Search

Existing texture filtering techniques acquiesce in the removal of all undesired background texture patterns simultaneously, and lead to the disappearance of small-scale foreground structures. Differently, this paper considers both multi-scale structures and specified texture patterns within an image as significant image contents, and presents a content-aware texture filtering approach via an easy-to-use interactive interface, named intuitive single click. The user’s effort on content-aware texture filtering is reduced to only locating a single click, where it is intuitively considered as one type of texture patterns. This simple interface is made possible by first generating three-scale intermediate filtered images (i.e., high, medium and low) and then performing breadth-first search (BFS) to identify one or more texture regions. Specifically, we utilize BFS to purify image edges in the medium-scale filtered image, on top of which a mask representing texture and non-texture regions is abstracted through the intuitive single-click texture locating and BFS. The mask is applied to effectively maintain multi-scale structures in the low-scale filtered image, and guides to fully smooth textures in the high-scale filtered image. Finally, a clean image is yielded by simply synthesizing the results of two mask operations in the low- and high-scale filtered images. We illustrate by extensive experiments that our approach advances the state-of-the-arts in terms of both the image cleanness and content-preserving ability. Our code is publicly available.

Texture analysis plays an important role in image processing. In the texture analysis field, the regular texture has been studied a lot, but the natural texture with complex backgrounds is less studied. In this paper we bring texture analysis into the study of rice paper's classification. First of all we show the processing flow chart of rice paper classification. By comparing the different kinds of texture analysis methods we choose the LAWS texture method and uncertainty texture spectrum method to achieve the rice paper classification. When we use the two texture analysis methods separately, the classification accuracy of rice paper is lower. So we try to combine the two texture analysis methods. The experimental results show that the classification result when we combine the two texture analysis methods is better than the result when we only use one single texture analysis method. The classification accuracy of rice paper has been distinctly improved after the combination of the two texture analysis methods.

The simulation of aging appearance is essential in the realistic modeling, We propose a method using existing aging image to simulate aging appearance, We establish the classification library, extract the texture feature of different types aging and save them, set the weight of the texture feature and synthesis texture according to the environment. In addition, user can specify the constrained texture mapping position optionally. This method can simulate various rusty phenomenon simultaneously. It can also be combined with physically-based method and ton spreading-based method to implement rendering of realistic 3D model rapidly.

Semi-variogram estimators and distortion measures of signal spectra are utilized in this paper for image texture retrieval. On the use of the complete Brodatz database, most high retrieval rates are reportedly based on multiple features and the combinations of multiple algorithms, while the classification using single features is still a challenge to the retrieval of diverse texture images. The semi-variogram, which is theoretically sound and the cornerstone of spatial statistics, has the characteristics shared between true randomness and complete determinism and, therefore, can be used as a useful tool for both the structural and statistical analysis of texture images. Meanwhile, spectral distortion measures derived from the theory of linear predictive coding provide a rigorously mathematical model for signal-based similarity matching and have been proven useful for many practical pattern classification systems. Experimental results obtained from testing the proposed approach using the complete Brodatz database, and the the University of Illinois at Urbana-Champaign texture database suggests the effectiveness of the proposed approach as a single-feature-based dissimilarity measure for real-time texture retrieval.

This study evaluated image texture within a multi-scale context. The multi-scale texture profile is a representation of texture as a function of scale. It is related to the variogram, but has the advantage of being a direct measure of local texture over a specified kernel. Directional texture is a method to capture the dependence of texture on orientation using 1-dimensional kernels, and is potentially useful for quantifying texture for classes with a linear shape, such as roads or rivers. Seven derived texture attributes were proposed from the multi-scale texture profile and directional texture: the minimum, median, mean, maximum, range, minimum directional, and minimum multi-scale and directional texture. The derived texture attributes were found to be useful in developing a locally adaptive texture measure, which uses different texture attributes for different locations in the image. In order to develop the rules for the adaptive method, seven characteristic multi-scale texture profiles were identified...

Extracting meaningful structures from an image is an important task and benefits a wide range of image application tasks. However, it is typically very challenging to distinguish between noisy or textural patterns from image structures, especially when such patterns do not exhibit regularity (e.g., irregular textural patterns or those with varying scales). While existing edge-preserving image filters like bilateral, guided, or propagation filters aim at observing strong image edges, they cannot be easily applied to solve the above texture filtering tasks. In this paper, we propose robust propagated filter, which is an extension to propagation filters while exhibiting excellent ability in eliminating the aforementioned textural patterns when performing filtering. We show in our experimental results that our filter provides promising results on image filtering. Additional experiments on inverse image half toning and detail enhancement further verify the effectiveness of our proposed method.

In this letter, we propose an efficient one-nearest-neighbor classifier of texture via the contrast of local energy histograms of all the wavelet subbands between an input texture patch and each sample texture patch in a given training set. In particular, the contrast is realized with a discrepancy measure which is just a sum of symmetrized Kullback-Leibler divergences between the input and sample local energy histograms on all the wavelet subbands. It is demonstrated by various experiments that our proposed method obtains a satisfactory texture classification accuracy in comparison with several current state-of-the-art texture classification approaches.

Texture is an important image feature in image analysis, which is related to qualitative properties of surfaces and corresponds to both brightness value and pixel locations. Image texture has been introduced into a wide range of applications such as metal surface analysis, textiles characterization, ultrasonic images processing, and food qualities evaluation. One of the most common methods for texture analysis is the grey level co-occurrence matrix (GLCM), which has a large number of texture features. In this work, an investigation of the relationship between GLCM texture features and the cutting conditions in milling operations (typically, feed, speed, and depth of cut) has been carried out. A vision system was employed to capture images for specimens with various known cutting conditions; then, the images were analysed by a software, which has been fully developed in-house to calculate 22 texture features. The relationship between each texture feature and the three cutting conditions are discussed and the correlation coefficients are introduced. The results showed that 15 texture features have good correlations with the feed, nine have good correlations with the speed, while only two have good correlations with the depth of cut.

The texture analysis is often discussed in image processing domain, but most methods are limited within gray-level image or color image, and the present conception of texture is defined mainly based on gray-level image of single band. One of the essential characters of remote sensing image is multidimensional or even high-dimensional, and the traditional texture conception cannot contain enough information for these. Therefore, it is necessary to pursuit a proper texture definition based on remote sensing images, which is the first discussion in this paper. This paper describes the mapping model of spectral vector in two-dimensional image space using Markov random field (MRF), establishes a texture model of multiband remote sensing image based on MRF, and analyzes the calculations of Gibbs potential energy and Gibbs parameters. Further, this paper also analyzes the limitations of the traditional Gibbs model, prefers a new Gibbs model avoiding estimation of parameters, and presents a new texture segmentation algorithm for hyperspectral remote sensing image later.

Texture is an important image feature and has been applied greatly in the food industry for quality evaluation and inspection. Recent applications of texture-analysis techniques in the food industry were reviewed in the current paper. There are four different types of texture, i.e. statistical texture, structural texture, model-based texture, and transform-based texture. Statistical texture can be obtained using statistical approachs from the higher-order of pixel grey values of images while structural texture is acquired through some structural primitives constructed from grey values of pixels. Transform-based texture can be obtained by using statistical measurements from the images transformed with certain methods. Model-based texture can be attained by calculating coefficients from a model based on the relationship of the grey values between a pixel and its neighboring pixels. Among them, statistical texture is the most widely used one in the food industry for its high accuracy and less computation time. Transform-based texture and model-based texture are also used, although not as popular as statistical texture. However, structural texture is rarely used in the food industry. Due to the limited understanding of image texture up to now, it is difficult to identify which kind of method is the best for certain application. Therefore, it is expected that by employing different methods together, final results can be improved.

The Poisson equation is a class of partial differential equations which describe a steady-state temperature distribution in a bounded object. This paper applies this equation to the modelling of image textures by constructing specific heat source functions and boundary conditions. The heat source function can be considered as an image transform function such that a set of texture features at different frequencies and orientations can be extracted from the transformed image, in conjunction with using a Gabor wavelet filer bank. Better performance of image texture retrieval by these features is achieved than using the features extracted directly from the original image texture.

A new structural method was developed to characterise texture of colour meat images. Structural primitives or region-primitives were constructed using a region-growing method with a global and automatically calculated threshold. A total of eight textural features under the consideration of size, orientation, and contrast, which are the primary factors of human perception of texture pattern, were obtained from the region-primitives. An experiment was set up to classify the texture pattern of a set of 60 cooked meat colour images consisting of 20 beef images, 20 lamb images, and 20 pork images using texture features obtained from the proposed region-primitive method and run-length matrix method, respectively. The proposed region-primitive method (error rate: 18.3%) was found to perform better than the run-length matrix method (error rate: 35.0%). Results also indicated that the texture features obtained by the proposed method carried adequate texture information for classification and that the proposed region-primitive method could be potentially used for the characterisation of irregular textural pattern in cooked meat images, which was inapproachable by conventional structural texture analysis techniques.

Surface texture is an important quality characteristic of many products. This paper provides an overview of several different approaches to image texture analysis and demonstrates their use on the problem of classifying a set of rolled steel sheets into various quality grades. Methods covered include traditional statistical approaches such as gray level co-occurrence matrix (GLCM) methods, multivariate statistical approaches based on PCA and PLS, and wavelet texture analysis. Traditional multivariate classification approaches, such as PLS-DA, applied directly to the images are shown to fail because of the loss of spatial identity of the variables (pixels) in those approaches, and the lack of congruency of the images. However, approaches that re-introduce spatial information, such as performing two-dimensional FFT on the images prior to applying multivariate methods can perform well. A new approach that re-introduces spatial information through image shifting and stacking, followed by multivariate image analysis (MIA) is presented and shown to work well. It can also be used to develop optimal spatial filters for extracting texture information. Wavelet texture analysis (WTA) methods are discussed and insight into their space/frequency decomposition behavior is used to show why they are generally considered to be state of the art in texture analysis.

Since image processing aimed at reducing image noise can also remove important texture, standard methods for evaluating the capture and retention of image texture are currently being developed. Concurrently, the evolution of the intelligence and performance of camera noise-reduction (NR) algorithms poses a challenge for these protocols. Many NR algorithms are ‘content-aware’, which can lead to different levels of NR being applied to various regions within the same digital image. We review the requirements for improved texture measurement. The challenge is to evaluate image signal (texture) content without having a test signal interfere with the processing of the natural scene. We describe an approach to texture reproduction analysis that uses embedded periodic test signals within image texture regions. We describe a target that uses natural image texture combined with a multi-frequency periodic signal. This low-amplitude signal region is embedded in the texture image. Two approaches for embedding periodic test signals in image texture are described. The stacked sine-wave method uses a single combined, or stacked, region with several frequencycomponents. The second method uses a low-amplitude version of the IEC-61146-1 sine-wave multi-burst chart, combined with image texture. A 3x3 grid of smaller regions, each with a single frequency, constitutes the test target. Both methods were evaluated using a simulated digital camera capture-path that included detector noise and optical MTF, for a range of camera exposure/ISO settings. Two types of image texture were used with the method, natural grass and a computed ‘dead-leaves’ region composed of random circles. The embedded-signal methods tested for accuracy with respect to image noise over a wide range of levels, and then further in an evaluation of an adaptive noise-reduction image processing.

Texture analysis has been used extensively in the computer-assisted interpretation of digital imagery. A popular texture feature extraction approach is the grey level co-occurrence probability (GLCP) method. Most investigations consider the use of the GLCP texture features for classification purposes only, and do not address segmentation performance. Specifically, for segmentation, the pixels in an image located near texture boundaries have a tendency to be misclassified. Boundary preservation when using the GLCP texture features for image segmentation is important. An advancement which exploits spatial relationships has been implemented. The generated features are referred to as weighted GLCP (WGLCP) texture features. In addition, an investigation for selecting suitable GLCP parameters for improved boundary preservation is presented. From the tests, WGLCP features provide improved boundary preservation and segmentation accuracy at a computational cost. As well, the GLCP correlation statistical parameter should not be used when segmenting images with high contrast texture boundaries.

This paper proposes to use three-dimensional information derived from the graph of an image function for texture description. The graph of an image function is a rumpled surface appearing like a landscape. To characterize the texture through this landscape, six novel texture feature curves based on the statistics of the geometrical and topological properties of the solids shaped by the graph and a variable horizontal plane are used. The proposed statistical landscape features have been shown by systematic experiments to offer very low error rates on a large subset of the Brodatz texture album having excluded some nonhomogeneous images, the entire Brodatz texture set, as well as the VisTex texture collection.

Medical image segmentation is a frequent processing step in image understanding and computer aided diagnosis. In this paper, we propose medical image texture segmentation using texture filter. Three different image enhancement technique s are utilized to remove strong speckle noise as well enhance the weak boundaries of medical images. We propose to exploit the concept of range filtering to extract the texture content of medical image. Experiment is conducted on ImageCLEF2010 database. Results show the efficacy of our proposed medical image texture segmentation.

This paper discusses the various methods used to analyze the texture property of an image. Texture analysis is broadly classified into three categories: Pixel based, local feature based and Region based. Pixel based method uses grey level co occurrence matrices, difference histogram and energy measures and Local Binary Patterns(LBP) Local feature based method uses edges of local features and generalization of co occurrence matrices. Region based method uses region growing and topographic models.

In image retrieval and identification technology, image texture feature extraction and recognition is a better way. Texture is a feature which is common and difficult to describe in the image. Texture feature extraction and recognition is an important part of image retrieval study based on content. This paper utilizes Wavelet Technology, proposes a texture feature extraction and recognition algorithm based on Wavelet packet transform and singular value decomposition, and by measuring calculation distance of European space, obtain the image which is closest to sample image.

In the past one difficulty of texture analysis was the lack of adequate tools to characterize different scales of texture effectively. Recent developments in multiresolution analysis such as the Gabor and wavelet transforms, help to overcome this difficulty. This paper introduces a new approach to characterize texture at multiple scales. The performances of the wavelet transform are measured in terms of sensitivity and selectivity for the classification of 25 types of remote sensing texture relief images under the condition of different wavelet decomposition models, different filter lengths, different resolutions and different mother bodies. The reliability exhibited by texture signatures of wavelet transforms are beneficial for accomplishing segmentation, classification and subtle discrimination of texture.

This paper introduces a new computer recognition algorithm for image texture recognition. Using this algorithm, the texture characteristic of image is extracted exactly for analysis. 2-d bidirectional median filtering is proposed to make image de-noising an ideal state. An image correction algorithm with Hough transfer on local computation is presented to obtain an image with better position state. An approach based on local extreme value is proposed to search for feature points of image texture to further recognize each parameter of image texture. At last, this paper takes example for fabric density recognition, and gives the result and discusses detail process. The result shows that this approach is correct and efficient, and has extensive application.

We present a variational model for image texture identification. We first use a second order model introduced in Bergounioux and Piffet (Set Valued Variational Anal. 18(3–4):277–306 (2010)) for image denoising. The model involves a L 2-data fitting term and a Tychonov-like regularization. We choose here the BV 2 norm, where BV 2 is the bounded hessian function space (see Bergounioux and Piffet, Set Valued Variational Anal. 18(3–4):277–306 (2010)). We observe that results are not satisfying since geometrical information appears in the oscillating component and should not. So we propose an anisotropic strategy, by setting components of the discrete hessian operator to 0 in order to focus on gradient directions. We precisely describe and illustrate the numerical methodology. Finally, we propose some numerical tests.

A novel approach of blind steganalysis is proposed, which is based on image texture energy-entropy features. Image complexity describes the difference of image's content and texture. Steg-image's texture is ordinarily more complicated than that of cover image. For analyzing image complexity, using image texture features to measure the statistical differences between cover image and steg-image. In the paper, we analyze image complexity based on image texture segmentation technique, and use Laws' image texture energy-entropy features to measure the statistical differences between cover image and steg-image. Applying these texture features, blind steganalysis is implemented. Support Vector Machine (SVM) is used as classifier to distinguish whether a given image is embedded into the convert message. Experiment results show that the proposed approach is greatly valuable and our blind steganalysis method attains a good testing accurate rate.

Extraction of effective features of objects is an important area of research in the intelligent processing of image data. A well-known feature in images is texture which can be used for image description, segmentation and classification. This paper presents a novel texture extraction method using the principles of geostatistics and the concept of entropy in information theory. Experimental results on medical image data have shown the superior performance of the proposed approach over some popular texture extraction methods.

Texture analysis plays an important role in remote sensing image processing. Most methods of texture classification, for the identification of different texture surfaces, are based on wavelet features, MRF models, STFT features, and geometric shape of texels and PCA analysis. Kernel Fisher discriminant (KFD) is a state-of-the-art nonlinear machine learning method, and it has great potential to identify image texture. Meanwhile, Texture analysis is also an important technique of SAR image classification. In this paper, a nonlinear discriminative texture feature extraction method based on kernel Fisher discriminant (KFD) is proposed for SAR image texture classification.

In recent years, multi-resolution, multi-channel texture analysis algorithm get extensive attention, and become the important development direction of the texture analysis. According to the satellite cloud images, this paper put forward the cloud classification method based on texture feature of 2D-Gabor wavelet. Experiments show that the 2D-Gabor wavelet texture features can achieve better classification of clouds, take this method compared with the method of gray level co-occurrence matrix has higher classification accuracy.

In this paper, a multilinear approach based on image texture for face recognition is present. First, we extract the texture features of the facial images using the local binary pattern (LBP) algorithm. Then, we apply the high-order orthogonal iteration (HOOI) algorithm, the algebra of higher-order tensors, to obtain a compact and effective representation of the facial images based on the texture features. Our representation yields improved facial recognition rates relative to standard eigenface and tensorface especially when the facial images are confronted by a variety of viewpoints and illuminations. To evaluate the validity of our approach, a series of experiments are performed on the CMU PIE facial databases.

Modern content-based image retrieval systems use different features to represent properties (e.g., color, shape, texture) of the visual content of an image. Retrieval is performed by example where a query image is given as input and an appropriate metric is used to find the best matches in the corresponding feature space. Both selecting the features and the distance metric continue to be active areas of research. In this paper, we propose a new approach, based on the recently proposed Multidimensional Dynamic Time Warping (MD-DTW) distance [1], for assessing the texture similarity of images with structured textures. The MD-DTW allows the detection and comparison of arbitrarily shifted patterns between multi-dimensional series, such as those found in structured textures. Chaos theory tools are used as a preprocessing step to uncover and characterize regularities in structured textures. The main advantage of the proposed approach is that explicit selection and extraction of texture features is not required (i.e., similarity comparisons are performed directly on the raw pixel data alone). The method proposed in this preliminary investigation is shown to be valid by proving that it creates a statistically significant image texture similarity measure.

Image texture is important in human and machine vision. A taxonomy of image texture that classifies textures the same way humans do psychophysically can be used in many fields. This paper deals with one attribute of texture, namely orderliness. To determine what underlying factors influence humans to perceive orderliness in textures, psychophysical direct magnitude estimation ratings of the orderliness of 27 Brodatz images were collected from 44 subjects. The images: a) were either tiled, locally oriented, or granular, b) had large, medium, or small scale elements, and c) contained either high, medium, or low regularity. Multidimensional scaling revealed three underlying factors that determined the perception of texture orderliness: uniformity of element shape and distribution, element size, and element dimensionality. Some of these results contradict predictions made by earlier computational models of texture. Such models should be revised to incorporate the results of our experiment.

A texture analysis method for handwriting image based on integer-to-integer wavelet transform is proposed. As the main features of texture image are present in the details and the WT coefficients of the high-frequency part denotes the details of the image, the features deduced from WT coefficients can be used to retrieve the texture image, and this method is not sensitive to luminance, these virtues cannot be found in the traditional literature. Adopting the features extracted of such texture images employed this method as main basis, it can not only simplify the computation, and also do not lose main characteristics of handwriting images.

This paper presents a new method for texture analysis through Bidimensional Empirical Mode Decomposition (BEMD). Although there have been many filter based methods for texture analysis, problems of non-adaptively and redundancy are still hard to solve. The BEMD is a locally adaptive method and suitable for the analysis of nonlinear or nonstationary signals. The texture image can be decomposed to several IMFs (intrinsic mode functions) by BEMD, which present new characters of the images. But for the BEMD, the boundary interference is a main limit for its application. In this paper, we proposed a new BEMD method based on the self-similar extend method and the neighbor local extremes to reduce the boundary interference. This new method can get a lower orthogonality index (OI) of the IMF, which present more clearly features of the texture images. The experiment result shown the new method also reduced the computation complex compared to other surface interpolation based methods.

This paper deals with the development of a new texture analysis method based on both spatial and spectral information for texture classification purposes. The idea of the Spatial and Spectral Gray Level Dependence Method (SSGLDM) is to extend the concept of spatial gray level dependence method by assuming texture joint information between spectral bands. In addition, new texture features measurement related to (SSGLDM) which define the image properties have been also proposed. Extensive experiments have been carried out on many multi-spectral images for use in prostate cancer diagnosis and quantitative results showed the efficiency of this method compared to the Gray Level Co-occurrence Matrix (GLCM). The results indicate a significant improvement in classification accuracy.

In order to effectively extract the detail of the image characteristics so as to obtain a better identification effect, the paper proposes a texture analysis method for handwriting image based on integral wavelet transform. This method not only can extract the detail features of the texture image, but also is not sensitive to brightness, this characteristic is difficult to achieve for the traditional texture analysis method. Adopting the features extracted of such texture images employed this method as main basis, it can not only simplify the computation, and also do not lose main characteristics of handwriting images.

Texture information of tongue image is one of the most important pathological features utilized in practical Tongue Diagnosis because it can reveal the severeness and change tendency of the illness. A texture segmentation method based on Gabor filter plus Normalized Cut is proposed in this paper. This method synthesizes the information of location, color and texture feature to be the weight for Normalized Cut, thus can make satisfactroy segmentation according to texture of tongue image. The experiments show that the overall rate of correctness for this method exceeds 81%.

We address the problem of efficiency in image texture analysis. Motivated by the statistical occupancy model, we introduce the notion of patch re-occurrences. Using the re-occurrences, we propose the use of approximate textural features in image analysis. We describe how the proposed approximate features can be extracted for Gabor filters, a popular texture analysis method. Preliminary results on image texture classification show that the proposed method can provide an improved efficiency in the processing, without introducing any significant degradation in the classification results.

Texture refers to properties that represent the surface or structure of an object and is defined as something consisting of mutually related elements. The main focus in this study is to do texture segmentation and classification for texture digital images. Grey level co-occurrence probabilities (GLCP) method is being used to extract features from texture image. Gaussian support vector machines (GSVM) have been proposed to do classification on the extracted features. A popular Brodatz texture album had been chosen to test out the result. In this study, a combined GLCP-GSVM shows an improvement over GLCP in terms of classification accuracy.

It is important to consider the role of scale for texture analysis since its a multiscale attribute of image texture. A textural detector based on the 2D Gabor function and visual textural perception is established first. Then, based on the textural detector and the theory of time-scale space decomposition-a general class extending wavelet transform, an energy distribution based multiscale texture analysis method is proposed. The multiscale texture analysis gives textural energy representation between the spatial space and scale space, and provides a hierarchical analysis framework for the image texture. They can detect different scale texture features, correspond to the visual texture perception, and have the ability to recognize texture image effectively.

A new scheme for texture segmentation based on Ant Colony Systems (ACS) is proposed in this paper. Texture segmentation is one of the important branches in image pattern recognition, which provides usefulness in many applications. Until now, how to find an effective way for accomplishing texture segmentation in practical applications is still a major task. In this paper, we employ wavelet coefficients and characteristics of different subbands to serve as the basis of characteristic vectors, and we use three feature-extraction elements, namely, the extrema, entropy, and energy, to compose the characteristic vector. To alleviate segmentation fragments caused from the information in high frequency bands of texture images, we integrate the fourth element, the mean variance, into the characteristic vector. Finally, we use ACS to find a trade-off between texture segmentation and fragments. Simulation results demonstrate the effectiveness and practicability of the proposed algorithm.

This paper addresses the problem of image texture classification. A novel texture feature called "characteristic view", which is directly extracted from a kernel corresponding to each texture class, was developed. The K-views template method was used to classify the texture pixels based on these features. The characteristic view concept is based on the assumption that an image taken from the nature scenes, a specific texture class in this image will frequently reveal the repetitions of some certain patterns of features. Different "views" can be obtained as features from different spatial locations. The datagram concept is developed in this paper. Experimental results using datagrams are provided.

A simple model for preattentive texture perception/discrimination is suggested. According to this model and with the introduction of local vector mapping (LVM) for texture description, a new algorithm for texture edge detection is developed. The algorithm basically consists of two steps, namely, intensity slicing to decompose complex forms into simple features, called textons, and textural transform to transform textural differences into gray-tone differences. Experimental results show that the algorithm has a number of desirable features such as the simplicity of the structures it uses, the consistency of the approach with the human visual system, and the accuracy of the detected texture edges. >

The usefulness of multichannel filtering for image texture segmentation has been demonstrated in the literature. Based on psychophysical studies, it has been largely believed that Gabor filters are optimal for segmentation. In this paper, the application of maximally decimated, non-separable, perfect reconstruction filter banks (subband systems) are investigated for texture segmentation. The subband decomposition used in this paper forms a maximally decimated filter bank and therefore computation time and complexity is much less than the full rate case of the Gabor filters. The directional selectivity achieved by the non-separable, multirate filter bank results in performance similar to that achievable with the Gabor filters, but with a drastically reduced computational load.

A novel neural network architecture for image texture classification is introduced. The proposed kernel modifying neural network (KM Net), which incorporates a convolution filter kernel array and a classifier in one, enables an automated texture feature extraction in the multichannel texture classification through modification of the kernels and the connection weights by a backpropagation-based training rule. The first layer units working as the convolution kernels are constrained to be an array of Gabor filters, which achieves the most efficient texture feature localization. The following layers work as a classifier of the extracted texture feature vectors. The capability of the KM Net and its training rule is verified with basic problems of synthetic and fabric texture images, and also with a biological tissue classification problem in an ultrasonic echo image.

This work presents a texture based image retrieval method, called RAH, in the frequency domain. In RAH, texture features are represented as power spectral histograms of images. RAH mainly consists of two parts: the extraction of texture features based on the histogram of image power spectral in the frequency domain and the similarity measurement function based on the texture features. The experimental results indicate that RAH outperforms the traditional cooccurrence matrix texture representation.

In this paper we investigate the role of colour spaces on texture analysis. We extract a range of correlogram and colour moment features for the VisTex colour texture benchmark in different colour spaces and find the average probabilistic distance of separation across different objects for different features and suggest the colour spaces that are best suited for the classification process. We also show the results of k-nearest neighbour classification for different features and their combined set.

Homomorphic filtering using the complex cepstrum has been used in many applications of one-dimensional (1D) signal processing, notably speech signal analysis and seismic signal analysis. Increasing attention is now being directed towards extension of cepstral filtering to the 2D domain. An application that we are particularly interested in is the extraction of texture elements or textons from an image with quasi-periodic or ordered texture. We show in this paper that 1D cepstral filtering techniques may be applied to the projections of textured images at various angles, i.e., in the Radon domain. The filtered projections may then be used to reconstruct an image of a single texton. Preliminary results with a synthetic image are presented to illustrate the application of the method. The technique should be easier to apply than 2D cepstral filtering techniques, and find applications in analysis of textured images and images obtained via reconstruction from projections.

Texture is an important feature that aids in identifying objects of interest or region of interest irrespective of the source of the image. In this paper, a novel and simple isopattern-based texture feature is introduced. Spatial gray scale dependencies represented by bit plane is analyzed for specific patterns and are accumulated in bins. These are scaled by half-normal weighting function to provide isopattern texture feature. The ability of this texture feature in capturing textural variations of the images despite the presence of illumination, scale and rotation is demonstrated by conducting texture analysis on Brodatz, OuTex texture datasets and its classification accuracy on Kylberg dataset. The results of these two experimentation indicate that the proposed textural feature picks variation in texture significantly and has a better texture classification accuracy of 98.26% when compared with the state-of-the-art features like Gabor, GLCM and LBP.

Texture description is important for image retrieval from various large image databases. A method for representing texture information in images utilizing dual tree complex wavelet transform (DT-CWT) is introduced. Approximate shift invariance, good directional selectivity and computational efficiency properties of DT-CWT are effective in describing texture patterns. The perfect reconstruction property of DT-CWT enables us to store the image data using DT-CWT coefficients. Experiments are performed with Brodatz (1966) texture database and nature images that have significant amount of texture information to investigate the efficiency of our texture description algorithm.

Several frequency domain or joint spatial/frequency domain techniques for image texture classification havebeen published. We formulate these techniques within a common signal processing framework based on digitalfilter banks. The usefulness of computationally efficient hR filter banks as channel filters in texture classifiers isdemonstrated. Using estimates of local energy in the frequency channels we also propose a technique for selectingoptimum filter banks by maximizing a between class distance measure. This optimization is particularly simplewhen using the hR based filter banks. 1 Introduction The problem of reliable texture classification based on digital images has received considerable attention for more thantwo decades. The importance of this problem is highlighted by the numerous applications in such areas as medicaldiagnosis, remote sensing and industrial quality inspection. Several techniques emphasizing different approachesto image texture description have been presented earlier, —

Texture plays an important role in image analysis and understanding, with many applications in medical imaging and computer vision. However, analysis of texture by image processing is a rather difficult issue, with most techniques being oriented towards statistical analysis which may not have readily comprehensible perceptual correlates. We propose new methods for auditory display (AD) and sonification of (quasi-) periodic texture (where a basic texture element or `texton' is repeated over the image field) and random texture (which could be modeled as filtered or `spot' noise). Although the AD designed is not intended to be speech- like or musical, we draw analogies between the two types of texture mentioned above and voiced/unvoiced speech, and design a sonification algorithm which incorporates physical and perceptual concepts of texture and speech. More specifically, we present a method for AD of texture where the projections of the image at various angles (Radon transforms or integrals) are mapped to audible signals and played in sequence. In the case of random texture, the spectral envelopes of the projections are related to the filter spot characteristics, and convey the essential information for texture discrimination. In the case of periodic texture, the AD provides timber and pitch related to the texton and periodicity. In another procedure for sonification of periodic texture, we propose to first deconvolve the image using cepstral analysis to extract information about the texton and horizontal and vertical periodicities. The projections of individual textons at various angles are used to create a voiced-speech-like signal with each projection mapped to a basic wavelet, the horizontal period to pitch, and the vertical period to rhythm on a longer time scale. The sound pattern then consists of a serial, melody-like sonification of the patterns for each projection. We believe that our approaches provide the much-desired `natural' connection between the image data and the sounds generated. We have evaluated the sonification techniques with a number of synthetic textures. The sound patterns created have demonstrated the potential of the methods in distinguishing between different types of texture. We are investigating the application of these techniques to auditory analysis of texture in medical images such as magnetic resonance images.

Several approaches to multichannel filtering for texture classification and segmentation with Gabor filters havebeen proposed. The rationale presented for the use of the Gabor filters is their relation to models for the earlyvision of mammals as well as their joint optimum resolution in time and frequency. In this work we present acritical evaluation of the Gabor filters as opposed to filter banks used in image coding — in both full rate and critically sampled realizations. In the critically sampled case, tremendous computational savings can be realized.We further evaluate the commonly used octave band decomposition versus alternative decompositions.We conclude that for a texture segmentation task is it possible to use a wider range of filters than just theGabor class of filters, it is possible to use alternative decompositions and, most important, it is possible to usesubsampled filters. 1 Introduction In several recent papers the successful application of multichannel filtering for texture segmentation is reported [1,2, 3, 4, 5, 6, 7]. The channel filters utilized are Gabor filters [8] designed to respond to different spatial frequencies.The rationale presented for the use of the Gabor filters is their relation to models for the early vision of mammals aswell as their joint optimum resolution in time and frequency [8, 9] .