Bidirectional Texture Function Compression Based on Multi‐Level Vector Quantization

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 multi-dimensional 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.

WebGL-based streaming and presentation of objects with bidirectional texture functions

“…Additionally, this technique has the considerable advantage that the texture mapping units of the GPU can be utilized to perform interpolation both in the angular and spatial domain. This reduces the decompression costs considerably in comparison with techniques using clustering [Müller et al 2003], sparse representations [Ruiters and Klein 2009] or vector quantization [Havran et al 2010], to name just a few. There are also tensor factorization based approaches [Wu et al 2008], which are, like our wavelet compression, capable of further compressing the spatial dimensions.…”

Effects of Approximate Filtering on the Appearance of Bidirectional Texture Functions

“…Tsai et al [7] further pushes the idea to a k-clustered tensor approximation, which is suitable for efficient real-time rendering of the compressed BTFs. Havran et al [8] compressed the BTF by adopting a multidimensional conditional probability density function in conjunction with vector quantization. Mipmapping was handled by directly applying the same algorithm to averaged BTF data.…”

Interactive high fidelity visualization of complex materials on the GPU

“…The major limitation with matrix-based approaches is that the high dimensional structure of the BTF data set is not exploited on the compression stage; matrices are two-dimensional and only correlations among columns are being exploited. Havran et al [6] decompose individual ABRDFs into multidimensional conditional probability density functions, which are then further compressed using multi-level vector quantization. The proposed approach achieves high compression ratios, supports mip-mapping and importance sampling for Monte Carlo based renderers, and the authors report rendering rates of up to 170 fps on the GPU for point based lighting.…”

“…The efficiency of these compressed approaches, together with improvements in BTF acquisition systems and the increasing computational power of GPUs, contributed to the increasing popularity of BTF based rendering systems [1,3,4]. However, most of these highly efficient compressed representations do not allow intuitive editing of the BTF data [5,6,7], as required by product designers. Interactive global illumination rendering with ordinary computing resources using an intuitively editable representation of the BTF remains an elusive goal.…”

“…Alternative representations based on factorizing the BTF tensor (or subsets thereof) [5,6,7] have the potential to achieve better frame rates, since they only require evaluating dot products. The SPMM representation has, however, the advantage of allowing direct intuitive editing of the BTF.…”