A Comic Retrieval System Based on Multilayer Graph Representation and Graph Mining
Résumé
Comics analysis offers a lot of interesting Content-based Image Retrieval (CBIR) applications focusing in this special type of document images. In this paper, we propose a scheme to represent and to retrieve comic-page images using attributed Region Adjacency Graphs (RAGs) and their frequent subgraphs. We first extract the graphical structures and local features of each panel in the whole comic volume, then separate different categories of local features to different layers of attributed RAGs. After that, a list of frequent subgraphs is obtained using Frequent Subgraph Mining (FSM) techniques. For CBIR purpose, the recognition and ranking are done by checking for isomorphism between the graphs representing the query image versus the list of discovered frequent subgraphs. Our experimental results show that the proposed approach can achieve reliable retrieval results for comic images browsing using query-by-example (QBE) model.