Jeremy S. De Bonet

This paper outlines a technique for treating input texture images as probability density estimators from which new textures, with similar appearance and structural properties, can be sampled. In a two-phase process, the input texture is first analyzed by measuring the joint occurrence of texture discrimination features at multiple resolutions. In the second phase, a new texture is synthesized by sampling successive spatial frequency bands from the input texture, conditioned on the similar joint occurrence of features at lower spatial frequencies. Textures synthesized with this method more successfully capture the characteristics of input textures than do previous techniques.

In many optimization problems, the structure of solutions reflects complex relationships between the different input parameters. For example, experience may tell us that certain parameters are closely related and should not be explored independently. Similarly, ex-perience may establish that a subset of parameters must take on particular values. Any search of the cost landscape should take advantage of these relationships. We present MIMIC, a framework in which we analyze the global structure of the optimization land-scape. A novel and efficient algorithm for the estimation of this structure is derived. We use knowledge of this structure to guide a randomized search through the solution space and, in turn, to re-fine our estimate ofthe structure. Our technique obtains significant speed gains over other randomized optimization procedures. 1

The observed distribution of natural images is far from uniform. On the contrary, real images have complex and important structure that can be exploited for image processing, recognition and analysis. There have been many proposed approaches to the principled statistical modeling of images, but each has been limited in either the complexity of the models or the complexity of the images. We present a non-parametric multi-scale statistical model for images that can be used for recognition, image de-noising, and in a "generative mode" to synthesize high quality textures. Accepted Advanced in Neural Information Processing 10 (1997). 1 Introduction In this paper we describe a multi-scale statistical model which can capture the structure of natural images across many scales. Once trained on example images, it can be used to recognize novel images, or to generate new images. Each of these tasks is reasonably efficient, requiring no more than a few seconds or minutes on a workstation. The sta...

We describe a technique for using the joint occurrence of local features at multiple resolutions to measure the similarity between texture images. Though superficially similar to a number of "Gabor" style techniques, which recognize textures through the extraction of multi-scale feature vectors, our approach is derived from an accurate generative model of texture, which is explicitly multiscale and non-parametric. The resulting recognition procedure is similarly non-parametric, and can model complex non-homogeneous textures. We report results on publicly available texture databases. In addition, experiments indicate that this approach may have sufficient discrimination power to perform target detection in synthetic aperture radar images (SAR).

This paper examines the problem of reconstructing a voxelized representation of 3D space from a series of im-ages. An iterative algorithm is used to find the scene model which jointly explains all the observed images by determin-ing which region of space is responsible for each of the ob-servations. The current approach formulates the problem as one of optimization over estimates of these responsibil-ities. The process converges to a distribution of responsi-bility which accurately reflects the constraints provided by the observations, the positions and shape of both solid and transparent objects, and the uncertainty which remains. Reconstruction is robust, and gracefully represents regions of space in which there is little certainty about the ex-act structure due to limited, non-existent, or contradicting data. Rendered images of voxel spaces recovered from syn-thetic and real observation images are shown. 1