Compression and Real-Time Rendering of Measured BTFs Using Local PCA

T. Ertl, B. Girod, G. Greiner, H. Niemann, H.-P. Seidel, E. Steinbach, and R. Westermann (Editors)
In proceedings of Vision, Modeling and Visualisation 2003, pages 271-280, Akademische Verlagsgesellschaft Aka GmbH, Berlin, Nov. 2003
Presented at The 8th International Fall Workshop Vision, Modeling and Visualisation 2003
 

Abstract

The Bidirectional Texture Function (BTF) is a suitable representation for the appearance of highly detailed surface structures under varying illumination and viewing conditions. Real-time rendering of measurements from this six-dimensional function requires approximation strategies, because of the huge size of the dataset.

In this paper we present a framework for BTF-compression and rendering enabling high-quality real-time rendering using much less memory than other comparable data-driven approaches. Our method exploits a BRDF-wise arrangement of the data and employs a flexible generalization of Principal Component Analysis (PCA) named local PCA for the data compression.

Keywords: BTF rendering, data compression, material representation, real-time rendering, textures

Images

Download Paper

Download Paper

Additional Material

Bibtex

@INPROCEEDINGS{mueller-2003-compression,
      author = {M{\"u}ller, Gero and Meseth, Jan and Klein, Reinhard},
      editor = {Ertl, T. and Girod, B. and Greiner, G. and Niemann, H. and Seidel, H.-P. and Steinbach, E. and
                Westermann, R.},
       pages = {271--280},
       title = {Compression and Real-Time Rendering of Measured BTFs Using Local PCA},
   booktitle = {Vision, Modeling and Visualisation 2003},
        year = {2003},
       month = nov,
   publisher = {Akademische Verlagsgesellschaft Aka GmbH, Berlin},
    keywords = {BTF rendering, data compression, material representation, real-time rendering, textures},
    abstract = {The Bidirectional Texture Function (BTF) is a suitable representation for the appearance of highly
                detailed surface structures under varying illumination and viewing conditions. Real-time rendering
                of measurements from this six-dimensional
                function requires approximation strategies, because of the huge size of the dataset.
                
                In this paper we present a framework for BTF-compression and rendering enabling high-quality
                real-time rendering using much less memory than other comparable data-driven approaches. Our method
                exploits a BRDF-wise arrangement of the data and employs a flexible generalization of Principal
                Component Analysis (PCA) named local PCA for the data compression.},
        isbn = {3-89838-048-3},
  conference = {The 8th International Fall Workshop Vision, Modeling and Visualisation 2003}
}