Automatic rating of movies using an arousal curve extracted from video features

Author

Tan, Daniel Stanley ; See, Solomon ; Tiam-Lee, Thomas James

Author_Institution

Coll. of Comput. Studies, De La Salle Univ. - Manila, Manila, Philippines

fYear

2014

fDate

12-16 Nov. 2014

Firstpage

1

Lastpage

6

Abstract

This paper discusses the extraction of film structure features from action films to build an arousal curve. The arousal curve is used as training data for building a Hidden Markov Model for predicting the rating of a movie. Evaluation of the model resulted in a 70% accuracy, which shows that there is some form of correlation between the structure of a film and its perceived rating. Interesting similarities were also observed in the arousal curve patterns between different movies in the same classifications.

Keywords

entertainment; feature extraction; hidden Markov models; image classification; video signal processing; action films; arousal curve pattern extraction; automatic movie rating; film structure; film structure feature extraction; hidden Markov model; model evaluation; movie classification; movie rating prediction; perceived rating; training data; video features; Educational institutions; Feature extraction; Films; Hidden Markov models; Motion pictures; Predictive models; Rhythm; Automatic movie rating; arousal curve; image and video processing; motion; rhythm; sound; video content modeling;

fLanguage

English

Publisher

ieee

Conference_Titel

Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment and Management (HNICEM), 2014 International Conference on

Conference_Location

Palawan

Print_ISBN

978-1-4799-4021-9

Type

conf

DOI

10.1109/HNICEM.2014.7016211

Filename

7016211