Bidirectional Texture Function Compression based on Multi-Level Vector Quantization


Vlastimil Havran1 Jiří Filip2 Karol Myszkowski3
1Faculty of Electrical Engineering, Czech Technical University in Prague
2Institute of Information Theory and Automation of the AS CR
3MPI Informatik, Saarbrucken, Germany

Published by Computer Graphics Forum, volume 29, issue 1, pages 175-190, March 2010,
final verstion by online access Wiley-Blackwell [PDF online] [BibTeX],
preliminary version of the publication in [PDF]


Abstract

The Bidirectional Texture Function (BTF) is becoming widely used for accurate representation of real-world material appearance. In this paper a novel BTF compression model is proposed. The model resamples input BTF data into a parametrization, allowing decomposition of individual view and illumination dependent texels into a set of multidimensional conditional probability density functions. These functions are compressed in turn using a novel multi-level vector quantization algorithm. The result of this algorithm is a set of index and scale code-books for individual dimensions. BTF reconstruction from the model is then based on fast chained indexing into the nested stored code-books. In the proposed model, luminance and chromaticity are treated separately to achieve further compression. The proposed model achieves low distortion and compression ratios 1:233-1:2040, depending on BTF sample variability. These results compare well with several other BTF compression methods with predefined compression ratios, usually smaller than 1:200. We carried out a psychophysical experiment comparing our method with LPCA method. BTF synthesis from the model was implemented on a standard GPU, yielded interactive framerates. The proposed method allows the fast importance sampling required by eye-path tracing algorithms in image synthesis.

Compression Algorithm Conceptual Scheme



Supplemental Materials:


(1) Research Report

V. Havran, J. Filip, K. Myszkowski: Bidirectional Texture Function Compression Based on Multi-Level Vector Quantization SUPPLEMENTAL MATERIAL, Research Report 2265, Institute of Information Theory and Automation, Academy of Sciences of the Czech Republic, 2009. [PDF] [BibTex]

(2) Images and Movies - LDR and HDR materials

LDR samples - point and environment lighting:
Note: LPCA is in the resolution 128x128 only, the original data and the proposed method in the resolution 256x256.
alucorduroy
movie movie movie movie
fabric01fabric02
movie movie movie movie
foil01foil02
movie movie movie movie
impallaleather02
movie movie movie movie
propostepulli
movie movie movie movie
wallpaperwool
movie movie movie movie
wood01wood02
movie movie movie movie
HDR samples - point lighting:
ceilingHDRfloortileHDR pinktileHDRwalkwayHDR
movie movie movie movie

(3) Source code for the BTFBASE project for decompression is available.

The distribution BTFBASEShader V1.0 allows real-time GPU visualization of selected BTF samples (source codes + project files to MS VC++, compressed samples parameters, objects of sphere and bunny included).

Creative
                                                                   Commons
                                                                   License
BTFBASEShaderV1.0 by J. Filip is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported License.
Updated 9/March/2011