DocumentCode :
2952060
Title :
A Textural Based Hidden Markov Model for Animation Genre Discrimination
Author :
Santarcangelo, Joseph ; Zhang, Xiao-Ping
Author_Institution :
Ryerson Univ., Toronto, ON, Canada
fYear :
2012
fDate :
9-13 July 2012
Firstpage :
552
Lastpage :
557
Abstract :
This paper develops a novel method to automatically categorize different animation genres in a video database made for children, this is the first such research done in animation genre categorization. The method is based on statistically modeling the temporal texture attributes of the video sequence. The features are extracted from gray-level co-occurrence matrices and a hidden Markov models (HMM) are used as a classifier. It was found the method had 16.66% better accuracy compared to other methods with the same number of parameters and dimensions of feature vector.
Keywords :
computer animation; hidden Markov models; image sequences; image texture; video signal processing; visual databases; HMM; animation genre discrimination; feature vector dimension; gray level cooccurrence matrices; statistically modeling; temporal texture attributes; textural based Hidden Markov model; video database; video sequence; Accuracy; Animation; Equations; Feature extraction; Hidden Markov models; Image color analysis; Mathematical model; Indexing; Video Classification; Video Retrieval;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Multimedia and Expo Workshops (ICMEW), 2012 IEEE International Conference on
Conference_Location :
Melbourne, VIC
Print_ISBN :
978-1-4673-2027-6
Type :
conf
DOI :
10.1109/ICMEW.2012.102
Filename :
6266443
Link To Document :
بازگشت