Chair: Edward J. Delp, Purdue University (USA)
Yue Wu, Kyoto Institute of Technology (JAPAN)
Yasuo Yoshida, Kyoto Institute of Technology (JAPAN)
This paper presents a new approach for texture classification using rotation and scaling invariant parameters. A test textured image can be correctly classified even if it is rotated and scaled. Based on a 2-D Wold-like decomposition of homogenenous random fields, the texture field can be decomposed into a deterministic component and an indeterministic component. The spectral density function(SDF) of the former is a sum of 1-D or 2-D delta functions. The 2-D autocorrelation function(ACF) of the latter is fitted to the assumed anisotropic ACF that has an elliptical contour. Invariant parameters applicable to the classification of rotated and scaled textured images can be estimated by combining the parameters representing the ellipse and those representing the delta functions. The effectiveness of this method is illustrated through experimemtal results on natural textures.
Sergio Barbarossa, University of Rome (ITALY)
Laura Parodi, University of Rome (ITALY)
The aim of this work is to propose a method for optimizing the decomposition law of a tree-structured wavelet transform in order to maximize the capability of discriminating different textures. The optimization criterion is the maximization of the Fisher's distance. The analysis is carried out theoretically and by simulation on gaussian Markov random fields and is then applied to the classification of real Synthetic Aperture Radar images.
Ashit Talukder, Carnegie Mellon University
Jennifer Davidson, Iowa State University (USA)
Texture is a phenomenon in image data that continues to receive attention due to its wide-spread applications, ranging from remotely sensed data, to medical imaging, to military applications. In this paper we use a new class of spatial stochastic models called partially ordered Markov models (POMMs) for texture analysis and model selection. POMMs are a generalization of Markov mesh models that have the property that their joint probability density function is an exact, closed form expression in terms of conditional probabilities. Markov random fields (MRFs) do not, in general, have this property. This property of the POMMs has lead to exact and fast computations involving the joint probabilities. We show how these fast algorithms allow POMMs to be used for fitting models to textures, and for supervised texture segmentation. Applications to real data show that the model selection technique gives very good results. POMMs are a broad and general class of models, and have the potential to be applied to diverse areas beyond imaging, such as probabilistic expert systems and artificial intelligence.
Martin Pfeiffer, University of Kaiserslautern (GERMANY)
Madhukar Pandit, University of Kaiserslautern (GERMANY)
In this paper we propose and illustrate the application of Wavelet descriptors for the quantitative description of shapes. Fourier descriptors are known to be useful for describing shapes based on their boundaries. As these basically operate at one given scale or resolution, their application leads to loss of information of salient features of a shape contained at other scales. Due to their inherent multiscale properties Wavelet descriptors are potentially suitable for shape discrimination in such situations. The paper gives the basic techniques of applying Wavelet descriptors in practical applications.
Anamitra Makur, Indian Institute of Science (INDIA)
M. R. Raghuveer, Rochester Institute of Technology (USA)
Approaches for digital halftoning of images using dithering threshold the input image with additive dithering noise. The paper presents a technique which thresholds the noise directly. The threshold is modified at each step such that the expected value of the output is equal to the input pixel's gray value. Further, error feedback is used to correct the threshold. Tests on images show the method's ability to retain features in the high frequency regions such as edges as well as low frequency features such as slow variations in intensity.
O. Alata, ENSERB (FRANCE)
P. Baylou, ENSERB (FRANCE)
M. Najim, ENSERB (FRANCE)
In the framework of model based image processing, we propose a new parametric approach for classifying textured images. The image, considered as a two-dimensional stochastic process, is characterized by a set of reflection coefficients computing using a two-dimensional adaptive lattice filter based on RLS criterion. The corresponding algorithm is named Two-Dimensional Fast Lattice Recursive Least Squares algorithm (TDFLRLS). In order to evaluate our method, classification rates are calculated on a set of 8 different textures from the Brodatz Album. We carry out performance comparisons with methods of characterization based on two-dimensional AR coefficients computed with two-dimensional AR filters or statistical features calculated from co-occurrence matrices and neighbouring matrices.
Srinivas R. Kadaba, Purdue University (USA)
Saul B. Gelfand, Purdue University (USA)
R.L. Kashyap, Purdue University (USA)
We present real-time algorithms for the segmentation of binary images modeled by Markov Mesh Random Fields (MMRF's) and corrupted by independent noise. The goal is to find a recursive algorithm to compute the MAP estimate of each pixel of the scene using a fixed lookahead of D rows and D columns of the observations. The optimal algorithm for this is computationally expensive. Using both hard and soft (conditional) decision feedback, the complexity is reduced in a principled manner to allow a performance/complexity tradeoff. Simulation results demonstrate the viability of the algorithm and its subjective relevance to the image segmentation problem.
Kamran Etemad, University of Maryland (USA)
Rama Chellappa, University of Maryland (USA)
An algorithm for classification task dependent multi- scale feature extraction is suggested. The algorithm focuses on dimensionality reduction of the feature space subject to maximum preservation of classification information. It has been shown that, for classification tasks, class separability based features are appropriate alternatives to features selected based on energy and entropy criteria. Application of this idea to feature extraction from multi-scale wavelet packets is presented. At each level of decomposition an optimal linear transform that preserves class separabilities and results in a reduced dimensional feature space is obtained. Classification and feature extraction is performed at each scale and resulting soft decisions are integrated across scales. The suggested scheme can also be applied to other orthogonal or non-orthogonal multiscale transforms e.g. local cosine transform or Gabor transform. The suggested algorithm has been tested on classification and segmentation of some radar target signatures as well as textured and document images.
Mahmoud R. El-Sakka, University of Waterloo (CANADA)
Mohamed S. Kamel, University of Waterloo (CANADA)
This paper is concerned with segmenting light intensity images for the sake of compressing them using lossy compression techniques. Among the most commonly used techniques for image segmentation is Quad-tree partitioning. In this technique, block variance based criteria are usually used to measure the smoothness of the segmented blocks and to consequently classify them. Block variance, however, does not consider the pixel value distribution within the block. Instead of using the block variance as a segmentation and classification measure, we propose using the mean squared deviation from the neighboring pixels mean. The proposed measure is capable of differentiating between blocks not only according to block pixel values but also according to their distribution within the block. This leads to a much better image segmentation and consequently to higher image compression ratios with lower image degradation. The results show the superiority of the proposed measure over the block variance measure.
D. Hernandez-Sosa, Universidad de Las Palmas de Gran Canaria (SPAIN)
J. Cabrera-Gamez, Universidad de Las Palmas de Gran Canaria (SPAIN)
A. Falcon- Martel, Universidad de Las Palmas de Gran Canaria (SPAIN)
M. Hernandez-Tejera, Universidad de Las Palmas de Gran Canaria (SPAIN)
SVEX is a multilevel knowledge-based tool for developing applications in image segmentation. Both numerical and symbolic computations take place at each level, being the transitions between these two domains defined by the computational structure itself. SVEX incorporates evidence combination and uncertainty control mechanisms. SVEX is programmed by means of a specific purpose declarative language based on a reduced set of objects. All the knowledge involved in the solution of a given segmentation problem is made explicit due to the declarative nature of the programming language. The results obtained by the application of SVEX in the segmentation of a set of outdoor images are also shown in this article.