We introduce a simple texture synthesis method based on Support Vector Machines (SVM). Although SVM has been effectively used for various pattern recognition tasks, there is no report available on directly applying SVM for texture synthesis. The advantage of using SVM is that the sample can be simply modeled by a linear model and is not required during the synthesis stage. In addition, the method can be further extended to synthesize 3D surface texture or Bidirectional Texture Functions. Our experimental results show that the method can successfully model and synthesize semi or highly structured textures, which can be difficult subjects for previous texture synthesis methods based on parametric models.

In content-based image retrieval, the representation of local properties in an image is one of the most active research issues. This paper introduces a salient region detector based on wavelet transform. The detector can extract the visually meaningful regions on an image and reflect local characteristics. An annular segmentation algorithm based on the distribution of salient regions is designed. It takes not only local image features into account, but also the spatial distribution information of the salient regions. Then, Color moments and the introduced visual perception texture features of the regions around the salient points were computed as a features vector used for indexing the image. We tested the proposed scheme using a wide range image samples from the Corel Image Library, the experimental results indicating that the method has produced promising results.

This paper presents a new approach to measure texture directions and estimate illumination tilt angle of 3D surface textures by using mojette transform. Feature vectors are generated from variances of 72 mojette transform projections with different projection angles. The measured texture directions are compared with human perceptual judgement. Furthermore, we estimate illumination tilt angles by minimizing the Euclidean distance of the feature vector between the test image and the training sets. Experimental results show the effectiveness and accuracy of our proposed approach.

Representing the appearances of surfaces illuminated from different directions has long been an active research topic. While many representation methods have been proposed, the relationships and conversion between different representations have been less well researched. These relationships are important, as they provide (a) an insight as to the different capabilities of the surface representations, and (b) a means by which they may be converted to common formats for computer graphic applications. In this paper, we introduce a single mathematical framework and use it to express three commonly used surface texture relighting representations: surface gradients (Gradient), Polynomial Texture Maps (PTM) and eigen base images (Eigen). The framework explicitly reveals the relations between the three methods, and from this we propose a set of conversion methods. We use 26 rough surface textures illuminated from 36 directions for our experiments and perform both quantitative and qualitative assessments to evaluate the conversion methods. The quantitative assessment uses a normalized root-mean-squared error as metric to compare the original images and those produced by proposed representation methods. The qualitative assessment is based on psychophysical experiments and non-parametric statistics. The results from the two assessments are consistent and show that the original Eigen representation produces the best performance. The second best performances are achieved by the original PTM representation and the conversion between Polynomial Texture Maps (PTM) and eigen base images (Eigen), while the performances of other representations are not significantly different.

We present a new method for simultaneously determining three-dimensional (3D) motion and structure of a nonrigid object from its uncalibrated two-dimensional (2D) data with Gaussian or non-Gaussian distributions. A nonrigid motion can be treated as a combination of a rigid component and a nonrigid deformation. To reduce the high dimensionality of the deformable structure or shape, we estimate the probability distribution function (PDF) of the structure through random sampling, integrating an established probabilistic model. The fitting between the observations and the estimated 3D structure will be evaluated using the pooled variance estimator. The recovered structure is only available when the 2D feature points have been properly corresponded over two image frames. Applications of the proposed method to both synthetic and real image sequences are demonstrated with promising results.