The Animation and Comics Content Retrieval Model Based on Analysis of Clustered Group
Xin Lü, Mao-Quan Zhang · 2010
In content-based multimedia data retrieval model, relying solely on cluster analysis blind search retrieval model has poor robustness, low recall rate problems. In order to address these problems, this paper proposed a new retrieval model for multimedia material. Combining with the features of animation and comics material, the model introduces a clustered group analyzing method which obeys the instruction of background-knowledge. Utilizing the clustered group, we can extract the effective semantic character of objective image. It aims to realize robustness, low-dimension and rapidly-converging so as to achieve high-quality retrieval.