Face identification of characters in movies is useful as it provides information about the content of the movie. It's a difficult process since each of the characters may appear differently in different scenes. The proposed method is robust since the noises in each scene are eliminated. Each of the faces are tracked, clustered and saved. External scripts are utilized to identify the characters. This paper therefore introduces a new method for considering each frame one by one for identification of characters in each scene. It considers frame by frame detection of each character and also uses file storage method as the basic image storing mechanism. It provides an excellent result since external scripts are utilized in order to name each character that appears in the movie.
Cluster, Character identification, External script, Face recognition, Storing mechanism
