comic page images for query by example based focused content retrieval. 10th IAPR International Workshop on Graphics Recognition, Aug 2013, United States. ⟨hal-00944427⟩