The technique outlined in this paper extends the ability of current warping motion estimation schemes to allow occlusion and uncovering effects to be modelled. Current methods, [1], [2], [4], use a continuous rubber s...
详细信息
ISBN:
(纸本)0818679204
The technique outlined in this paper extends the ability of current warping motion estimation schemes to allow occlusion and uncovering effects to be modelled. Current methods, [1], [2], [4], use a continuous rubber sheet approach, consisting of non-overlapping polygons. The new technique estimates and compensates affine and then translational motion within the scene. The latter is achieved by allowing polygons to overlap through the introduction of rips into the sheet which are located in areas where occlusion or uncovering is thought to occur. Results show a reduction in the prediction error when compared to both traditional block-based methods and recently developed warping schemes (without rips).
We presents results obtained by different contrast enhancement methods applied to medical images. We take into account classical histogram specification, local and wavelet-based techniques and a novel approach for mul...
详细信息
ISBN:
(纸本)0818679204
We presents results obtained by different contrast enhancement methods applied to medical images. We take into account classical histogram specification, local and wavelet-based techniques and a novel approach for multiscale contrast enhancement. The latter, whose rationale grounds in theories of visual perception, exploits a local definition of the Fechner-Weber's contrast within the-context of a non-linear scale-space representation generated by anisotropic diffusion. Our experimental fields concerns a difficult kind of medical images, namely digital mammographic images.
A limitation of the existing maximum likelihood (ML) based methods for blur identification is that the estimate of blur is poor when the blurring is severe. In this paper, we propose an ML-based method for blur identi...
详细信息
ISBN:
(纸本)0818679204
A limitation of the existing maximum likelihood (ML) based methods for blur identification is that the estimate of blur is poor when the blurring is severe. In this paper, we propose an ML-based method for blur identification from multiple observations of a scene. When the relations among the blurring functions of these observations me known, we show that the estimate of blur obtained by using the proposed method is very good. The improvement is particularly significant under severe blurring conditions. With an increase in the number of images, direct computation of the likelihood function, however, becomes difficult as it involves calculating the determinant and the inverse of the cross-correlation matrix. To tackle this problem, we propose an algorithm that computes the likelihood function recursively as more observations are added.
The proceedings contain 29 papers. The topic discussed include: algorithm for classification of multispectral data and its implementation on a massively parallel computer;convergence measure and some parallel aspects ...
The proceedings contain 29 papers. The topic discussed include: algorithm for classification of multispectral data and its implementation on a massively parallel computer;convergence measure and some parallel aspects of Markov-chain Monte Carlo algorithms;evidential reasoning based on Dempster-Shafer theory and its application to medical image analysis;cluster approximations for statistical imageprocessing;image recovery and segmentation using competitive learning in a layered network;feature competition and domain of attraction in artificial-perceptron pattern recognizer;and probabilistic spectral feature extraction technique for neural networks.
Usually, we begin the analysis of an image by partitioning it into several homogeneous regions; this is the image segmentation. The homogeneity can be defined by many properties such as gray level intensity, color, te...
详细信息
Usually, we begin the analysis of an image by partitioning it into several homogeneous regions; this is the image segmentation. The homogeneity can be defined by many properties such as gray level intensity, color, texture, etc. In many cases, the texture is the only available information. Texture analysis, as a result, has received considerable attention for the last two decades. A large number of approaches for texture classification and segmentation have been suggested. Commonly, two types of approaches are distinguished, adapted respectively to macro- and microtextures, namely, the structural and statistical approaches. As far as the latter is concerned, we can cite probabilistic methods based on texture modelling, statistical methods which characterize an image in terms of numerical attributes or features and new tools like neural networks, wavelets, multiresolution and multiscale approaches, and fuzzy modelling. A few methods also come from signalprocessing and seem to be promising: bidimensional autoregressive modelling and, time-frequency and time-scale representations. We focus on stochastic approaches and, specifically, on texture modelling by bidimensional autoregressive models (2D-AR models). We describe the AR model; we propose our method for choosing an adapted neighbourhood and evaluation. Then, our segmentation algorithm is presented with the classification criterion and the contextual information. Finally, we present experimental results showing the influence of the context and demonstrating the improvements brought by the adapted models and the multiscale approach.
An engineering annealing method, called hardware annealing method, for optimal solutions on cellular neural networks is presented, Cellular neural networks have great potential in solving many important scientific pro...
详细信息
An engineering annealing method, called hardware annealing method, for optimal solutions on cellular neural networks is presented, Cellular neural networks have great potential in solving many important scientific problems in signalprocessing and optimization by use of pre-determined templates, Hardware annealing, which is a parallel version of effective mean-field annealing in analog networks, is a highly efficient method to find optimal solutions on cellular neural networks, It does not require any iterative stochastic procedure and henceforth can be very fast, The landscape of the network energy function is first adapted so that the whole annealing process does not get stuck at a local minimum at the beginning of searching for optimal solutions by adjusting each neuron to a low voltage gain, Then, the hardware annealing searches for the globally minimum energy state by continuously increasing the gain of neurons, The globally optimal equilibrium state is reached when each neuron is at its maximum voltage gain. The robustness of this proposed hardware annealing method for global optimization is assessed by analysis of the eigenvalues in a corresponding dynamical system model.
A unified Eigenfilter approach is proposed for determining the mean-square-optimal coefficients of the McClellan transformation. The approach applies to all filter shapes without the use of prior knowledge of the prop...
详细信息
A unified Eigenfilter approach is proposed for determining the mean-square-optimal coefficients of the McClellan transformation. The approach applies to all filter shapes without the use of prior knowledge of the properties of the coefficients. Several design examples for arbitrarily shaped and oriented 2-D fan, elliptical, and diamond filters are given to demonstrate the results achieved with this method. Comparisons with other recently published methods are made to demonstrate the advantages of our method.
Weighted Median (WM) filters have attracted a growing number of interest in the past few years. They inherent the robustness and edge preserving capability of the classical median filter and resemble linear FIR fitter...
详细信息
Weighted Median (WM) filters have attracted a growing number of interest in the past few years. They inherent the robustness and edge preserving capability of the classical median filter and resemble linear FIR fitters in certain properties. Furthermore, WM filters belong to the broad class of nonlinear filters called stack filters. This enables the use of the tools developed for the latter class in characterizing and analyzing the behavior and properties of WM filters, e.g. noise attenuation capability. The fact that WM filters are threshold functions allows the use of neural network training methods to obtain adaptive WM filters. In this tutorial paper we trace the development of the theory of WM filtering from its beginnings in the median filter to the recently developed theory of optimal weighted median filtering. The following one and multidimensional applications are presented in this paper: idempotent weighted median filters for speech processing, adaptive weighted median and optimal weighted median filters for image and image sequence restoration, weighted medians as robust predictors in DPCM coding and Quincunx coding, and weighted median filters in scan rate conversion in normal TV and HDTV systems.
This paper is concerned with the identification of corresponding points in curves related together, as time series, or in epipolar lines of stereo images. The properties of the corresponding points are defined by the ...
详细信息
ISBN:
(纸本)7505338900
This paper is concerned with the identification of corresponding points in curves related together, as time series, or in epipolar lines of stereo images. The properties of the corresponding points are defined by the near neighbourhood but also further points which are related to the corresponding point. Filtering with special non-linearities is applied to suppress non-relevant information and to find interrelations. For this the curves are represented by non-linear stochastic differential equations. The properties of the curves are obtained by the estimation of the expectation values of these stochastic equations and represented by different fuzzy measures. The corresponding points are determined by fusion of information obtained from different properties related to a special corresponding point with the help of fuzzy integrals. Using filtering and fuzzy integration of properties, feature and intensity based methods are combined.
Automatic sorting in the canning industry is searching for simple methods for classifying automatically the products. Reliable classifiers can use discriminant analysis but the actual technology available with neural ...
详细信息
ISBN:
(纸本)7505338900
Automatic sorting in the canning industry is searching for simple methods for classifying automatically the products. Reliable classifiers can use discriminant analysis but the actual technology available with neural networks is growing everyday due to its advantages of experience-based learning, generalization, graceful degradation and fault tolerance. In this paper we present a neural network classifier of jalapeno chile based on chile images obtained with a CCD camera. after being singulated. The image is processed and the noise outside from its perimeter eliminated before calculating features as area length and angle. The area, length under different angles are introduced as training data to a propagation algorithm which classifies chiles in three different categories.
暂无评论