Eurographics 2009.
Jianye Lu, Yale University
Julie Dorsey, Yale University
Holly Rushmeier, Yale University
Abstract
Texture synthesis techniques require nearly uniform texture samples, however identifying suitable texture samples in an image requires significant data preprocessing. To eliminate this work, we introduce a fully automatic pipeline to detect dominant texture samples based on a manifold generated using the diffusion distance. We define the characteristics of dominant texture and three different types of outliers that allow us to efficiently identify dominant texture in feature space. We demonstrate how this method enables the analysis/synthesis of a wide range of natural textures. We compare textures synthesized from a sample image, with and without dominant texture detection. We also compare our approach to that of using a texture segmentation technique alone, and to using Euclidean, rather than diffusion, distances between texture features.
Paper and Supplementary Materials
Full-resolution PDF file (~10MB)
Low-resolution PDF file (~2MB)
Bibtex-entry
Animated rotational view of Figure 4(c) (~8MB)
Additional examples (~22MB)
EG'09 presentation slides (2MB)
Source Data
(coming soon...)
A Psychophysical Study
Following our EG'09 paper, we conducted a psychophysical experiment to
quantitatively validate our selection of diffusion distance
manifolds for dominant texture detection, and proposed a systematic
way to select appropriate detection methods for different textures.
Please see our latest technical report,
A Psychophysical Study of Dominant Texture Detection, about this work.
Acknowledgement
The authors would like to thank
Steven Gortler for commenting on an early version of the paper and
Franco Woolfe for invaluable discussions. This material is
based upon work supported by the National Science Foundation
under Grant No. 0528204.
Some of the source texture images used in this paper were provided by http://grungetextures.com and http://mayang.com/textures.
