Title :
Automatic analysis of movies for content characterization
Author :
Doudpota, Sher Muhammad ; Guha, Sumanta
Author_Institution :
Sch. of Eng. & Technol., Asian Inst. of Technol., Bangkok, Thailand
Abstract :
This paper describes an approach to automatically detect music and song sequences in movies. Particularly, the system is designed to identify songs in movies, as well as their genre, playback singer and on-screen performing actors. We define the so-called song grammar that a musical sequence typically follows by means of a finite state machine. The observation of 20 most popular movies shows that 98% of the songs in these movies are based on grammar defined in this paper, hence, it can be applied for song sequences detection.
Keywords :
data analysis; entertainment; finite state machines; music; automatic movies analysis; content characterization; finite state machine; music sequences; song grammar; song sequences; Acoustic noise; Automata; Information analysis; Layout; Motion pictures; Multiple signal classification; Speech; Thyristors; Action; content characterization; dialogue; finite state machine; movie; song;
Conference_Titel :
Networking and Information Technology (ICNIT), 2010 International Conference on
Conference_Location :
Manila
Print_ISBN :
978-1-4244-7579-7
Electronic_ISBN :
978-1-4244-7578-0
DOI :
10.1109/ICNIT.2010.5508472
