Brodatz Album Research Articles

Texture classification based on EMD and FFT

Empirical mode decomposition (EMD) is an adaptive and approximately orthogonal filtering process that reflects human’s visual mechanism of differentiating textures. In this paper, we present a modified 2D EMD algorithm using the FastRBF and an appropriate number of iterations in the shifting process (SP), then apply it to texture classification. Rotation-invariant texture feature vectors are extracted using auto-registration and circular regions of magnitude spectra of 2D fast Fourier transform (FFT). In the experiments, we employ a Bayesion classifier to classify a set of 15 distinct natural textures selected from the Brodatz album. The experimental results, based on different testing datasets for images with different orientations, show the effectiveness of the proposed classification scheme.

Texture Classification and Retrieval Using the Random Neural Network Model

We extend the application of the random neural network to texture classification and retrieval. A neuron in the network corresponds to an image pixel and the neurons are connected according to neighboring relationship between pixels. A texture is represented with the weights of the network and the random neural network is used as an associative memory. In order to assess the performance of the method, texture mosaic images are generated from the popular Brodatz album. We also present a real life application in which a specified texture type is retrieved within a large remote sensing image. Results show that error is comparable to that of previous studies on texture classification which make use of other approaches.

Texture image retrieval and image segmentation using composite sub-band gradient vectors

A new texture descriptor, called CSG vector, is proposed for image retrieval and image segmentation in this paper. The descriptor can be generated by composing the gradient vectors obtained from the sub-images through a wavelet decomposition of a texture image. By exercising a database containing 2400 images which were cropped from a set of 150 types of textures selected from the Brodatz Album, we demonstrated that 93% efficacy can be achieved in image retrieval. Moreover, using CSG vectors as the texture descriptor for image segmentation can generate very successful results for both synthesized and natural scene images.

Texture Image Retrieval Using New Rotated Complex Wavelet Filters

A new set of two-dimensional (2-D) rotated complex wavelet filters (RCWFs) are designed with complex wavelet filter coefficients, which gives texture information strongly oriented in six different directions (45 degrees apart from complex wavelet transform). The 2-D RCWFs are nonseparable and oriented, which improves characterization of oriented textures. Most texture image retrieval systems are still incapable of providing retrieval result with high retrieval accuracy and less computational complexity. To address this problem, we propose a novel approach for texture image retrieval by using a set of dual-tree rotated complex wavelet filter (DT-RCWF) and dual-tree-complex wavelet transform (DT-CWT) jointly, which obtains texture features in 12 different directions. The information provided by DT-RCWF complements the information generated by DT-CWT. Features are obtained by computing the energy and standard deviation on each subband of the decomposed image. To check the retrieval performance, texture database D1 of 1856 textures from Brodatz album and database D2 of 640 texture images from VisTex image database is created. Experimental results indicates that the proposed method improves retrieval rate from 69.61% to 77.75% on database D1, and from 64.83% to 82.81% on database D2, in comparing with traditional discrete wavelet transform based approach. The proposed method also retains comparable levels of computational complexity.

Cosine-Modulated Wavelet Packet based Texture Features for Content based Image Retrieval

Content-based image retrieval (CBIR) is an active research area with application to digital libraries and multimedia databases. The focus of this paper is on the image processing aspects and in particular texture features for CBIR. We propose Cosine modulated wavelet packet based texture features for CBIR. The major advantages of Cosine modulated wavelet transform are less implementation complexity and good filter quality. Texture features are obtained by computing the energy, standard deviation and their combination on each subband of the decomposed image. To check the retrieval performance, texture database of 1856 textures is created from Brodatz album. Results are compared with standard Daubechies wavelet packet, Af-band wavelet packet and Gabor wavelet. Result indicates that proposed method is superior to other methods in terms of retrieval accuracy and retrieval time.

An efficient computation method for the texture browsing descriptor of MPEG-7

In this paper, an efficient computation method for computing the texture browsing descriptor of MPEG-7 is provided. Texture browsing descriptor is used to characterize a texture's regularity, directionality and coarseness. To compute the regularity of textures, Fourier transform is first performed. To get more discriminative features for regularity computation, the Fourier spectrum is treated as an image and the Fourier transform is performed again to produce an enhanced Fourier spectrum. A regularity measure based on the variance of the radial wedge distribution is then calculated to determine the regularity of textures. For regular textures, the texture primitives are assumed to be parallelograms, the two dominant directions are extracted by Hough transform. A scale computation method is then provided to determine the scales corresponding to the two dominant directions. In addition, principal component analysis is provided to detect textures with only one dominant direction. Experiments of texture browsing, coarse classification of textures and similarity-based image-to-image matching are performed on the texture images of Brodatz album and Corel Gallery image database to demonstrate the efficiency and effectiveness of the proposed method. The proposed method can be used in the applications of texture browsing and texture retrieval.

Cloud classification using the textural features of Meteosat images

The sum and difference histogram approach is applied to the assessment of the textural features of Meteosat images and the resulting textural parameters are used to classify the clouds appearing over North Africa. The images under consideration were taken by Meteosat in the visible and infrared bands during the month of December 1994. They cover North Africa, the Mediterranean Sea, Europe and the Atlantic Ocean. The visible images are first processed following four directions, namely 0°, 45°, 90° and 135°. For each direction, the textural parameters are calculated from the sum and difference histograms of grey levels. To take into account the effect of the texture anisotropy in the classification, the Karhunen-Loeve transformation (KLT) is then used. The minimum number of components representing these parameters is obtained with the slightest loss of information. Finally, after averaging the textural parameters over the four directions, each image of the database is divided into homogeneous classes by using the K-means algorithm. The approach thus described, tested on one of the images of the Brodatz album, shows that the classification ratio is greater than 96%. The segmentation of the Meteosat images is performed using both the textural parameters of visible images and the brightness of infrared images. It is then found that the different ground and cloud types are classified with proper accuracy. The implementation of this method to satellite images advantageously reduces the classification time, which is found to be three times smaller than that required by classical techniques of image processing.

Color and Texture Image Retrieval Using Chromaticity Histograms and Wavelet Frames

In this paper, we explore image retrieval mechanisms based on a combination of texture and color features. Texture features are extracted using Discrete Wavelet Frames (DWF) analysis, an over-complete decomposition in scale and orientation. Two-dimensional (2-D) or one-dimensional (1-D) histograms of the CIE Lab chromaticity coordinates are used as color features. The 1-D histograms of the a, b coordinates were modeled according to the generalized Gaussian distribution. The similarity measure defined on the feature distribution is based on the Bhattacharya distance. Retrieval benchmarking is performed over the Brodatz album and on images from natural scenes, obtained from the VisTex database of MIT Media Laboratory and from the Corel Photo Gallery. As a performance indicator recall (relative number of correct images retrieved) is measured on both texture and color separately and in combination. Experiments show this approach to be as effective as other methods while computationally more tractable.

Cosine-modulated wavelet based texture features for content-based image retrieval

Feature extraction is one of the most important tasks for efficient and accurate image retrieval purpose. In this paper we have presented a Cosine-modulated wavelet transform based technique for extraction of texture features. The major advantages of Cosine-modulated wavelet transform are less implementation complexity, good filter quality, and ease in imposing the regularity conditions. Texture features are obtained by computing the energy, standard deviation and their combination on each subband of the decomposed image. To check the retrieval performance, texture database of 1856 textures is created from Brodatz album. Retrieval efficiency and accuracy using Cosine-modulated wavelet based features is found to be superior to other existing methods.

Log-polar wavelet energy signatures for rotation and scale invariant texture classification

Classification of texture images is important in image analysis and classification. This paper proposes an effective scheme for rotation and scale invariant texture classification using log-polar wavelet signatures. The rotation and scale invariant feature extraction for a given image involves applying a log-polar transform to eliminate the rotation and scale effects, but at same time produce a row shifted log-polar image, which is then passed to an adaptive row shift invariant wavelet packet transform to eliminate the row shift effects. So, the output wavelet coefficients are rotation and scale invariant. The adaptive row shift invariant wavelet packet transform is quite efficient with only O(n /spl middot/ log n) complexity. A feature vector of the most dominant log-polar wavelet energy signatures extracted from each subband of wavelet coefficients is constructed for rotation and scale invariant texture classification. In the experiments, we employed a Mahalanobis classifier to classify a set of 25 distinct natural textures selected from the Brodatz album. The experimental results, based on different testing data sets for images with different orientations and scales, show that the proposed classification scheme using log-polar wavelet signatures outperforms two other texture classification methods, its overall accuracy rate for joint rotation and scale invariance being 90.8 percent, demonstrating that the extracted energy signatures are effective rotation and scale invariant features. Concerning its robustness to noise, the classification scheme also performs better than the other methods.

Design of a two-stage content-based image retrieval system using texture similarity

Efficacy and efficiency are two important issues in designing content-based image retrieval systems. In this paper, we present an efficient two-stage image retrieval system with high performance of efficacy based on two novel texture features, the composite sub-band gradient (CSG) vector and the energy distribution pattern (EDP)-string. Both features are generated from the sub-images of a wavelet decomposition of the original image. At the first stage, a fuzzy matching process based on EDP-strings is performed and serves as a signature filter to quickly remove a large number of non-promising database images from further consideration. At the second stage, the images passing through the filter will be compared with the query image based on their CSG vectors for detailed feature inspection. By exercising a database of 2400 images obtained from the Brodatz album, we demonstrated that both high efficacy and high efficiency can be achieved simultaneously by our proposed system.

A Generative Probabilistic Oriented Wavelet Model for Texture Segmentation

This Letter addresses image segmentation via a generative model approach. A Bayesian network (BNT) in the space of dyadic wavelet transform coefficients is introduced to model texture images. The model is similar to a Hidden Markov model (HMM), but with non-stationary transitive conditional probability distributions. It is composed of discrete hidden variables and observable Gaussian outputs for wavelet coefficients. In particular, the Gabor wavelet transform is considered. The introduced model is compared with the simplest joint Gaussian probabilistic model for Gabor wavelet coefficients for several textures from the Brodatz album [1]. The comparison is based on cross-validation and includes probabilistic model ensembles instead of single models. In addition, the robustness of the models to cope with additive Gaussian noise is investigated. We further study the feasibility of the introduced generative model for image segmentation in the novelty detection framework [2]. Two examples are considered: (i) sea surface pollution detection from intensity images and (ii) image segmentation of the still images with varying illumination across the scene.

TEXTURE ALGORITHMS: PERFORMANCE VARIABILITY ACROSS DATA SETS

Texture analysis plays a vital role in the area of image understanding research. One of the key areas of research is to compare how well these algorithms rank in their ability to differentiate between different textures. Traditionally, texture algorithms have been applied mostly to benchmark data such as the Brodatz album and the studies have found that certain algorithms are better suited for differentiating between certain types of textures. It would be presumptuous to believe that these chosen algorithms will perform well on other types of image data. We investigate the variability in performance across different benchmark data sets. Four benchmarks are chosen for analysis, including Brodatz album, MeasTex, VisTex, and Minerva. We take a total of seven texture feature extraction methods and compare the algorithms on the basis of their rankings on texture classification recognition rate and draw some important conclusions.

Integrating Cortex Transform and Brightness Based Features for Multi-texture classification.

This paper proposes a new scheme of fusing cortex transform and brightness based features obtained by local windowing operation. Energy features are obtained by applying popular cortex transform technique within a sliding window rather than the conventional way, while we define three features namely directional surface density (DSD), normalised sharpness index (NSI), and normalized frequency index (NFI) as mesures for pixel brightness variation. Fusion by simply vector tagging as well as by correlation is performed in the feature space and then classification is done using minimum distance classifier on the fused vectors. The interesting point is that brightness features, though inferior on some natural images, often produces smoother texture boundary in mosaic images, whereas energy features show the opposite behavior.This symmetrically inverse properties are combined through vector fusion for robust classification of muti-texture images obtained from Brodatz album and VisTex database. Results and comparison in terms of edge smoothness and confusion matrix based accuracy metrics show the robustness of the scheme.

On indexing the periodicity of image textures

Abstract Texture analysis is an important part in many computer vision tasks, and periodicity (regularity) is one of the most important texture features. Based on the structural view of texture, we describe a new way to index the periodicity of image textures in this paper. We suggest that a texture's periodicity can be further divided into two aspects — the regularity of the placement of its texels and the similarity among the texels. For a given texture image, we first find its gradient field. Then we derive the autocorrelation function of the gradient field. Two placement directions are determined from the autocorrelation function. For each direction, we extract a profile of correlation values. From each profile, we derive a placement regularity value, a texel similarity value, and a general periodicity value. Lastly, we combine the corresponding index values obtained from the two placement directions, using the total gradient magnitude along the two directions as weights, to obtain a final set of indices that describe the periodicity of the texture. We have applied the approach to the entire 112 textures in the Brodatz album. The results show that the suggested indices effectively reflect the periodicity of the test images.

Rotation-invariant texture classification using a two-stage wavelet packet feature approach

A novel two-stage wavelet packet feature approach for classification of rotated textured images is discussed. In the first stage, a set of sorted and dominant wavelet packet features is extracted from a texture image and a Mahalanobis distance classifier is employed to output N best classes. In the second stage, another set of wavelet packet features is extracted from the polarised form of the sample texture image and the most dominant wavelet packet features are selected and passed to the radial basis function (RBF) classifier with the N best classes to output the final matched class. Experimental results, based on a large sample data set of twenty distinct natural textures selected from the Brodatz album with different orientations, show that the proposed method outperforms the similar wavelet methods and the other rotation invariant texture classification schemes, and an overall accuracy rate of 91.4% was achieved.

Classification of noisy signals using fuzzy ARTMAP neural networks

This paper describes an approach to classification of noisy signals using a technique based on the fuzzy ARTMAP neural network (FAMNN). The proposed method is a modification of the testing phase of the fuzzy ARTMAP that exhibits superior generalization performance compared to the generalization performance of the standard fuzzy ARTMAP in the presence of noise. An application to textured gray-scale image segmentation is presented. The superiority of the proposed modification over the standard fuzzy ARTMAP is established by a number of experiments using various texture sets, feature vectors and noise types. The texture sets include various aerial photos and also samples obtained from the Brodatz album. Furthermore, the classification performance of the standard and the modified fuzzy ARTMAP is compared for different network sizes. Classification results that illustrate the performance of the modified algorithm and the FAMNN are presented.

Unsupervised texture segmentation via wavelet transform

In this paper, a mechanism for unsupervised texture segmentation is presented. The approach is based on the multiscale representation of the discrete (dyadic) wavelet transform which can be implemented by a fast iterative algorithm. For unsupervised segmentation it is generally difficult to determine the number of classes to be identified. The proposed approach offers an approach to circumvent this problem. Our method utilizes a set of high-frequency channel energies to characterize texture features, followed by a multi-thresholding technique for coarse segmentation. The coarsely segmented results at the same scale are incorporated by an intea-scale fusion procedure. A fine segmentation technique is then used to reclassify the ambiguously labeled pixels generated from the intea-scale fusion step. Finally, the number of texture classes is determined by an inter-scale fusion in which the segmentation results at multiple scales are integrated. The performance of this method is demonstrated by several experiments on synthetic images, natural textures from Brodatz's album and real-world textured images. Since the choice of wavelets is very extensive and open, we further explore various types of wavelets for texture segmentation. The time cost of the proposed method is also measured.

Towards a texture naming system: Identifying relevant dimensions of texture

Recently, researchers have started using texture for data visualization. The rationale behind this is to exploit the sensitivity of the human visual system to texture in order to overcome the limitations inherent in the display of multidimensional data. A fundamental issue that must be addressed is what textural features are important in texture perception, and how they are used. We designed an experiment to help identifiy the relevant higher order features of texture perceived by humans. We used twenty subjects, who were asked to rate 56 pictures from Brodatz's album on 12 nine-point Likert scales. Each subject was also asked to group these pictures into as many classes as desired. We applied the techniques of hierarchical cluster analysis and non-parametric multidimensional scaling (MDS) to the pooled similarity matrix generated from the subjects' groupings. We used Classification and Regression Tree Analysis (CART), discriminant analysis, and principal component analysis on the data from the scale ratings. The clusters generated from hierarchical cluster analysis remained intact in the MDS plots. We found that the MDS solutions fit the data well. The stress in the three-dimensional case is 0.12. The CART and discriminant analyses provided further justification for our interpretation. The three orthogonal dimensions we identified for texture are repetitive vs non-repetitive; high-contrast and non-directional vs low-contrast and directional; granular, coarse and low-complexity vs non-granular, fine and high-complexity.

Multifractal and generalized dimensions of gray-tone digital images

Fractal geometry has found widespread applications in image processing problems. One interesting fractal parameter is the fractal dimension that characterizes the roughness of the image. Higher-order fractal dimensions called multifractal dimensions are used to characterize the underlying inhomogeneity of texture in an image. This note deals with the methods of estimating fractal and multifractal dimensions of gray-tone digital images. Experimental results on texture images from the Brodatz album are presented and discussed.