BROWSE
Detail
Conference Paper
PCA/LDA based Image Analysis for 3D Graphic Shader Retrieval
Cited - time in 
Share 
Share
- Authors
- Hee-Kwon Kim, Jae-Ho Lee, Seung-Woo Nam
- Issue Date
- 2010-02
- Citation
- Korea-Japan Joint Workshop on Frontiers of Computer Vision (FCV) 2010, pp.247-250
- Language
- English
- Type
- Conference Paper
- Abstract
- This paper proposed an image analysis method that based on PCA/LDA (Principal Component Analysis/Linear Discriminant Analysis) for 3D graphic shader retrieval. Our previous shader retrieval system was based on similar shader search which utilized query shader. However, a query shader also had to be prepared for the shader retrieval in the previous method. For this reason, requests for a new shader search mechanism has increased from content productions. In this paper, we provided the new shader search method that adapts images for shader retrieval. Basically, the 3-D shaders and the images have totally different structures, and it makes direct comparison impossible on retrieval. We solve this problem using PCA/LDA based classification with rendered shader images. In the developed system, shader retrieval is activated from the classification of input images. The features for shader retrieval is divided as color, material, and pattern.
- KSP Keywords
- Different structures, Principal Component analysis, Search mechanism, Three dimensional(3D), direct comparison, image analysis method, linear discriminant analysis(LDA), retrieval system, search method