The Brodatz Album has become the de facto standard for evaluating texture algorithms, with hundreds of studies having been applied to small sets of its images. This paper compares two powerful recognition algorithms, principal components analysis and multiscale autoregressive models, by evaluating them on a 999image database derived from the entire Brodatz Album. The variety of homogeneous and non-homogeneous images studied is thus nearly an order of magnitude larger than has been compared before, giving one snapshot of the "state of the art" in real-time texture recognition. 1 Introduction Image recognition applications are shifting rapidly from traditional areas of target recognition and satellite imagery to new areas in multi-media image/video analysis and retrieval of visual information. Many of the old applications can be typified by having a small number of "classes" of patterns, e.g., wheat, grass, water, and a large availability of training samples of each. In contrast, many ...

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). 1 Introduction The notion of texture is difficult to capture formally. Textures usually can be described informally as the output of some physical process wherein local structure is repeated seemingly at random. Two texture patches are consider...

The average person with a computer will soon have access to the world's collections of digital video and images. However, unlike text which can be alphabetized or numbers which can be ordered, image and video has no general language to aid in its organization. Although tools which can "see" and "understand" the content of imagery are still in their infancy, they are now at the point where they can provide substantial assistance to users in navigating through visual media. This paper describes new tools based on "vision texture" for modeling image and video. The focus of this research is the use of a society of low-level models for performing relatively high-level tasks, such as retrieval and annotation of image and video libraries. This paper surveys our recent and present research in this fast-growing area. 1 Introduction: Vision Texture Suppose you have a set of vacation photos of Paris and the surrounding countryside, and you accidentally drop them on the floor. They get out of or...

In this paper we propose a technique for segmenting images by texture content with application to indexing images in a large image database. Using a quad-tree decomposition, texture features are extracted from spatial blocks at a hierarchy of scales in each image. The quad-tree is grown by iteratively testing conditions for splitting parent blocks based on texture content of children blocks. While this approach does not achieve smooth identification of texture region borders, homogeneous blocks of texture are extracted which can be used in a database index. Furthermore, this technique performs the segmentation directly using image spatial-frequency data. In the segmentation reported here, texture features are extracted from the wavelet representation of the image. This method however, can use other subband decompositions including Discrete Cosine Transform (DCT), which has been adopted by the JPEG standard for image coding. This makes our segmentation method extremely applicable to databases containing compressed image data. We show application of the texture segmentation towards providing a new method for searching for images in large image databases using “Query-by-texture.” 1.

Image texture is useful in image browsing, search and retrieval. A texture descriptor based on a multiresolution decomposition using Gabor wavelets is proposed. The descriptor consists of two parts: a perceptual browsing component (PBC) and a similarity retrieval component (SRC). The extraction methods of both PBC and SRC are based on a multiresolution decomposition using Gabor wavelets. PBC provides a quantitative characterization of the texture's structuredness and directionality for browsing application, and the SRC characterizes the distribution of texture energy in different subbands, and supports similarity retrieval. This representation is quite robust to illumination variations and compares favorably with other texture descriptors for similarity retrieval. Experimental results are provided.

Textures are classified based on the change in their properties with changing resolution. The area of the gray level surface is measured at several resolutions. This area decreases at coarser resolution since fine details that contribute to that area disappear. Fractal properties of the picture are computed from the rate of this decrease in area, and are used for texture comparison and classification. The relation of a texture picture to its negative, and directional properties, are also discussed.

We present the global k-means algorithm which is an incremental approach to clustering that dynamically adds one cluster center at a time through a deterministic global search procedure consisting of N (with N being the size of the data set) executions of the k-means algorithm from suitable initial positions. We also propose modifications of the method to reduce the computational load without significantly affecting solution quality. The proposed clustering methods are tested on well-known data sets and they compare favorably to the k-means algorithm with random restarts.

The recent development of computing hardware has resulted in a rapid increase of visual information such as databases of images. To successfully utilize this increasing amount of data, we need eoeective ways to process it. Content-based image retrieval utilizes the visual content of images directly in the process of retrieving relevant images from a database. The retrieval is based on visual features such as the colors, textures, shapes, and spatial relations the image contains rather than traditional textual keywords. These features are usually extracted automatically, without the need for a human operator. In the literature survey part o...

A method of rotation-invariant texture classification based on a complete space-frequency model is introduced. A polar, analytic form of a two-dimensional (2-D) Gabor wavelet is developed, and a multiresolution family of these wavelets is used to compute information-conserving microfeatures.From these microfeatures a micromodel, which characterizes spatially localized amplitude, frequency, and directional behavior of the texture, is formed. The essential characteristics of a texture sample, its macrofeatures, are derived from the estimated selected parameters of the micromodel. Classification of texture samples is based on the macromodel derived from a rotation invariant subset of macrofeatures. In experiments, comparatively high correct classification rates were obtained using large sample sets. Index Terms---Gabor filters, texture classification, wavelets. I. INTRODUCTION T HE SPECTRUM of texture analysis techniques ranges from those focusing on structural features to those emph...

We address a new and rapidly growing application, automated searching through large sets of images to find a pattern "similar to this one." Classical matched filtering fails at this problem since patterns, particularly textures, can differ in every pixel and still be perceptually similar. Most potential recognition methods have not been tested on large sets of imagery. This paper evaluates a key recognition method on a library of almost 1000 images, based on the entire Brodatz texture album. The features used for searching rely on a significant improvement to the traditional Karhunen-Lo'eve (KL) transform which makes it shift-invariant. Results are shown for a variety of false alarm rates and for different subsets of KL features. 1 Introduction As vastly increasing amounts of image and video are stored in computers it becomes harder for humans to locate a particular scene or video clip. It is currently impossible, in the general case, to semantically describe an image to the computer...