Retrieval of movie character using semantic graph matching techniques

In today´s scenario identification based updating are on demand, especially automatic face identification. By this identification we can collect the related data efficiently, means it works like a data collector based on identification. The face distinguishing mark of notation in filter has tense magnanimous log in study interests and lead to many interesting applications. There are several areas where the above notation is leading including movies. It is a strenuous problem due to the huge variation in the appearance of each character and the potential mismatch between images and their captions. Instead of the fact that realized methods prevail upon sparing in besom mood, the performances are not up to the presumptions because of the noises generated during the face tracking and face clustering process. In this paper, global face-name matching scheme instead of the local matching scheme is proposed. The global matching is done between faces from video and names from text script file of movie. We have proposed a hybrid frame work which is based on Principle Component Analysis classification and VF2 graph matching for improving the robustness and reduce the mismatch between images and captions respectively. The proposed system gives better results in terms of robustness in identifying characters in less time as compared with existing systems.