DocumentCode
2917090
Title
Fast video analysis by genetic programming
Author
Song, Andy
Author_Institution
Sch. of Comput. Sci. & Inf. Technol., RMIT Univ., Melbourne, VIC
fYear
2008
fDate
1-6 June 2008
Firstpage
3237
Lastpage
3243
Abstract
Genetic programming has been applied to various types of vision tasks. This paper extends the use of this powerful problem solving method to a more complex but more common domain, video analysis. We present the methodology as well as the experiments on two video analysis tasks: segmenting texture regions and detecting moving objects. The advantages of GP in this domain can be shown by this study. Firstly GP methods are less dependent on knowledge from domain experts. One methodology is suitable for both tasks. Secondly GP can generate fast video frame analyzers which are highly desirable or even critical in real time vision applications.
Keywords
genetic algorithms; image segmentation; image texture; object detection; video signal processing; fast video analysis; genetic programming; moving object detection; texture region segmentation; vision tasks; Data mining; Delay; Face detection; Genetic programming; Information analysis; Mobile handsets; Motion detection; Object detection; Problem-solving; Streaming media;
fLanguage
English
Publisher
ieee
Conference_Titel
Evolutionary Computation, 2008. CEC 2008. (IEEE World Congress on Computational Intelligence). IEEE Congress on
Conference_Location
Hong Kong
Print_ISBN
978-1-4244-1822-0
Electronic_ISBN
978-1-4244-1823-7
Type
conf
DOI
10.1109/CEC.2008.4631236
Filename
4631236
Link To Document