We present ViPRA-Haplo, a de novo strain-specific assembly workflow for reconstructing viral haplotypes in a viral population from paired-end next generation sequencing (NGS) data. The proposed Viral Path Reconstruction Algorithm (ViPRA) generates a subset of paths from a De Bruijn graph of reads using the pairing information of reads. The paths generated by ViPRA are an over-estimation of the true contigs. We propose two refinement methods to obtain an optimal set of contigs representing viral haplotypes. The first method clusters paths reconstructed by ViPRA using VSEARCH Deorowicz et al. 2015 based on sequence similarity, while the second method, MLEHaplo, generates a maximum likelihood estimate of viral populations. We evaluated our pipeline on both simulated and real viral quasispecies data from HIV (and real data from SARS-COV-2). Experimental results show that ViPRA-Haplo, although still an overestimation in the number of true contigs, outperforms the existing tool, PEHaplo, providing up to 9% better genome coverage on HIV real data. In addition, ViPRA-Haplo also retains higher diversity of the viral population as demonstrated by the presence of a higher percentage of contigs less than 1000 base pairs (bps), which also contain k-mers with counts less than 100 (representing rarer sequences), which are absent in PEHaplo. For SARS-CoV-2 sequencing data, ViPRA-Haplo reconstructs contigs that cover more than 90% of the reference genome and were able to validate known SARS-CoV-2 strains in the sequencing data.

Human Endogenous Retrovirus type K (HERV-K) is the only HERV known to be insertionally polymorphic; not all individuals have a retrovirus at a specific genomic location. It is possible that HERV-Ks contribute to human disease because people differ in both number and genomic location of these retroviruses. Indeed viral transcripts, proteins, and antibody against HERV-K are detected in cancers, auto-immune, and neurodegenerative diseases. However, attempts to link a polymorphic HERV-K with any disease have been frustrated in part because population prevalence of HERV-K provirus at each polymorphic site is lacking and it is challenging to identify closely related elements such as HERV-K from short read sequence data. We present an integrated and computationally robust approach that uses whole genome short read data to determine the occupation status at all sites reported to contain a HERV-K provirus. Our method estimates the proportion of fixed length genomic sequence (k-mers) from whole genome sequence data matching a reference set of k-mers unique to each HERV-K locus and applies mixture model-based clustering of these values to account for low depth sequence data. Our analysis of 1000 Genomes Project Data (KGP) reveals numerous differences among the five KGP super-populations in the prevalence of individual and co-occurring HERV-K proviruses; we provide a visualization tool to easily depict the proportion of the KGP populations with any combination of polymorphic HERV-K provirus. Further, because HERV-K is insertionally polymorphic, the genome burden of known polymorphic HERV-K is variable in humans; this burden is lowest in East Asian (EAS) individuals. Our study identifies population-specific sequence variation for HERV-K proviruses at several loci. We expect these resources will advance research on HERV-K contributions to human diseases.

Metagenomics involves the analysis of genomes of microorganisms sampled directly from their environment. Next Generation Sequencing allows a high-throughput sampling of small segments from genomes in the metagenome to generate reads. To study the properties and relationships of the microorganisms present, clustering can be performed based on the inherent composition of the sampled reads for unknown species. We propose a two-dimensional lattice based probabilistic model for clustering metagenomic datasets. The occurrence of a species in the metagenome is estimated using a lattice of probabilistic distributions over small sized genomic sequences. The two dimensions denote distributions for different sizes and groups of words respectively. The lattice structure allows for additional support for a node from its neighbors when the probabilistic support for the species using the parameters of the current node is deemed insufficient. We also show convergence for our algorithm. We test our algorithm on simulated metagenomic data containing bacterial species and observe more than 85% precision. We also evaluate our algorithm on an in vitro-simulated bacterial metagenome and on human patient data, and show a better clustering than other algorithms even for short reads and varied abundance. The software and datasets can be downloaded from https:// github.com/lattclus/lattice-metage.

The immune system is built to defend an organism against both known and new attacks, and functions as an adaptive distributed defense system. Artificial Immune Systems abstract the structure of immune systems to incorporate memory, fault detection and adaptive learning. We propose an immune system based real time intrusion detection system using unsupervised clustering. The model consists of two layers: a probabilistic model based T-cell algorithm which identifies possible attacks, and a decision tree based B-cell model which uses the output from T-cells together with feature information to confirm true attacks. The algorithm is tested on the KDD 99 data, where it achieves a low false alarm rate while maintaining a high detection rate. This is true even in case of novel attacks,which is a significant improvement over other algorithms.

Literature about cloud computing often makes the assumption that the resource demands of computing clouds (and the virtual machines that constitute them) are unpredictable in the short term. There are, however, specific use cases where resource demands can be anticipated. This paper discusses dissertation work-in-progress which shows that, in certain predictable environments, preemptive virtual machine migration can improve both computational resource utilization and the overall user experience. A novel algorithm which reacts to anticipated future resource demands based on past behavior of virtual machines is presented. Simulations are used to quantify performance improvements.

Health care is a highly regulated industry in which much value is placed upon privacy and confidentiality. The business of health care, particularly in certain academic environments, requires access to data of varying sensitivities, including information from the public Internet. This paper proposes a VLAN-based architecture for segregating data of varying sensitivities, a list of components that facilitate access to and distillation of data, and a method for one-way promotion of individual nodes from areas of lower security to areas of higher security. The proposed solution is an implementable and pragmatic approach to reducing the risk of data leakage. Quality of experience (QoE) measures of two methods for access (node promotion and porthole-based access) are compared. The node promotion method improves the user-perceived responsiveness of applications over the porthole-based method while reducing flexibility.

Development and application of pattern recognition techniques in the field of bioinformatics is of utmost importance for gaining new insights about phenomena in life sciences through the analysis of biological data. In this special section, three research manuscripts in their significantly extended form were selected from the papers presented at the Eighth IAPR International Conference on Pattern Recognition in Bioinformatics (PRIB 2013), which was held in Nice, France.

This paper describes doctoral research in progress that is focused on the optimization of private computing clouds for use in health care. Three areas of research will be examined: appropriate levels of hardware provisioning, algorithms for virtual machine to physical machine mapping and appropriate levels of redundancy.

There has been a growing interest in designing mobile systems consisting of special relay nodes whose mobility can be controlled by the underlying network. In this paper, we consider the design of a heterogeneous mobile ad hoc network (MANET) consisting of two kinds of mobile nodes-traditional nodes with limited energy and a few controllable mobile relay nodes with relatively abundant energy resources. We propose a novel relay deployment framework that utilizes mobility prediction and works in tandem with the underlying MANET routing protocol to optimally define the movement of the relay nodes. We present two instances of the relay deployment problem, together with the solutions, to achieve different goals. Instance 1, termed Min-Total, aims to minimize the total energy consumed across all the traditional nodes during data transmission, while instance 2, termed Min-Max, aims to minimize the maximum energy consumed by a traditional node during data transmission. Our solutions also enable the prioritization of individual nodes in the network based on residual energy profiles and contextual significance. We perform an extensive simulation study to understand the trade-offs involved in deploying an increasing fraction of such relay nodes in the network. We also investigate the performance of the proposed framework under different mobility prediction schemes. Results indicate that even when the relay nodes constitute a small fraction of the total nodes in the network, the proposed framework results in significant energy savings. Further, we observed that while both the schemes have their potential advantages, the differences between the two optimization schemes are clearly highlighted in a sparse network.

The study and inference of biological pathways and gene regulation mechanisms has become a vital component of modern medicine and drug discovery. Gene expression studies make it possible to understand these mechanisms by simultaneously measuring the expression level of thousands of genes. These data though rich in information are also prone to many quality control issues that ultimately result in noisy data. A new method to smooth the data and measure expression dissimilarity between genes is proposed in this paper. A new dissimilarity measure is defined as an approximation of the Kullback-Leibler divergence between mixture models. Further, a noise reduction method is also proposed for use with data from time-course experiments. Results from real data and simulated data demonstrate that the method is well suited for clustering gene expression profiles.

In recent years, there has been a growing interest in designing mobile systems consisting of special relay nodes whose mobility can be controlled by the underlying network. In this paper, we consider the design of a heterogeneous mobile ad hoc network (MANET) consisting of two kinds of mobile nodes - the traditional nodes with limited energy and a few mobility controllable relay nodes with relatively abundant energy resources. We propose the relay deployment problem that aims to optimally position these relay nodes in the network so as to minimize the overall power consumption for data transmission at the traditional nodes. We present a mobility prediction based framework to solve to the relay deployment problem in a truly mobile network. We investigate the performance of the proposed framework through extensive simulation study using three different mobility prediction schemes. We also perform experiments to understand the tradeoffs involved in deploying an increasing fraction of such relay nodes in the network. Results indicate that even when the relay nodes constitute a small percentage of the total nodes in the network, the proposed framework results in significant energy savings.

Quality-of-service (QoS) routing between domains is an essential component for providing service differentiation in the Internet. State aggregation is a technique that makes QoS routing scalable to large internetworks, by presenting a concise and accurate representation of the domain to the routing process. The concept of using a domain's routing capacity as a bandwidth aggregate has been in existence for some time. However, no methodology has been suggested in the literature for its estimation and usage. Also, the impact of a domain's routing capacity on the routing performance has not been studied before. This paper aims to fill these voids by presenting a framework based on "network flows" for estimating a domain's routing capacity and evaluating its efficacy on the routing performance. The routing capacity is used in conjunction with the conventional widest path bandwidth as the domain aggregate. Analytical and experimental results show that appropriate use of routing capacity along with the widest path bandwidth reduces the tendency of the advertised aggregate to overestimate bandwidth availability, and makes the routing process more robust to the frequency of domain state updates. During periods of congestion, the use of routing capacity can improve the bandwidth admitted into the network by as much as 20%

In this correspondence, we examine the relationship between Dempster-Shafer (DS) decision making systems and conventional Bayesian (centralized or distributed) detectors. We establish theoretically that under certain intuitively pleasing DS probability mass assignments the two systems become equivalent. A few numerical studies illustrate the theoretical developments.

Current research in biomedical informatics involves analysis of multiple heterogeneous data sets. This includes patient demographics, clinical and pathology data, treatment history, patient outcomes as well as gene expression, DNA sequences and other information sources such as gene ontologies. Analysis of these data sets could lead to better disease diagnosis, prognosis, treatment and drug discovery. However, the extent of knowledge that can be extracted from individual data sets is limited Recently, there has been a lot of focus on techniques that analyze genomic data sources in an integrated manner through information fusion. This places a need for an online platform to analyze biomedical informatics data sets using these techniques. We present here an online data warehouse to perform data exploration and analysis across heterogeneous biomedical informatics data sets with the aid of information fusion. The prototype platform is available at http://biogeowarehouse.cse.psu.edu.

MOTIVATION: Genome sequencing projects and high-through-put technologies like DNA and Protein arrays have resulted in a very large amount of information-rich data. Microarray experimental data are a valuable, but limited source for inferring gene regulation mechanisms on a genomic scale. Additional information such as promoter sequences of genes/DNA binding motifs, gene ontologies, and location data, when combined with gene expression analysis can increase the statistical significance of the finding. This paper introduces a machine learning approach to information fusion for combining heterogeneous genomic data. The algorithm uses an unsupervised joint learning mechanism that identifies clusters of genes using the combined data. RESULTS: The correlation between gene expression time-series patterns obtained from different experimental conditions and the presence of several distinct and repeated motifs in their upstream sequences is examined here using publicly available yeast cell-cycle data. The results show that the combined learning approach taken here identifies correlated genes effectively. The algorithm provides an automated clustering method, but allows the user to specify apriori the influence of each data type on the final clustering using probabilities. AVAILABILITY: Software code is available by request from the first author. CONTACT: [email protected].

Quality of service (QoS) routing between domains is indispensable for deploying QoS in the Internet. State aggregation is a technique that makes QoS routing scalable to large networks. As the impreciseness in the advertised aggregate adversely affects the performance of the QoS routing process, it is important to investigate the accuracies of various aggregation schemes. In this paper, we study the efficacy of a new bandwidth aggregation scheme, and compare our scheme's performance with that of the existing schemes. Our analytical and experimental studies show that for active networks, the proposed aggregation scheme is more accurate, and gives a better routing performance than the current schemes.

MOTIVATION: Arrays allow measurements of the expression levels of thousands of mRNAs to be made simultaneously. The resulting data sets are information rich but require extensive mining to enhance their usefulness. Information theoretic methods are capable of assessing similarities and dissimilarities between data distributions and may be suited to the analysis of gene expression experiments. The purpose of this study was to investigate information theoretic data mining approaches to discover temporal patterns of gene expression from array-derived gene expression data. RESULTS: The Kullback-Leibler divergence, an information-theoretic distance that measures the relative dissimilarity between two data distribution profiles, was used in conjunction with an unsupervised self-organizing map algorithm. Two published, array-derived gene expression data sets were analyzed. The patterns obtained with the KL clustering method were found to be superior to those obtained with the hierarchical clustering algorithm using the Pearson correlation distance measure. The biological significance of the results was also examined. AVAILABILITY: Software code is available by request from the authors. All programs were written in ANSI C and Matlab (Mathworks Inc., Natick, MA).

QoS routing is the process of routing a connection based on the connection's resource requirements. The overhead involved in QoS routing increases with the network size. State aggregation is an important technique that makes QoS routing scalable. We propose a new domain state aggregation technique based on "network-flows", for stochastic networks. We propose a new fast heuristic algorithm to estimate the PMF of "maximum-flow" through a probabilistic flow network. Using this algorithm, we aggregate the bandwidth information in a domain and use it along with the path aggregate for routing. Through simulations, we show that stochastic estimation of traffic capacity is better than deterministic estimation and gives a better bandwidth admission ratio when compared with existing state aggregation techniques.

QoS routing is the process of routing a connection based on the connection's resource requirements. The overhead involved in QoS routing increases with the network size. State aggregation is an important technique that helps to reduce the overhead. We propose a new state aggregation technique based on "network-flow". Our approach allows a domain to update the aggregate sent by other domains and keep track of resource availability in other domains. We study the efficacy of our approach with respect to various network and traffic parameters. Preliminary simulation results show that our approach gives a better bandwidth admission ratio when compared with existing techniques.

Multimedia applications such as video-conferencing, telemedicine, HDTV, etc. have very stringent quality-of-service (QoS) demands and require a connection-oriented service. For these applications, a path satisfying their requirements in terms of bandwidth, delay buffer, etc. needs to be found. As conventional IP routing is based only on hop counts, it is not suitable for multimedia applications. It is clear that, to route requests that have QoS requirements, existing routers should be made QoS aware and the packet forwarding should be based on QoS parameters. Also, routing protocols like OSPF and RIP must be extended suitably to facilitate QoS routing. The goal of QoS routing algorithms is to find a loop-less path satisfying a given set of constraints on parameters like bandwidth, delay, etc. The path selection process could return either the entire path to the destination or the best next hop for the request. The first case is called "source routing" and the second is referred to as "distributed routing". In this paper, we propose a new distributed QoS routing algorithm for unicast flows, which has a very low call establishment overhead. Our algorithm makes use of existing IP routing protocols such as OSPF and RIP with minimal modifications.

In this paper, we propose a new deformable model a robust snake model, which solves the primary problems suffered by the conventional snake, such as contour initialization, proper internal parameter setting and the limited capture range of the external energy. A reformulated internal energy is used to serve the smoothness of snake contour without a contraction of the contour. The external energy combines both region and edge information to enlarge the capture range, and also reduces the requirement of initial contour. Both synthetic and real gray-level images are selected to evaluate the performance of the proposed model. Its implementation show it robust, fast and accurate. Initial experimental results are encouraging.

The goal of QoS routing is to find a loopless path through a network that can satisfy a given set of constraints on parameters such as bandwidth, delay and cost. QoS routing algorithms can be classified into three categories. They are source routing, distributed routing and hierarchical routing. In this paper we propose a distributed QoS routing algorithm that gives high call admission rates in a congested network for constraints on metrics like bandwidth, buffer, and CPU usage with a low message complexity.

In this paper, a knowledge-based, fully automatic method for identifying lung regions in digital chest radiographs is described. The method uses an object-oriented knowledge model to appropriately integrate the anatomical knowledge and image processing routines in lung detection. A series of chest radiographs are employed to test the proposed method; the experimental results are encouraging.

This paper describes the design and analysis of the scheduling algorithm for energy conserving medium access control (EC-MAC), which is a low-power medium access control (MAC) protocol for wireless and mobile ATM networks. We evaluate the scheduling algorithms that have been proposed for traditional ATM networks. Based on the structure of EC-MAC and the characteristics of wireless channel, we propose a new algorithm that can deal with the burst errors and the location-dependent errors. Most scheduling algorithms proposed for either wired or wireless networks were analyzed with homogeneous traffic or multimedia services with simplified traffic models. We analyze our scheduling algorithm with more realistic multimedia traffic models based on H.263 video traces and self-similar data traffic. One of the key goals of the scheduling algorithms is simplicity and fast implementation. Unlike the time-stamped based algorithms, our algorithm does not need to sort the virtual time, and thus, the complexity of the algorithm is reduced significantly.

A new family of scale, rotation and affine transformation invariant parameters, denoted as multiresolution moments, is introduced. The basic underlying theory of this generalized and extensible family is described. Experimental results obtained by using a specific subset of this family of new parameters towards texture similarity measurement are included to illustrate the applicability and robustness of this family of parameters.

This paper describes the design and analysis of the scheduling algorithm for EC-MAC (energy conserving medium access control) (Sivalingam et al.), a low-power medium access control (MAC) protocol for wireless and mobile ATM networks. Based on the structure of EC-MAC and the characteristics of the wireless channel, we propose a new algorithm which can deal with the bursty errors and the location-dependent errors. Most scheduling algorithms proposed for either wired or wireless networks were analyzed with homogeneous traffic or multimedia services with simplified traffic models. We analyze our scheduling algorithm with more realistic multimedia traffic models. One of the key goals of the scheduling algorithm is simplicity and fast implementation. Unlike the time-stamp based algorithm, our algorithm does not need to sort the virtual time, thus reducing the complexity of the algorithm significantly.

Efficient means of content based image indexing is an important problem in large visual information systems. Research in this area is still in its preliminary stages. We present a new method for content based image query from image databases. Our retrieval approach is based on fractal dimension estimation of image features which are generated using various image data transformation techniques. Results are presented for image query using the wavelet transform as the image transform and alternating sequential filters (ASF) for the fractal dimension estimation.

Fractional Brownian motion has been used to model self-similar textures. While using the fractal model, the most important procedure is measuring the Hurst parameter H, which is directly related to the fractal dimension. A maximum likelihood estimator has been applied to estimate the Hurst parameter H on a self-similar texture image. Much of the work done so far has concentrated in the spatial domain. In this paper, we propose an approximate MLE method for estimating H in the Fourier domain. The proposed Fourier-domain MLE method saves computational time, as the spatial-domain MLE needs extensive computations to obtain an inverse matrix. We use synthetic fractal datasets and a human tibia image to study the performance of our method.

Fractals have been successfully used to model "natural" shapes and forms. While using the fractal model, the most important procedure is measuring the fractal parameter H (the Hurst coefficient), which is directly related to the fractal dimension. The Wigner-Ville distribution (WVD) is a time-frequency representation with excellent time and frequency resolutions. We propose a one dimensional WVD method to measure the fractal parameter H. Synthetic fractal images and a human tibia image were used to compare the performance of the WVD method to that of the maximum likelihood estimator (MLE) method and the power spectra method.

One of the basic characteristics of a fractal is its dimension. Estimates of the dimension tend to be very inaccurate, as data size is reduced. This paper presents a novel method for fractal dimension estimation based on the use of continuous pyramids of alternative sequential filters (CASF). The method assumes only the self-similarity property of fractals. It was tested over synthetic images and results are presented. Performance of CASF was compared to that of the power spectra method and the wavelet method for large size images (512/spl times/512). Results show that CASF yields much more accurate estimates than the other two methods. When applied to small images, (16/spl times/16) it was observed that the accuracy of the method was still comparable to that obtained with large images.

The computation of the fractal dimension of a 2-dimensional image has become an important issue in various image processing applications such as texture analysis. The fractal dimension is a measure of the similarity in a multiresolution representation. An alternating sequential filter (ASF), a morphological transformation, is used to generate such a representation. The surface areas at different resolutions, computed using a piecewise planar approximation, are used to estimate the fractal dimension. Comparisons with existing methods are done to show the effectiveness of the proposed method. The robustness to noise is also demonstrated. Testing was done on synthetic and natural textured images.

An improved methodology is presented for the localization of tumors for radiation therapy treatment planning applications. The proposed approach focuses on the integration of MRI, CT, and PET imaging data towards the generation of a mathematically optimal tumor boundary. Fuzzy edges derived from CT and MRI are fused to form an integrated edge map which is employed in conjunction with a region based analysis of the PET scan to guide the evolution of a tumor segmenting model. The presented algorithm yields tumor contours which may be employed directly in the radiation therapy treatment planning process.

The aim of this paper is to find a relationship between alternating sequential filters (ASF) and the morphological sampling theorem (MST) developed by Haralick et al. (1987). The motivation behind this approach is to take advantage of the computational efficiency offered by the MST to implement morphological operations. First, we show alternative proofs for opening and closing in the sampled and unsampled domain using the basis functions. These proofs are important because they show that it possible to obtain any level of a morphological pyramid in one step rather than the traditional two-step procedure. This decomposition is then used to show the relationship of the open-closing in the sampled and unsampled domain. An upper and a lower bound, for the above relationships, are presented. Under certain circumstances, an equivalence is shown for open-closing between the sampled and the unsampled domain. An extension to more complicated algorithms using a union of openings and an intersection of closings is also proposed. Using the Hausdorff metric, it is shown that a morphologically reconstructed image cannot have a better accuracy than twice the radius of the reconstruction structuring element. Binary and gray scale examples are presented.

The storage, display, and manipulation of three dimensional volumetric information requires large amounts of computing resources, both in terms of memory, and processing power. Most existing serial algorithms that display 3-D objects on a 2-D screen are found to be too slow to process the large amounts of volume data in a reasonable time. Hence, one way to increase the performance of the display algorithm is to process individual volume elements (voxels) in parallel. The first part of this paper presents a brief over view of the linear octree data structure which represents 3-D objects by an eight-way branching tree, while the second part focusses on the parallel display of such objects. We have shown that, for an object represented by a linear octree and enclosed in a 2/sup nspl times/2/sup nspl times/2/sup n/ universe, the maximum number of voxels that can be processed in parallel is 3/sup n/, and the maximum number of time steps required to display such an object is 4/sup n/. This paper presents a set of formulae which identify the processing element (PE) as well as the time step in which a given linear octree node is processed. Similarly, a procedure which determines the locational code of a linear octree node which must be processed by a given PE, at some specific time step, is presented, along with a strategy for determining whether a PE is active or idle.>

Modeling textures using Gaussian Markov random fields (GMRF) has been successfully used in classifying textures. However, these models do not perform well for self-similar textures such as those generated from fractional Brownian motion. The authors show that by using the difference images at different scales instead of the original image, one can significantly increase the performance of classifying self-similar texture patterns using GMRF models.>

This paper presents a method for range-speed imaging of moving targets. The approach is based on illuminating the target region with an FM-CW signal. The inversion in this imaging problem involves a partitioning of the echoed signals in the frequency domain followed by a transformation. The system model and inversion in this imaging method does not approximate a moving target's Doppler shift to be a constant over the bandwidth of the transmitted signal and does not put any restrictions on the target's range.

The authors have applied mathematical morphology for fractal analysis on bone X-ray images. The digitized gray level image is treated as a three-dimensional surface whose fractal dimension is calculated by performing a series of dilations on this surface and plotting the area of the resulting set of surfaces against the size of the structuring element. This approach has the added advantage of encoding structural information via the use of a structuring element. The algorithm has been applied to several bone radiographs, and the results demonstrate that the fractal dimension using mathematical morphology gives a robust texture measure of trabecular bone structures.

A multiscale filtering scheme based on the three Matheron axioms for morphological openings is developed. It is shown that opening a signal with a gray scale operator does not introduce additional zero-crossings as one moves to coarser scales. Within this framework, the problem of choosing an appropriate structuring element is studied. In order to obtain a measure of the performance of different structuring elements, the statistical properties of gray scale opening are studied, using a powerful tool in mathematical morphology, namely, basis functions.>

Mathematical morphology was applied for fractal analysis of bone X-ray images. Fractal dimension is calculated by performing a series of dilations on the given image and plotting the surface area of the resulting object against the size of the structuring element. This approach gives the added advantage of encoding structural information via the use of a structuring element. The algorithm was applied to several bone radiographs and the results demonstrate that fractal dimension using mathematical morphology gives us a robust texture measure of trabecular bone structures.>

The aim of this research is to obtain a morphological pyramid with the aid of alternating sequential filters. The authors establish a relationship between alternating sequential filters and the morphological sampling theorem developed by R. Haralick et al. (1987). An alternative proof for opening and closing in the sampled and unsampled domain using the basis functions is shown. This decomposition is then used to show the relationship of the compound mapping opening-closing in the sampled and unsampled domain. An upper and a lower bound for the above relationships are presented. Under certain circumstances, an equivalence is shown for opening-closing between the sampled and the unsampled domain. An extension to more complicated algorithms using union of openings and intersection of closings is also proposed.>

A neural architecture is presented for fusion of multisensory information at the feature level. The inputs to the network are in the form of binary edge patterns from multiple sensors. Each input is processed by a hierarchical neural structure similar to the forward path of MARA, which was previously proposed by the authors (1990). Since the neurons at higher levels are insensitive to distortion, noise, scaling, and displacement of the activities at lower levels, the activities of higher-level neurons in these networks processing different sensor information can be combined. The proposed architecture consists of a fusion path to carry the combined information to indicate the final decision. The fusion architecture is used to recognize the traces of trained patterns in low-quality images from different sensors.>

Presents a hierarchical neural network architecture which can achieve multisensor data fusion at feature level by combining the coherent activities of higher-level neurons in the architecture. The architecture learns very quickly and is highly stable and plastic. Initial results show promise. The authors utilize the fusion architecture to recognize traces of trained patterns in low-quality images after fusion through the architecture

The multilayer adaptive resonance architecture (MARA) is a highly stable and plastic self-organizing neural network which is capable of recognizing, reconstructing, and segmenting the traces of previously learned binary patterns. The recognition and reconstruction properties of the network are invariant with respect to distortion, noise, translation, scaling and partial rotation of the original training patterns. If multiple traces of trained patterns are simultaneously presented, the network can separate and reconstruct the traces separately. A description is given of MARA and an image segmentation scheme using MARA. MARA's ability to segment some biological structures in high-speed synchronous volume CT images of the heart is demonstrated

A 3-D segmentation algorithm is presented to extract object surfaces from computed tomography (CT) and magnetic resonance imaging (MRI) scans. The algorithm is formulated in the framework of mathematical morphology. The generalized morphological operators are proposed for multidimensional segmentation. A priori knowledge of the approximate location of the object surface is communicated to the algorithm via the definition of the search space. The algorithm uses this definition of the search space to obtain the surface candidate elements. The search-space specification reduces the computational cost and increases the reliability of the detected features. The detected surface is represented as a hierarchical combination of patches. Initial results obtained by using the algorithm to segment 4-D images of the beating heart are presented.>