BROWSE
Detail
Conference Paper
a Video Classification Algorithm Using Appearance Feature and Group of Frame Feature for Contents Filtering Software
Cited - time in 
Share 
Share
- Authors
- Ho Gyun Lee, Seung Min Lee, Taek Yong Nam
- Issue Date
- 2006-11
- Citation
- International Conference on Hybrid Information Technology (ICHIT) 2006, pp.1-10
- Language
- English
- Type
- Conference Paper
- Abstract
- The illegal movie sharing is increased due to the supply of the portable multimedia device like the Video i-pod and P2P file sharing. It becomes the social problem that the porno video accessing of the minors using illegal movie sharing. The internet filtering software based on image and text was developed already. However, the video classification algorithm which can be commercialized was not developed. In this paper, the video classification algorithm which can be used for harmful contents filtering software is devised. For this, the image classification algorithm based on appearance feature and the video classification algorithm based on Group of Frame feature are used. Each classification performances are compared and a performance of the combination algorithm of two algorithms is found out. For the exact harmfulness determination of one scene, the different features including the shape information and skin color information, and etc. has to be assembled but the only Group of Frame feature is adequate for the fast harmfulness determination of the obscene video.
- KSP Keywords
- Appearance features, Classification algorithm, Combination algorithm, Skin color information, Social problems, Video classification, shape information