Super-resolution target identification from remotely sensed imagesusing a Hopfield neural network

Fuzzy classification techniques have been developed recently to estimate the class composition of image pixels, but their output provides no indication of how these classes are distributed spatially within the instantaneous field of view represented by the pixel. As such, while the accuracy of land cover target identification has been improved using fuzzy classification, it remains for robust techniques that provide better spatial representation of land cover to be developed. Such techniques could provide more accurate land cover metrics for determining social or environmental policy, for example. The use of a Hopfield neural network to map the spatial distribution of classes more reliably using prior information of pixel composition determined from fuzzy classification was investigated. An approach was adopted that used the output from a fuzzy classification to constrain a Hopfield neural network formulated as an energy minimization tool. The network converges to a minimum of an energy function, defined as a goal and several constraints. Extracting the spatial distribution of target class components within each pixel was, therefore, formulated as a constraint satisfaction problem with an optimal solution determined by the minimum of the energy function. This energy minimum represents a “best guess” map of the spatial distribution of class components in each pixel. The technique was applied to both synthetic and simulated Landsat TM imagery, and the resultant maps provided an accurate and improved representation of the land covers studied, with root mean square errors (RMSEs) for Landsat imagery of the order of 0.09 pixels in the new fine resolution image recorded     

Chromosomal changes in primary myelodysplasias

The data from literature on chromosome abnormalities in various subgroups of the myelodysplastic syndrome (according to FAB-classification) are presented. Special attention is paid to information about changes in the number and structure of chromosomes during the disease and their prognostic significance. The relation of definite chromosome aberrations to the clinical course and prognosis of the disease, in particular to its transformation into acute leukosis is under discussion.     

Application of neural networks to AVHRR cloud segmentation

The application of neural networks to cloud screening of AVHRR data over the ocean is investigated. Two approaches are considered, interactive cloud screening and automated cloud screening. In interactive cloud screening a neural network is trained on a set of data points which are interactively selected from the image to be screened. Because the data variability is limited within a single image, a very simple neural network topology is sufficient to generate an effective cloud screen. Consequently, network training is very quick and only a few training samples are required. In automated cloud screening, where a general network is designed to handle all images, the data variability can be significant and the resulting neural network topology is more complex. The latitudinal, seasonal and spatial dependence of cloud screening large AVHRR data sets is studied using an extensive data set spanning 7 years. A neural network and associated feature set are designed to cloud screen this data set. The sensitivity of the thermal infrared bands to high atmospheric water vapor concentration was found to limit the accuracy of cloud screening methods which rely solely on data from these channels. These limitations are removed when the visible channel data is used in combination with the thermal infrared data. A post processing algorithm is developed to improve the cloud screening results of the network in the presence of high atmospheric water vapor concentration. Post processing also is effective in identifying pixels contaminated by subpixel clouds and/or amplifier hysteresis effects at cloud-ocean boundaries. The neural network, when combined with the post processing algorithm, produces accurate cloud screens for the large, regionally distributed AVHRR data set     

Baltic Sea ice SAR segmentation and classification using modified pulse-coupled neural networks

A method for segmentation and classification of Baltic Sea ice synthetic aperture radar (SAR) images, based on pulse-coupled neural networks (PCNNs), is presented. Also, automated training, which is based on decomposing the total pixel value distribution into a mixture of class distributions, is presented and discussed. The algorithms have been trained and tested using logarithmic scale Radarsat-1 ScanSAR Wide mode images over the Baltic Sea ice. Before the decomposition into mixture of class distributions, an incidence angle correction, specifically designed for these Baltic Sea ice SAR images, is applied. Because the data distributions in the uniform areas of these images are very close to Gaussian distributions, the data are decomposed into a mixture of Gaussian distributions, using the Expectation-Maximazation algorithm. Only uniform image areas are used in the decomposition phase. The mixture of distributions is compared to the distributions of the Baltic Sea ice classes, based on earlier scatterometer measurements and visual video interpretations of the sea ice classes. The parameter values for the PCNN segmentation are defined based on this mixture of distributions. The PCNN segmentation results are also compared to the operational sea ice information of digitized ice charts and to visual interpretation of the sea ice class.     

A Modified SI Epidemic Model for Combating Virus Spread in Wireless Sensor Networks

We study the dynamics of virus spread in wireless sensor networks (WSNs). We first analyze the susceptible-infective (SI) epidemic model for WSNs. In the SI model, once a sensor node is attacked by a virus, the infective node then, using normal communications, spreads the virus to its neighboring nodes, which further spread the virus to their neighbors, the process continues until the whole network fails. To combat this drawback, we propose a modified SI model by leveraging the sleep mode of WSNs to perform system maintenance. The modified SI model can improve the network anti-virus capability and flexibly adapt to different types of virus, without causing any additional hardware effort and signaling overhead. We derive the explicit analytical solutions for the modified SI model, which can capture both the spatial and temporal dynamics of the virus spread process. Extensive numerical results are presented to validate our analysis. The proposed model and analysis method are expected to be used for analysis and design of information (including virus) propagation mechanisms in distributed wireless or computer networks.     

Performance surfaces of minimum effort estimators and controllers

The performance surfaces of two classes of minimum effort adaptive filter are studied. Eigenvalue spreads of the cost functions of minimum weight vector norm and minimum output variance algorithms are examined for both estimation and feedforward control applications. Results support the widespread practical use of the “leaky LMS” algorithm     

应用神经网络的图像分类矢量量化编码

矢量量化作为一种有效的图像数据压缩技术，越来越受到人们的重视。设计矢量量化器的经典算法ＬＢＧ算法，由于运算复杂，从而限制了矢量量化的实用性。本文讨论了应用神经网络实现的基于边缘特征分类的矢量量化技术。它是根据人的视觉系统对图象的边缘的敏感性，应用模式识别技术，在对图像编码前，以边缘为特征对图像内容分类，然后再对每类进行矢量量化。除特征提取是采用离散余弦变换外，图像的分类和矢量量化都是由神经网络完成     

Morphology neural networks: An introduction with applications

The area of artificial neural networks has recently seen an explosion of theoretical and practical results. In this paper, we present an artificial neural network that is algebraically distinct from the classical artificial neural networks, and several applications which are different from the typical ones. In fact, this new class of networks, calledmorphology neural networks, is a special case of a general theory of artificial neural nets, which includes the classical neural nets. The main difference between a classical neural net and a morphology neural net lies in the way each node algebraically combines the numerical information. Each node in a classical neural net combines information by multiplying output values and corresponding weights and summing, while in a morphology neural net, the combining operation consists of adding values and corresponding weights, and taking the maximum value. We lay a theoretical foundation for morphology neural nets, describe their roots, and give several applications in image processing. In addition, theoretical results on the convergence issues for two networks are presented.This research was supported in part by National Science Foundation, Contract No. ECS-9010403.     

Hierarchical universal coding

In an earlier paper, we proved a strong version of the redundancy-capacity converse theorem of universal coding, stating that for “most” sources in a given class, the universal coding redundancy is essentially lower-bounded by the capacity of the channel induced by this class. Since this result holds for general classes of sources, it extends Rissanen's (1986) strong converse theorem for parametric families. While our earlier result has established strong optimality only for mixture codes weighted by the capacity-achieving prior, our first result herein extends this finding to a general prior. For some cases our technique also leads to a simplified proof of the above mentioned strong converse theorem. The major interest in this paper, however, is in extending the theory of universal coding to hierarchical structures of classes, where each class may have a different capacity. In this setting, one wishes to incur redundancy essentially as small as that corresponding to the active class, and not the union of classes. Our main result is that the redundancy of a code based on a two-stage mixture (first, within each class, and then over the classes), is no worse than that of any other code for “most” sources of “most” classes. If, in addition, the classes can be efficiently distinguished by a certain decision rule, then the best attainable redundancy is given explicitly by the capacity of the active class plus the normalized negative logarithm of the prior probability assigned to this class. These results suggest some interesting guidelines as for the choice of the prior. We also discuss some examples with a natural hierarchical partition into classes     

Double-weight signature pattern codes for multicore-fibercode-division multiple-access networks

To transmit digitized image pixels in optical code-division multiple-access (CDMA) networks with multicore fiber, classes of two-dimensional (2-D) patterns, so-called optical orthogonal signature pattern codes (OOSPCs), have previously been proposed. The new technology enables parallel transmission and simultaneous access of multiple 2-D images in multiple-access environments. In this letter, we construct two new families of “double-weight” OOSPCs without the assumption of identical weight for all signature patterns in a code set. Since the performance of a signature pattern varies with its weight, these new double-weight OOSPCs are especially useful for optical CDMA networks with multiple performance requirements     

Balanced probe extends high-frequency measurements

The balanced coaxial probe is a valuable high frequency measurement tool that avoids the limitations of conventional probes, active and passive. Its resistive input impedance and balanced design are the keys to making clean measurements in the 1 to 500 MHz range. In fact, this probe represents an extremely effective method to accurately measure a high frequency voltage between nodes a few inches apart, as the required lead lengths for such measurements cause severe overshoots for “normal” probes with capacitive input impedances     

Data mapping by probabilistic modular networks andinformation-theoretic criteria

The quantitative mapping of a database that represents a finite set of classified and/or unclassified data points may be decomposed into three distinctive learning tasks: (1) detection of the structure of each class model with locally mixture clusters; (2) estimation of the data distributions for each induced cluster inside each class; and (3) classification of the data into classes that realizes the data memberships. The mapping function accomplished by the probabilistic modular networks may then be constructed as the optimal estimator with respect to information theory, and each of the three tasks can be interpreted as an independent objective in real-world applications. We adapt a model fitting scheme that determines both the number and kernel of local clusters using information-theoretic criteria. The class distribution functions are then obtained by learning generalized Gaussian mixtures, where a soft classification of the data is performed by an efficient incremental algorithm. Further classification of the data is treated as a hard Bayesian detection problem, in particular, the decision boundaries between the classes are fine tuned by a reinforce or antireinforce supervised learning scheme. Examples of the application of this framework to medical image quantification, automated face recognition, and featured database analysis, are presented as well     

Classification of multisensor remote-sensing images by structuredneural networks

Proposes the application of structured neural networks to classification of multisensor remote-sensing images. The purpose of the approach is to allow the interpretation of the “network behavior”, as it can be utilized by photointerpreters for the validation of the neural classifier. In addition, this approach gives a criterion for defining the network architecture, so avoiding the classical trial-and-error process. First of all, the architecture of structured multilayer feedforward networks is tailored to a multisensor classification problem. Then, such networks are trained to solve the problem by the error backpropagation algorithm. Finally, they are transformed into equivalent networks to obtain a simplified representation. The resulting equivalent networks may be interpreted as a hierarchical arrangement of “committees” that accomplish the classification task by checking on a set of explicit constraints on input data. Experimental results on a multisensor (optical and SAR) data set are described in terms of both classification accuracy and network interpretation. Comparisons with fully connected neural networks and with the k-nearest neighbor classifier are also made     

基于遗传算法的BP神经网络在图像压缩中的应用

以神经网络和遗传算法为代表的进化算法都基于智能信息处理的理论,但是各自都存在一些缺陷.设计并实现了基于遗传算法的BP神经网络算法BP-GA,该算法将遗传算法和BP算法相结合,用基于实数编码的遗传算法优化神经网络的权值后,应用于图像压缩.实验证明,利用此混合神经网络进行图像压缩,压缩比高,图像恢复质量效果好.     

A multichannel watershed-based segmentation method for multispectral chromosome classification

Multiplex fluorescent in situ hybridization M-FISH is a recently developed chromosome imaging technique where each chromosome class appears to have a distinct color. This technique not only facilitates the detection of subtle chromosomal aberrations but also makes the analysis of chromosome images easier; both for human inspection and computerized analysis. In this paper, a novel method for segmentation and classification of M-FISH chromosome images is presented. The segmentation is based on the multichannel watershed transform in order to define regions of similar spatial and spectral characteristics. Then, a Bayes classifier, task-specific on region classification, is applied. Our method consists of four basic steps: 1 computation of the gradient magnitude of the image, 2 application of the watershed transform to decompose the image into a set of homogenous regions, 3 classification of each region, and 4 merging of similar adjacent regions. The method is evaluated using a publicly available chromosome image database and the obtained overall accuracy is 82.4%. By introducing the classification of each watershed region, the proposed method achieves substantially better results compared to other methods at a lower computational cost. The combination of the multichannel segmentation and the region-based classification is found to improve the overall classification accuracy compared to pixel-by-pixel approaches.     

Maximum-likelihood techniques for joint segmentation-classification of multispectral chromosome images

Traditional chromosome imaging has been limited to grayscale images, but recently a 5-fluorophore combinatorial labeling technique (M-FISH) was developed wherein each class of chromosomes binds with a different combination of fluorophores. This results in a multispectral image, where each class of chromosomes has distinct spectral components. In this paper, we develop new methods for automatic chromosome identification by exploiting the multispectral information in M-FISH chromosome images and by jointly performing chromosome segmentation and classification. We (1) develop a maximum-likelihood hypothesis test that uses multispectral information, together with conventional criteria, to select the best segmentation possibility; (2) use this likelihood function to combine chromosome segmentation and classification into a robust chromosome identification system; and (3) show that the proposed likelihood function can also be used as a reliable indicator of errors in segmentation, errors in classification, and chromosome anomalies, which can be indicators of radiation damage, cancer, and a wide variety of inherited diseases. We show that the proposed multispectral joint segmentation-classification method outperforms past grayscale segmentation methods when decomposing touching chromosomes. We also show that it outperforms past M-FISH classification techniques that do not use segmentation information.     

Generating ROC curves for artificial neural networks

Receiver operating characteristic (ROC) analysis is an established method of measuring diagnostic performance in medical imaging studies. Traditionally, artificial neural networks (ANN's) have been applied as a classifier to find one “best” detection rate. Recently researchers have begun to report ROC curve results for ANN classifiers. The current standard method of generating ROC curves for an ANN is to vary the output node threshold for classification. Here, the authors propose a different technique for generating ROC curves for a two class ANN classifier. They show that this new technique generates better ROC curves in the sense of having greater area under the ROC curve (AUC), and in the sense of being composed of a better distribution of operating points     

Fault Diagnosis of Nonlinear Analog Circuits Using Neural Networks with Wavelet and Fourier Transforms as Preprocessors

A neural-network based analog fault diagnostic system is developed for nonlinear circuits. This system uses wavelet and Fourier transforms, normalization and principal component analysis as preprocessors to extract an optimal number of features from the circuit node voltages. These features are then used to train a neural network to diagnose soft and hard faulty components in nonlinear circuits. Our neural network architecture has as many outputs as there are fault classes where these outputs estimate the probabilities that input features belong to different fault classes. Application of this system to two sample circuits using SPICE simulations shows its capability to correctly classify soft and hard faulty components in 95% of the test data. The accuracy of our proposed system on test data to diagnose a circuit as faulty or fault-free, without identifying the fault classes, is 99%. Because of poor diagnostic accuracy of backpropagation neural networks reported in the literature (Yu et al., Electron. Lett., Vol. 30, 1994), it has been suggested that such an architecture is not suitable for analog fault diagnosis (Yang et al., IEEE Trans. on CAD, Vol. 19, 2000). The results of the work presented here clearly do not support this claim and indicate this architecture can provide a robust fault diagnostic system.     

Feature normalization via expectation maximization and unsupervised nonparametric classification for M-FISH chromosome images

Multicolor fluorescence in situ hybridization (M-FISH) techniques provide color karyotyping that allows simultaneous analysis of numerical and structural abnormalities of whole human chromosomes. Chromosomes are stained combinatorially in M-FISH. By analyzing the intensity combinations of each pixel, all chromosome pixels in an image are classified. Often, the intensity distributions between different images are found to be considerably different and the difference becomes the source of misclassifications of the pixels. Improved pixel classification accuracy is the most important task to ensure the success of the M-FISH technique. In this paper, we introduce a new feature normalization method for M-FISH images that reduces the difference in the feature distributions among different images using the expectation maximization (EM) algorithm. We also introduce a new unsupervised, nonparametric classification method for M-FISH images. The performance of the classifier is as accurate as the maximum-likelihood classifier, whose accuracy also significantly improved after the EM normalization. We would expect that any classifier will likely produce an improved classification accuracy following the EM normalization. Since the developed classification method does not require training data, it is highly convenient when ground truth does not exist. A significant improvement was achieved on the pixel classification accuracy after the new feature normalization. Indeed, the overall pixel classification accuracy improved by 20% after EM normalization.     

基于组合差异图和卷积小波神经网络的SAR图像变化检测

合成孔径雷达(Synthetic aperture radar,SAR)图像变化检测可以看作是一个分类过程,像素被分类为变化类和不变类。但是,差异图的质量影响了现有方法的检测精度。为了提供高质量的差异图,提出一种基于组合差异图和卷积小波神经网络(Convolutional-Wavelet Neural Network,CWNN)结合的SAR图像变化检测方法。首先,使用对数比算子、均值比算子和差分算子产生差异图,并使用简单的组合方法得到最终的差异图。然后,在差异图上使用分层模糊C均值获得变化类和不变化类的训练样本。最后,使用训练好的CWNN对所有像素进行分类。该方法提高了差异图的质量,为网络提供高质量的训练样本。真实SAR图像数据集上的实验结果表明,该方法提高了变化检测精度。