Search

Diffusion weighted imaging (DWI) is a non-invasive method for investigating the brain white matter structure. It can be used to evaluate fiber bundles in the brain. However, clinical acquisitions are often low resolution. This paper proposes a method for improving the resolution using sparse representation. In this method a non-diffusion weighted image (bO) is utilized to learn the patches and then diffusion weighted images are reconstructed based on the trained dictionary. Our method is compared with bilinear, nearest neighbor and bicubic interpolation methods. The proposed method shows improvement in terms of Peak Signal-to-Noise Ratio (PSNR) and Structural SIMilarity (SSIM).

Fuzzy systems that can automatically derive fuzzy if then rules and membership functions from numeric data have been developed previously. In this paper, we propose two new fuzzy learning methods for automatically deriving membership functions and fuzzy if-then rules from a set of given training examples. The proposed methods first select relevant attributes and build appropriate initial membership functions. They then simplify the intervals and the membership functions of each attribute before the decision table is formed. These attributes and membership functions are then used in a decision table to derive the final fuzzy if-then rules and membership functions. Experimental results on Iris data show that our methods can achieve a high accuracy. The proposed methods are thus useful in constructing membership functions and in managing uncertainty and vagueness. They can also reduce the time and effort needed to develop a fuzzy knowledge base.

The aim of this paper is to introduce a novel phase-based feature representation for robust speech recognition. This method consists of four main parts: autoregressive (AR) model extraction, group delay function (GDF) computation, compression, and scale information augmentation. Coupling GDF with an AR model results in a high-resolution estimate of the power spectrum with low frequency leakage. The compression step includes two stages similar to MFCC without taking a logarithm of the output energies. The fourth part augments the phase-based feature vector with scale information which is based on the Hilbert transform relations and complements the phase spectrum information. In the presence of additive and convolutional noises, the proposed method has led to 15% and 12% reductions in the averaged error rates, respectively (SNR ranging from 0 to 20 dB), compared to the standard MFCCs.

The 3D reconstruction of a face from a single frontal image is an ill-posed problem. This is further accentuated when the face image is captured under different poses and/or complex illumination conditions. In this paper, we aim to solve the shape recovery problem from a single facial image under these challenging conditions. The local image models for each patch of facial images and the local surface models for each patch of 3D shape are learned using a non-linear dimensionality reduction technique, and the correspondences between these local models are then learned by a manifold alignment method. By combining the local shapes, the global shape of a face can be reconstructed directly using a single least-square system of equations. We perform experiments on synthetic and real data, and validate the algorithm against the ground truth. Experimental results show that our method can yield accurate shape recovery from out-of-training samples with a variety of pose and illumination variations.

The segmentation of the lung tissues in chest Computed Tomography (CT) images is an important step for developing any Computer-Aided Diagnostic (CAD) system for lung cancer and other pulmonary diseases. In this paper, we introduce a new framework to generate 3D realistic synthetic phantoms to validate our developed Joint Markov-Gibbs based lung segmentation approach from CT data. Our framework is based on using a 3D generalized Gauss-Markov Random Field (GGMRF) model of voxel intensities with pairwise interaction to model the 3D appearance of the lung tissues. Then, the appearance of the generated 3D phantoms is simulated based on iterative minimization of an energy function that is based on using the learned 3D-GGMRF image model. These 3D realistic phantoms can be used to evaluate the performance of any lung segmentation approach. In this paper, we used the 3D realistic phantoms to evaluate the performance of our developed lung segmentation approach based on using the Dice Similarity Coefficient (DSC) metric and the Receiver Operating Characteristics (ROC). The DSC demonstrated that our approach achieves a mean DSC value of 0.994 ± 0.0034. Moreover, the ROC analysis for our method showed the best performance (area 0.99), while intensity showed the worst performance (area 0.92).

The appearances of the tracked object and its surrounding background usually change during tracking. As for tracking methods using subspace analysis, fixed subspace basis tends to cause tracking failure. In this paper, a novel tracking method is proposed by using incremental 2D-LDA learning and Bayes inference. Incremental 2D-LDA formulates object tracking as online classification between foreground and background. It updates the row- or/and column- projected matrix efficiently. Based on the current object location and the prior knowledge, the possible locations of the object (candidates) in the next frame are predicted using simple sampling method. Applying 2D-LDA projection matrix and Bayes inference, candidate that maximizes the posterior probability is selected as the target object. Moreover, informative background samples are selected to update the subspace basis. Experiments are performed on image sequences with the object's appearance variations due to pose, lighting, etc. We also make comparison to incremental 2D-PCA and incremental FDA. The experimental results demonstrate that the proposed method is efficient and outperforms both the compared methods.

Signature verification and Identification has great importance in authentication systems. Persian signatures are different from other signature types because people usually do not use text in it and they draw a shape as their signature. In this paper, a method for off-line Persian signature identification and verification is proposed that is based on DWT (Discrete Wavelet Transform) and Image Fusion. To extract features, at first DWT is used to access details of signature; then several signatures instances of each person are fused together to generate reference pattern of person's signature. In the classification phase, Euclidean distance between the test image and each pattern is used in each sub band. Experimental results confirmed the effectiveness of proposed method.

The network robustness is defined by how well its vertices are connected to each other to keep the network strong and sustainable. The change of network robustness may reveal events as well as periodic trend patterns that affect the interactions among vertices in the network. The evaluation of network robustness may be helpful to many applications, such as event detection, disease transmission, and network security, etc. There are many existing metrics to evaluate the robustness of networks, for example, node connectivity, edge connectivity, algebraic connectivity, graph expansion, R-energy, and so on. It is a natural and urgent problem how to choose a reasonable metric to effectively measure and evaluate the network robustness in the real applications. In this paper, based on some general principles, we design and implement a benchmark, namely BMNR, for the metrics of network robustness. The benchmark consists of graph generator, graph attack and robustness metric evaluation. We find that R-energy can evaluate both connected and disconnected graphs, and can be computed more efficiently.

In this paper, a Dynamic Stochastic Resonance (DSR)-based robust watermark extraction technique has been presented in combined DWT-SVD domain. The singular values of grayscale watermark image have been embedded in the singular values of the approximation sub-band of audio signal, obtained after DWT decomposition. DSR has been used in the extraction process to improve the authenticity of the extracted watermark by utilizing the noise or degradation introduced during different signal processing attacks. DSR is an iterative process that tunes the coefficients of a possibly attacked watermarked audio signal so that effect of noise is suppressed and hidden information is enhanced. An adaptive optimization procedure has been adopted for selection of bistable parameters to achieve maximum correlation coefficient between original and extracted watermark to ensure minimum computational complexity. Resilience of this technique has been tested in presence of various signal processing attacks like amplification, zero-cross, exchange, smooth, stat, add dynamic noise and cut-sample. Using the proposed technique, a robust extraction of watermark is achieved without trading-off the audibility of audio signal. Comparison with plain DWT-SVD-based technique suggests that the proposed DSR-based audio watermarking scheme gives remarkably better performance in terms of correlation between original and extracted watermarks.

In this paper, we propose the color fuzzy decision algorithm for face segmentation in a color image. Our algorithm can segment without the user's interaction by fuzzy decision marking. It removes small parts such as a noise using wavelet morphology in the image obtained by applying the fuzzy decision algorithm. Also, it merges and chooses the face region in each quantization image through rough sets. This video object division algorithm is shown to be superior to a conventional algorithm.

Diffraction tomography is a technique for imaging with acoustic fields in which parameter, such as reflective index, sound velocity, etc., can be mapped from scatter wave resulting from insonifying the object with a plane wave at a single temporal frequency. By solving the direct scattering problem, the scattered field can be presented in term of scattering parameters. Different inversion techniques can be applied to take advantage of the linearization process of the non-linear wave equation describing wave propagation in heterogeneous media under a limited class of scattering. Specifically, when the scattering effect is weak, one can invoke the Born or Rytov approximation and thus derive the generalized Fourier slice theorem to reconstruct the cross-section of the insonified object. Although diffraction tomography is a promising technology for medical application as it provides a quantitative ultrasonic image, its realization toward medical use is still far-to-go, and this may be due to the complexity of the hardware involved. In this research we investigate a potential use of diffraction tomography for medical application by using a delicate-designed ultrasonic computerized tomographic system. The contribution of the paper is that we have purposed to frequency synthesis method (FSM) that uses 3 frequency components of the Fourier transform of scattered filed, i.e. the center frequency and the -3 dB from center frequency of both side, to fill Fourier space for the generalized Fourier slice theory. The reconstructed ultrasonic image from such theory provides a very promising result.

In this paper, we present a new approach to detecting, locating and visualizing the local bifurcations originated from the changes of critical points in a 2D time-dependent vector field. Via precisely tracking the changes of all the critical points in a flow and pictorially displaying the evolution paths of these changes with a color model, our method can find and visualize all of this kind of bifurcations. This work extends Theisel's research on detecting a few local bifurcations based on the feature flow fields, and provides more complete information for probing into the structural instability in complex flow fields.

This paper presents several uses of the logical connective of implication to problems of interest in switching theory. The implications which hold among the prime implicants of a function are examined. A new set of necessary and sufficient conditions for determining essential prime implicants and a rapid approximate method for obtaining minimal sums are included.

We informally define a modular (cellular, iterative) computer to be a device consisting of a large (or, in theory, infinite) number of identical circuit modules connected together in some uniform manner, that is, in such a fashion that every module is connected into the device in the same manner as every other. In this paper we propose a mathematically precise definition of, "connected together in a uniform manner". In brief, we show that the underlying linear graph whose vertices correspond to the modules and whose edges correspond to the cables connecting the modules, is a group-graphs that is, the vertices correspond to the elements of a group G and there is given a finite subset G0 of G such that {g, g1} is an edge of the graph If, and only if, there exists g0 e G0 such that g1 = gg0. We further investigate the effects of restricting G to be an Abelian group and indicate why we feel such a restriction is unwarranted at least in developing the theory of modular computers.

Whether arrived at by chance or by design, any network of adaptive logic elements starts with a fixed structure. Attendant with a fixed structure are constraints on performance capabilities. Any task such a network is to learn will be performed in some input environment, and in most practical situations the environment itself will be constrained. These latter constraints imply the existence of certain relationships among the inputs. It is held that for the net to perform any given task well, the structure of the net must be matched to the relationships in the environment which are pertinent to the task. This paper discusses these ideas and presents an approach, which uses the theory of decomposition of switching functions, for matching the structure of a network with the constraints in the environment pertinent to the task.

Power load forecasting is one of the important parts in power system planning. This paper focused on the research of load forecasting and proposed an improved fuzzy AHP for mid-long term load forecasting. Firstly, Triangular fuzzy number is introduced to the model for indicating the judgments of experts, and then AHP is used for dealing with the judgments so as to obtain optimal weights of different forecasting schemes. The factors which can influence power load forecasting are taken into consideration in the proposed algorithm, meanwhile, the experts' experiences are combinated into the the processing of load forcasting to improve the accuracy. Finally, the experiment results shown that the presented algorithm can fit the medium and long term load forecasting better than other traditional approaches.

This paper explains our approach to the problem of pattern recognition by serial computer. The rudimentary theory of vision presented here lies within the framework of automata theory. Our goal is to classify the types of patterns that can be recognized by an automaton that scans a finite 2-dimensional tape. For example, we would like to know if an automaton can decide whether or not a given pattern on a tape forms a connected region. Although we have solved a number of problems, we have failed to solve this connectedness problem. This paper merely begins to describe the action of automata in higher dimensions. Our goal now is to generalize the theory presented here and make it applicable to a wide variety of pattern-recognizing machines.

Cellular logic arrays are beginning to take on an increasingly greater importance in digital technology, mainly because of their numerous advantages for the design, manufacture, and use in digital systems employing large-scale integrated semiconductor arrays. Particularly significant among these advantages is the feature of testability. One would naturally expect that the iterative structure and the short intercell connections of a cellular logic array would allow it to be tested from its edge terminals much more easily than a relatively disorganized interconnection of the same number of gates. In this paper we confirm this conjecture, and we describe procedures for deriving minimal (or near-minimal) schedules of test inputs, to be applied to a combinational cellular array in order to detect the presence of any single faulty cell. In addition, necessary and sufficient conditions are presented for some types of arrays of unilaterally connected, identical cells to be completely testable for single faults. For one-dimensional arrays, these conditions are based upon known results in sequential network theory. For two-dimensional arrays, a relationship to the "domino problem" (which is known to be insoluble) is described, but is shown to be largely avoidable and nonrestricting in the present case.

The paper proposes probabilistic default reasoning as a suitable approach to inheritance and recognition in uncertain and fuzzy object-oriented models. Firstly, we introduce an uncertain and fuzzy object-oriented model where a class property (i.e., an attribute or a method) can contain fuzzy sets interpreted as families of probability distributions, and uncertain class membership and property applicability are measured by lower and upper bounds on probability. Each uncertainty applicable property is interpreted as a default probabilistic logic rule, which is defeasible. In order to reduce the computational complexity of general probabilistic default reasoning, we propose to use R. Jeffrey's (1965) rule for a weaker notion of consistency and for local inference, then apply them to uncertain inheritance of properties. Using the same approach but with inverse Jeffrey's rule, uncertain recognition as probabilistic default reasoning is also presented. The approach is illustrated by an example in Fril++, the uncertain and fuzzy object-oriented logic programming language that we have been developing.

Rice is one of the food commodities that is most needed by the Indonesian people. Its condition requires farmers to maximize rice harvest as a rice-producing plant which one of them by providing fertilizer with the right dose. One of the methods used by rice farmers is to use a Leaf Color Chart to compare the color of rice leaves manually which might cause an error. Several research topics of classification based on plant image processing have been done to help the agriculture sector including rice. In this paper, the classification of rice leaves to determine the fertilizer dose by processing the rice leaf image using the HSV method is proposed. Results of rice leaf image processing are classified using fuzzy logic to calculate the right dose of fertilizer and developed as a mobile-based application. The proposed method achieved an accuracy value of 90% for the color of rice leaf and an accuracy value of 82.5% for the determination of fertilizer dose.

Operational requirements often make it necessary for several transmitters and receivers to operate in close proximity with small frequency separations. The important interference parameters that limit operation under these conditions are transmitter emission spectra, intermodulation, cross-modulation, and desensitization. This paper discusses the EMC problems produced by strong interfering signals in the neighborhood of the receiver tuned frequency. The region of concern, in a typical receiver, may consist of a considerable number of channels on both sides of the receiver-tuned frequency. This paper outlines methods that may be used to calculate the interference resulting from adjacent signals; presents mathematical models that describe the significant effects (such as intermodulation, cross-modulation, and desensitization); and describes methods for utilizing spectrum signature data for evaluating the models for a specific equipment. In addition, for cases where specific receiver measurements are not available, the paper presents general receiver models that have been derived from an analysis of spectrum signature data. The results of the work described in this paper will help to provide an understanding of the basic factors that contribute to the receiver susceptibility to interfering signals in adjacent channels. This information may be used by interference analysts in the calculation of receiver susceptibility and the generation of realistic EMC specifications, and by equipment designers to develop less susceptible equipments.

The amplitude of an emission spectrum which is generated by a noise modulated transmitter is properly calibrated in the power spectral density terms of power per bandwidth or watts/kHz. When a photograph is taken of the measured spectrum, it sometimes becomes difficult to determine the power spectral density at a particular point of the distributed spectrum. The relative scale is easily obtainable in dB but the absolute value is seldom if ever given correctly on measured data. The purpose of this paper is to develop a method for determining the largest absolute value of power spectral density for four most used transmission modes in communications; namely, single sideband, amplitude modulation, narrowband and wideband frequency modulation. A single equation is developed which expresses the maximum power spectral density as a function of average power, bandwidth and modulation factor. These three quantities are easily obtainable from manufacturers' information or measured data.

This paper surveys the developments in the field of pattern classification to describe the current state-of-the-art. The author divides the pattern classification problem into deterministic and statistical approaches although, in many instances, both converge to the same result. The approaches are further subdivided into: Adalines and Madalines, linear discriminant functions, mathematical programming, mode-seeking, nearest neighbor, Bayes and minimax, statistical criteria, and fuzzy sets subclasses. A major effort is devoted to show relationships among procedures. The conditions under which procedures are equivalent are discussed. Such relationships are summarized through a graph. The problems of supervision, adaptive property, sample size, and sequential analysis are discussed.

In this paper we propose a new approach to determine the length of root canal in dental X-ray image semi-automatically. The approach consists of a sequence of the following procedures: preprocessing procedure using contrast stretching method, segmentation procedure using active shape model to find the area of root canal and thinning procedure to find centerline of the root canal area. The length of the extracted centerline to be the length of root canal. Active shape model able to find root canal area in the image, and the length of root canal can be determined. Hence, semi-automatic determination of root canal length in dental X-ray image is feasible.

Image interpolation refers to constructing a high-resolution (HR) image from a low-resolution (LR) image. Traditionally, an HR image can be produced from an observed LR image via the polynomial-based interpolation (bi-linear or bi-cubic interpolations, involving a small number of neighbors around each interpolated position). The advanced interpolation makes use of the so-called “geometric similarity” to design a set of optimal interpolation weighting coefficients. However, better geometric similarities can perhaps be found from a non-local area within the LR source image or even from other but similar images (possibly with higher resolutions). Based on this fact, we propose in this paper a non-local geometric similarity based interpolation scheme to construct HR images. In our proposed method, optimal weighting coefficients are determined by solving a regularized least squares problem which is built upon a number of dual reference patches drawn from the observed LR image and regularized by the variation of directional gradients of the image patch. Experimental results demonstrate that our proposed method offers a remarkable quality improvement, both objectively and subjectively.

The need to locate several radio transmitters and receivers in close proximity due to space limitations and yet be able to operate them simultaneously introduces severe interference. The purpose of this paper is to discuss the feasibility of using a time-sharing approach to alleviate or eliminate some of these problems. The central idea to be investigated is the concept of a switching arrangement such that individual equipments "take turns" in performing their designated function. That is, at any instant of time, only one equipment (transmitter or receiver) in the complex is communicating. The switching must be done at a rate such that no loss of information occurs. This switching rate is specified by the sampling theorem, where the rate is a function of the highest frequency component in the message. Attention is focused on the method of switching or sampling, with particular emphasis on the shape of the sampling signal. The trade-offs in terms of spectrum width, signal level, and numbers of equipments to be handled are examined for several sampling pulse shapes.

In order to evaluate receiver susceptibility to interference, it is necessary to have some means of measuring and specifying receiver characteristics. One major source of receiver susceptibility information is the spectrum signature measurement program performed under MIL-STD-449C. Although many of the important receiver interference effects result from the same basic cause (i.e., undesired nonlinearities) and are influenced by the same basic factors such as selectivity and gain, under the current spectrum signature measurement program, each type of interference is measured separately as if they were completely unrelated. As a result, it is expensive and time consuming to obtain a complete "spectrum signature" for a receiver. Furthermore, receiver interference specifications have adopted the same general philosophy of treating each of the interference effects separately, and consequently, it is prohibitively expensive and difficult to establish any form of production line EMC quality control. What is desired is a receiver "figure of merit" that is simple to measure and describes the overall susceptibility of a receiver to interference. If this elusive but needed receiver "figure of merit" can be defined, it would provide a basis for receiver EMC specification, measurement, comparison, and quality control. This paper discusses a single receiver parameter that is easily measured, and may be used directly to evaluate receiver susceptibility to desensitization, cross modulation, and intermodulation. Furthermore, because this parameter provides an overall measure of the combined effects of RF selectivity, gain, and nonlinearity, it provides some information on receiver spurious response characteristics.

In structure from motion, if two metric structures are given, the unknown scale between them can be resolved by constraining the rigidity of the metric space. There exist two well-known approaches. The first one is the pose estimation which aims to find the pose of a camera for known 3D points. The second one is the scale estimation whose goal is to resolve the scale after estimating the motion between cameras. Recently, the former way is preferred because of the vulnerability of the scale estimation, the weakness to the image noise. In this paper, we thus propose a robust method which can overcome the weakness of the scale estimation by considering the uncertainty of reconstructed 3D points. Additionally, the rotation matrix is directly corrected under the structural consistency constraint. To illustrate the performance of the method, we demonstrate some examples of large-scale reconstructions and compare the results.

This paper is concerned with the use of nonlinear regression analysis as a means for achieving pictorial pattern recognition. The object to be identified is presumed to be represented by a set of image plane coordinates representing points associated with its boundary. Regression analysis techniques are then used to produce a sequence of computationally generated templates for various types of objects. The best fitting template type determines the classification of the object. Experimental results relating to a four class problem involving ellipses, rectangles, concave cresents, and convex cresents are presented.

This paper investigates the identification of a class of rapidly varying parameters, b"@k ("n x 1 vector), by means of sequential identification algorithms. These algorithms are heuristic extensions of relatively well-known, constant-parameter, identification algorithms to the time-varying parameter situation. It is assumed that b"@k = P(@k)@b"@k, where P(@k) is an n x n invertible information matrix whose elements are either measurable, or specified ahead of time, at t"@k, and @b"@k is a collection of unknown parameters which vary less rapidly than b"@k. Such a decomposition is often possible in aerospace applications. A priori and a posteriori identification algorithms are described and synthesized for both the perfect and noisy measurement situations. These algorithms are then applied to the identification of aerodynamic parameters of a high-performance, aerodynamically controlled aerospace vehicle. The feasibility of tracking these parameters by means of the a posteriori algorithms is demonstrated.

This paper discusses methods of estimating the probability of error for the Bayes' classifier which must be designed and tested with a finite number of classified samples. The expected difference between estimators is discussed. A simplified algorithm to compute Lachenbruch's method is proposed for multivariate normal distributions with unequal covariance matrices. Also, the variances of the likelihood ratios are given so as to compare them with the differences between the estimates. The discussion is extended to nonparametric classifiers by using the Parzen approximation for the density functions. Experimental results are shown for both parametric and nonparametric cases.

The use of a gradient search algorithm for the computation of the minimum of a function is well understood. In particular, in continuous time, the algorithm can be formulated as a differential equation whose equilibrium solution is a local minimum of the function. Khas'minskii has proposed using the continuous gradient algorithm with a white-noise driving term. He shows, using an argument on the convergence of the probability density function, that the equilibrium solution of this differential equation is the global minimum of the function. This paper reviews that result from the point of view of the theory of diffusion processes. It is shown that the conclusion of Khas'minskii is not correct, and that in fact the operation of the stochastic gradient search is the same as the operation of the noise-free gradient search. In fact, the search with noise converges with probability one to any of the local minima in the same way as the noise-free search converges to a local minimum. Several stochastic adaptive control systems use random search algorithms for their operation. One of these, due to Barron, is analyzed to show that it meets the conditions for convergence imposed by the results which have been derived here.

This paper presents some preliminary results on minimax feature extraction and selection using the Bhattacharyya coefficient as the selection criterion. It is assumed that the only knowledge of the two classes of pattern vectors is the mean vectors and the covariance matrices, and the minimax approach is attractive for this case because it is relatively insensitive to the variation of the true distributions. The maximization of the Bhattacharyya coefficient is discussed briefly.

Edge detection is often used procedure in digital image processing. For some practical applications is desirable to detect edges with sub-pixel accuracy. In this paper we present comparison of four methods for edge detection with sub-pixel accuracy in 1-D images: method based on approximation of real image function with erf function, moment-based edge operator, technique using spatial moments of the image function and the method based on wavelet transform of image. The paper presents results of simulations as well as of experiments to compare methods in terms of accuracy, the standard deviation of the edge localization error is chosen as precision criterion.

In this paper, we propose a new transductive learning framework for image retrieval, in which images are taken as vertices in a weighted hypergraph and the task of image search is formulated as the problem of hypergraph ranking. Based on the similarity matrix computed from various feature descriptors, we take each image as a ‘centroid’ vertex and form a hyperedge by a centroid and its k-nearest neighbors. To further exploit the correlation information among images, we propose a probabilistic hypergraph, which assigns each vertex v i to a hyperedge e j in a probabilistic way. In the incidence structure of a probabilistic hypergraph, we describe both the higher order grouping information and the affinity relationship between vertices within each hy-peredge. After feedback images are provided, our retrieval system ranks image labels by a transductive inference approach, which tends to assign the same label to vertices that share many incidental hyperedges, with the constraints that predicted labels of feedback images should be similar to their initial labels. We compare the proposed method to several other methods and its effectiveness is demonstrated by extensive experiments on Corel5K, the Scene dataset and Caltech 101.

This paper describes a new approach by which a blind person can visualize the approximate shape of the image by hearing sound. Nowadays a lot of image shape approximation and representation algorithms are available, but these algorithms represent image contour as a sequence of straight line segments. This work proposed a digital image shape approximation method where a shape boundary is divided & approximated to a few numbers of half-elliptical arc segments. These half elliptical arc segments are then represented by three-phoneme speech sound pattern so that on hearing this sound a blind person can get information about the shape of an object.

This paper provides a comparison of some of the steganalysis methods proposed in the literature, and using these comparison results, a global steganalysis methodology is proposed. The secret message detection capacities of these steganalysis methods are evaluated using stegoimages generated by typical data hiding algorithms. The evaluation of steganalysis methods is realized in terms of false negative and false positive error rates using 100 images. There isn’t any steganalysis that can detect presence of secret message in all type of stegoimages. Therefore, to realize a reliable analysis about a suspicious image, several steganalysis methods must be efficiently combined. In this paper, some considerations about steganalysis are provided using the results obtained of the comparison of steganalysis methods.

In the study of an object or a system, one must first of all describe it through interrelating the variables which are considered most relevant to the modeling of the system. However in many cases, the attributes of a system are ambiguous as in the case of handwritten characters; for there is no clear-cut definition of, say, handwritten "five". The characters "5", "5", "5", and "5" are somehow all recognized as "five". For machine recognition of this seemingly ill-defined object, it is necessary to be able to describe it analytically. One attempt is made in this paper to model the handwritten characters. The handwritten characters are directed graphs[1], the branch set of which are fuzzily classified to the branch types such as straight lines, circles and parts of a circle; the branch types in turn are considered as fuzzy variables[2]. A fuzzy function is then synthesized to model each handwritten character by relating fuzzy variables.

This In this paper, we describe the image processing techniques needed to produce sharp images using microwave technology. A scanning system produces a set of images representing the return of an object at several angles. By treating each point as a vector and calculating the autocorrelation function, we can produce relatively sharp images of the target. The technology can be used for detecting hidden targets.

One of the earliest-developed and still widely-used dc to square-wave inverters is the Royeror Uchrin-Taylor configuration shown in Fig. 1. Although the circuit is composed of very few components, its proper operation depends on complex interactions of the transformer as it switches between its unsaturated and saturated regions, and the tow transistors as they switch between cut-off and saturation. Numerous qualitative descriptions of the operation of this circuit have been presented; however, the complexities introduced by the multiple nonlinearities are such that few mathematically based analyses have been presented, and these have been rather limited in scope (1,2,3,4). The analysis presented in this paper provides insight not only into steady-state operation but also into the transient behavior of the circuit. It also makes evident the influence of certain small, often neglected, parasitic elements such as winding capacitance, transformer saturation inductance, and semiconductor junction capacitances.

The necessity of filtering noisy data generated by multidimensional processes arises in many diverse settings. The direct application of the Kalman-Bucy results is hindered by dimensionality difficulties inherent in multidimensional problems. This paper shows that for linear steady-state problems significant dimensionality reductions can be accomplished thus making routine the solution of many interesting problems.

Pattern classes composed of multiple gray level imagery can be recognized through processing of feature sets extracted from measures of the gray level or intensity structures of the patterns. The absence of geometric dependence in the feature extraction method leads to the formation of feature sets which are invariant under the pattern perturbations of translation, rotation, and dilation. This paper considers the application of the gray level processing method to a two-class recognition problem where the object gray level patterns are embedded in a gray level structured background of unknown character. In attacking this problem, a preprocessing algorithm is established which dissects the pattern image into a number of subsets of connected regions as a function of gray level. A "one at a time" sequential minimization procedure is then employed to determine an optimal gray level subset based on a minimum distance criterion.

This paper presents a new algorithm of image de-noising. This new algorithm introduces both a new threshold function and multilevel threshold, and has not only the de-noising ability of the energy-adaptation, but also the capability of protecting edge information of image signal during the de-noising process. Simulation results based on CL2 multi-wavelet transformation indicate that the de-noising approach with both the new threshold function and multilevel threshold outperforms the previous image de-noising methods in both visual effect and signal-to-noise ratio.

This paper presents a general theory of convergence for optimization algorithms in which approximations are used. The theory indicates how the accuracy of the approximations can be controlled adaptively in order to insure convergence. The application of the theory to a parameter optimization problem for a non-linear dynamical system is briefly discussed.

In this paper, a blind video watermarking algorithm based on 3D wavelet transform and Human Visual System (HVS) model is proposed. The proposed method is based on extracting temporal characteristics of video signal and using it to adjust spatial features of each frame. These frames are designed based on HVS model embed the message. Then the message will be extracted from video watermarked through blind way. Experimental results verifies the robustness of this method against frame swapping, frame dropping, frame averaging and MJPEG and MPEG2 compression.

A family of four static dc-to-square-wave transistor inverters has been analyzed employing state-plane techniques to describe the quasi -harmonic mode of oscillation commonly encountered in this family, as well as the less familiar relaxation oscillation and its limiting case, discontinuous oscillations. Design guidelines are provided through a study of the sensitivity of parameter changes which affect the oscillation characteristics of the inverter and its frequency stability. The analysis of one particular inverter of this family in which the avalanche breakdown of the transistor emitter-to-base junction occurs is presented. Finally the starting characteristics of this family of inverters are discussed with two illustrative examples. It is shown in this paper that close relationships exist between the family of LC tuned inverters and the previously analyzed saturable-core transistor inverters.

Wide band digital signal processing systems are becoming increasily important due to the advance of high speed digital Fourier analyzers. This paper presents a new threshold scheme that adaptively reduces noise effect in a data processing system that receives signals corrupted by non-stationary colored noise. The corrupted signal is decomposed into spectral segments by a discrete Fourier transform (DFT) unit and the resulting spectrum is processed by an adaptive threshold unit. The adaptive threshold network is tasked to estimate the noise level and to subtract this noise estimate from the composite signal plus noise to arrive at a high quality signal. For each set of DFT outputs, a finite order polynomial is used to approximate the average noise level by least squares method. The key feature in this polynomial approach is the reduction of the number of parameters characterizing the threshold. The results of this work show that second order can be sufficient, hence comparing to actual number of spectral values (generally greater than 512), the polynomial approach represents a large reduction in computation time and hardware circuit elements. The adaptive algorithm used in this work to estimate the -parameters sequentially is derived from adaptive estimation theory, which has been successfully implemented in a digital signal processing system.

Receivers operating on a space vehicle or an aircraft having many on-board transmitters are subject to intermodulation interference from mixing in the transmitting antenna systems, the external environment, or the receiver front-ends. This paper presents the techniques utilized in RFI Math Model computer programs that were developed to aid in the prevention of interference by predicting problem areas prior to occurrence. Frequencies and amplitudes of possible intermodulation products generated in the external environment are calculated and compared to receiver sensitivities. Intermodulation products generated in receivers are evaluated to determine the adequacy of preselector ejection.

In this paper we apply undersampling techniques to capture the temporal input-output relationship of RF power amplifiers. This approach avoids the distortion introduced by the upconverter and downconverter stages. We develop polynomial models with memory from the available data and evaluate its performance estimating device parameters like Adjacent power Ratio (ACPR) and AM-AM curves. The estimated parameters show good agreement with the empirical ones.

This paper considers the problem of estimation of classification error in pattern recognition. A theorem is presented to obtain the changes in the eigenvalues and eigenvectors of matrices of the form S 2 −1 S 1 , when there are changes of first order of smallness in the real symmetric matrices S i , i = 1, 2. Based on this theory a computational algorithm is developed for the estimation of classification error of Fisher classifier, using leaving groups out method.