SEARCH · PubMed Health

Results for “Neural Networks, Computer”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 325 records · Page 18Linked to original sources

PubMed2001

Alopex-B: a new, simpler, but yet faster version of the alopex training algorithm.

Experimenting with some changes and simplifications to the Alopex algorithm, we obtained a new faster version (Alopex-B), that also shows lower failure rates on training attempts. Like Alopex, our version is network-architecture independent, does not require error or transfer functions to be differentiable, has a high potential for parallelism, and is stochastic (which helps avoid local minima), but unlike Alopex it follows no annealing scheme, and uses less parameters which makes it simpler to implement and to use.

Algorithms↗

PubMed2001

Segmentation of medical ultrasonic image using hybrid neural network.

Objective. To solve one of the most difficult problems in multi-dimensional reconstruction of medical ultrasonic images: image segmentation. Method. A new segmental method based on hybrid neural network was presented in this paper. The hybrid neural network comprised two phases. The first phase was Kohonens self-organization neural network, which was used to segment and label the image coarsely. The feature vectors of those pixels within a specified distance from the cluster centers were employed to train the second phase--a three-layer perception network using back-propagation (BP) technique. Then the trained BP network was used to label every pixel of the image. In the end, a post-processing stage was used to remove the small isolated points and smooth out the contours of the segmented image. Result. The segmented image had smooth continuous edges, few noises or speckles, and the contour of ventricle was clear and accurate. Conclusion. Our method could segment the ultrasonic images accurately and effectively, and had a lot of advantages compared to traditional methods. The unsupervised segmentation problems could be solved using supervised methods.

Cardiovascular Physiological Phenomena↗

PubMed1997

[Neuronal networks and psychopathology].

Within the past 10 years, computer simulations of neural networks have yielded new insights into the nature of information-processing neural systems. Since 1987, more and more network models relevant to psychiatry have been published, and such models have meanwhile been proposed for almost every psychopathological phenomenon. This paper provides an introduction to the theory of neural networks and gives examples of discoveries made via computer simulations. Network models of psychopathological phenomena are discussed with respect to formal thought disorder and delusions. The relevance of neural network models for psychopathological research and theory formulation is highlighted and it is argued that neural network modeling in psychopathology does not merely introduce a few mathematical gimmicks, but rather represents the beginning of a new basis for psychopathology.

Brain↗

PubMed1996

The dynamics of discrete-time computation, with application to recurrent neural networks and finite state machine extraction.

Recurrent neural networks (RNNs) can learn to perform finite state computations. It is shown that an RNN performing a finite state computation must organize its state space to mimic the states in the minimal deterministic finite state machine that can perform that computation, and a precise description of the attractor structure of such systems is given. This knowledge effectively predicts activation space dynamics, which allows one to understand RNN computation dynamics in spite of complexity in activation dynamics. This theory provides a theoretical framework for understanding finite state machine (FSM) extraction techniques and can be used to improve training methods for RNNs performing FSM computations. This provides an example of a successful approach to understanding a general class of complex systems that has not been explicitly designed, e.g., systems that have evolved or learned their internal structure.

Neural Networks, Computer↗

PubMed1998

Analytical and simulation results for stochastic Fitzhugh-Nagumo neurons and neural networks.

An analytical approach is presented for determining the response of a neuron or of the activity in a network of connected neurons, represented by systems of nonlinear ordinary stochastic differential equations--the Fitzhugh-Nagumo system with Gaussian white noise current. For a single neuron, five equations hold for the first- and second-order central moments of the voltage and recovery variables. From this system we obtain, under certain assumptions, five differential equations for the means, variances, and covariance of the two components. One may use these quantities to estimate the probability that a neuron is emitting an action potential at any given time. The differential equations are solved by numerical methods. We also perform simulations on the stochastic Fitzugh-Nagumo system and compare the results with those obtained from the differential equations for both sustained and intermittent deterministic current inputs with superimposed noise. For intermittent currents, which mimic synaptic input, the agreement between the analytical and simulation results for the moments is excellent. For sustained input, the analytical approximations perform well for small noise as there is excellent agreement for the moments. In addition, the probability that a neuron is spiking as obtained from the empirical distribution of the potential in the simulations gives a result almost identical to that obtained using the analytical approach. However, when there is sustained large-amplitude noise, the analytical method is only accurate for short time intervals. Using the simulation method, we study the distribution of the interspike interval directly from simulated sample paths. We confirm that noise extends the range of input currents over which (nonperiodic) spike trains may exist and investigate the dependence of such firing on the magnitude of the mean input current and the noise amplitude. For networks we find the differential equations for the means, variances, and covariances of the voltage and recovery variables and show how solving them leads to an expression for the probability that a given neuron, or given set of neurons, is firing at time t. Using such expressions one may implement dynamical rules for changing synaptic strengths directly without sampling. The present analytical method applies equally well to temporally nonhomogeneous input currents and is expected to be useful for computational studies of information processing in various nervous system centers.

Action Potentials↗

PubMed1995

High-quality compression of echographic images by neural networks and vector quantisation.

A compression method is presented for medical images based on neural networks and vector quantisation (VQ). The neural net is a perceptron and it is followed by some expressly designed VQ algorithms. The method devotes special attention to the quality of the reconstructed images and to the reduction of artefacts. The compression ratio obtained in the long run with echographic images is greater than 50:1 with good quality.

Algorithms↗

PubMed2002

Identification of Cryptosporidium parvum oocysts by an artificial neural network approach.

Microscopic detection of Cryptosporidium parvum oocysts is time-consuming, requires trained analysts, and is frequently subject to significant human errors. Artificial neural networks (ANN) were developed to help identify immunofluorescently labeled C. parvum oocysts. A total of 525 digitized images of immunofluorescently labeled oocysts, fluorescent microspheres, and other miscellaneous nonoocyst images were employed in the training of the ANN. The images were cropped to a 36- by 36-pixel image, and the cropped images were placed into two categories, oocyst and nonoocyst images. The images were converted to grayscale and processed into a histogram of gray color pixel intensity. Commercially available software was used to develop and train the ANN. The networks were optimized by varying the number of training images, number of hidden neurons, and a combination of these two parameters. The network performance was then evaluated using a set of 362 unique testing images which the network had never "seen" before. Under optimized conditions, the correct identification of authentic oocyst images ranged from 81 to 97%, and the correct identification of nonoocyst images ranged from 78 to 82%, depending on the type of fluorescent antibody that was employed. The results indicate that the ANN developed were able to generalize the training images and subsequently discern previously unseen oocyst images efficiently and reproducibly. Thus, ANN can be used to reduce human errors associated with the microscopic detection of Cryptosporidium oocysts.

Algorithms↗

PubMed1996

A nonlinear Hebbian network that learns to detect disparity in random-dot stereograms.

An intrinsic limitation of linear, Hebbian networks is that they are capable of learning only from the linear pairwise correlations within an input stream. To explore what higher forms of structure could be learned with a nonlinear Hebbian network, we constructed a model network containing a simple form of nonlinearity and we applied it to the problem of learning to detect the disparities present in random-dot stereograms. The network consists of three layers, with nonlinear sigmoidal activation functions in the second-layer units. The nonlinearities allow the second layer to transform the pixel-based representation in the input layer into a new representation based on coupled pairs of left-right inputs. The third layer of the network then clusters patterns occurring on the second-layer outputs according to their disparity via a standard competitive learning rule. Analysis of the network dynamics shows that the second-layer units' nonlinearities interact with the Hebbian learning rule to expand the region over which pairs of left-right inputs are stable. The learning rule is neurobiologically inspired and plausible, and the model may shed light on how the nervous system learns to use coincidence detection in general.

Artificial Intelligence↗

PubMed2000

Neural networks in analog hardware--design and implementation issues.

This paper presents a brief review of some analog hardware implementations of neural networks. Several criteria for the classification of general neural networks implementations are discussed and a taxonomy induced by these criteria is presented. The paper also discusses some characteristics of analog implementations as well as some trade-offs and issues identified in the work reviewed. Parameters such as precision, chip area, power consumption, speed and noise susceptibility are discussed in the context of neural implementations. A unified review of various "VLSI friendly" algorithms is also presented. The paper concludes with some conclusions drawn from the analysis of the implementations presented.

Artificial Intelligence↗

PubMed1993

Neural network feature detector for real-time video signal processing.

The application of artificial neural networks to real-time image processing tasks requires the use of dedicated, high performance hardware. A linear array processor called HANNIBAL has been developed which implements the backpropagation neural learning algorithm on-chip. This paper considers the design of a complete neural system which integrates HANNIBAL into an existing image processing environment. The goals for the design of the system have been set partly by the primary application, namely feature recognition, but mainly by the desire for a flexible, high performance hardware tool for the study and evaluation of range of neural image processing applications.

Algorithms↗

PubMed2003

Topology representing network enables highly accurate classification of protein images taken by cryo electron-microscope without masking.

In single-particle analysis, a three-dimensional (3-D) structure of a protein is constructed using electron microscopy (EM). As these images are very noisy in general, the primary process of this 3-D reconstruction is the classification of images according to their Euler angles, the images in each classified group then being averaged to reduce the noise level. In our newly developed strategy of classification, we introduce a topology representing network (TRN) method. It is a modified method of a growing neural gas network (GNG). In this system, a network structure is automatically determined in response to the images input through a growing process. After learning without a masking procedure, the GNG creates clear averages of the inputs as unit coordinates in multi-dimensional space, which are then utilized for classification. In the process, connections are automatically created between highly related units and their positions are shifted where the inputs are distributed in multi-dimensional space. Consequently, several separated groups of connected units are formed. Although the interrelationship of units in this space are not easily understood, we succeeded in solving this problem by converting the unit positions into two-dimensional (2-D) space, and by further optimizing the unit positions with the simulated annealing (SA) method. In the optimized 2-D map, visualization of the connections of units provided rich information about clustering. As demonstrated here, this method is clearly superior to both the multi-variate statistical analysis (MSA) and the self-organizing map (SOM) as a classification method and provides a first reliable classification method which can be used without masking for very noisy images.

Algorithms↗

PubMed1994

Discovering side-chain correlation in alpha-helices.

Using a new representation for interactions in protein sequences based on correlations between pairs of amino acids, we have examined alpha-helical segments from known protein structures for important interactions. Traditional techniques for representing protein sequences usually make an explicit assumption of conditional independence of residues in the sequences. Protein structure analyses, however, have repeatedly demonstrated the importance of amino acid interactions for structural stability. We have developed an automated program for discovering sequence correlations in sets of aligned protein sequences using standard statistical tests and for representing them with Bayesian networks. In this paper, we demonstrate the power of our discovery program and representation by analyzing pairs of residues from alpha-helices. The sequence correlations we find represent physical and chemical interactions among amino-acid side chains in helical structures. Furthermore, these local interactions are likely to be important for stabilizing and packing alpha-helices. Lastly, we have also detect correlations in side-chain comformations that indicate important structural interactions but which don't appear as sequence correlations.

Computer Simulation↗

PubMed1994

Computer-assisted image classification: use of neural networks in anatomic pathology.

Artificial neural networks (ANN) implemented on digital computers have received much attention for interpretation of images in pathology and cytology. Most such images are too complex for current ANN to interpret directly; instead, ANN usually classify the images according to numeric features extracted from them. In experiments on three distinct image classification problems, ANN classifiers performed as well or better than multivariate linear discriminant analysis (a traditional parametric statistical classifier). ANN empirically define non-linear multivariate decision boundaries, and can combine non-contiguous feature areas in mapping a classification. However, many training cases are required in order to map complex area boundaries precisely and with a low risk of 'overtraining.' Careful problem selection and attention to data dimensionality and format are important for efficient ANN use.

Adenocarcinoma↗

PubMed1998

Computer-aided diagnosis: a neural-network-based approach to lung nodule detection.

In this work, we have developed a computer-aided diagnosis system, based on a two-level artificial neural network (ANN) architecture. This was trained, tested, and evaluated specifically on the problem of detecting lung cancer nodules found on digitized chest radiographs. The first ANN performs the detection of suspicious regions in a low-resolution image. The input to the second ANN are the curvature peaks computed for all pixels in each suspicious region. This comes from the fact that small tumors possess and identifiable signature in curvature-peak feature space, where curvature is the local curvature of the image data when viewed as a relief map. The output of this network is thresholded at a chosen level of significance to give a positive detection. Tests are performed using 60 radiographs taken from routine clinic with 90 real nodules and 288 simulated nodules. We employed free-response receiver operating characteristics method with the mean number of false positives (FP's) and the sensitivity as performance indexes to evaluate all the simulation results. The combination of the two networks provide results of 89%-96% sensitivity and 5-7 FP's/image, depending on the size of the nodules.

Diagnosis, Computer-Assisted↗

PubMed1999

The Telemedicine benchmark--a general tool to measure and compare the performance of video conferencing equipment in the telemedicine area.

In this paper, we describe the 'Telemedicine Benchmark' (TMB), which is a set of standard procedures, protocols and measurements to test reliability and levels of performance of data exchange in a telemedicine session. We have put special emphasis on medical imaging, i.e. digital image transfer, joint viewing and editing and 3D manipulation. With the TMB, we can compare the aptitude of different video conferencing software systems for telemedicine issues and the effect of different network technologies (ISDN, xDSL, ATM, Ethernet). The evaluation criteria used are length of delays and functionality. For the application of the TMB, a data set containing radiological images and medical reports was set up. Considering the Benchmark protocol, this data set has to be exchanged between the partners of the session. The Benchmark covers file transfer, whiteboard usage, application sharing and volume data analysis and compression. The TMB has proven to be a useful tool in several evaluation issues.

Benchmarking↗

PubMed2001

A model of the interaction between mood and memory.

This paper investigates a neural network model of the interaction between mood and memory. The model has two attractor networks that represent the inferior temporal cortex (IT), which stores representations of visual stimuli, and the amygdala, the activity of which reflects the mood state. The two attractor networks are coupled by forward and backward projections. The model is however generic, and is relevant to understanding the interaction between different pairs of modules in the brain, particularly, as is the case with moods and memories, when there are fewer states represented in one module than in the other. During learning, a large number of patterns are presented to the IT, each paired with one of two mood states represented in the amygdala. The recurrent connections within each module, the forward connections from the memory module to the amygdala, and the backward connections from the amygdala to the memory module, are associatively modified. It is shown how the mood state in the amygdala can influence which memory patterns are recalled in the memory module. Further, it is shown that if there is an existing mood state in the amygdala, it can be difficult to change it even when a retrieval cue is presented to the memory module that is associated with a different mood state. It is also shown that the backprojections from the amygdala to the memory module must be relatively weak if memory retrieval in the memory module is not to be disrupted. The results are relevant to understanding the interaction between structures important in mood and emotion (such as the amygdala and orbitofrontal cortex) and other brain areas involved in storing objects and faces (such as the inferior temporal visual cortex) and memories (such as the hippocampus).

Affect↗

PubMed1997

Dynamic cell structures for the evaluation of keypoints in facial images.

In this contribution Dynamic Cell Structures (DCS network) are applied to classify local image structures at particular facial landmarks. The facial landmarks such as the corners of the eyes or intersections of the iris with the eyelid are computed in advance by a combined model and data driven sequential search strategy. To reduce the detection error after the processing of the sequential search strategy, the computed image positions are verified applying a DCS network. The DCS network is trained by supervised learning with feature vectors which encode spatially arranged edge and structural information at the keypoint position considered. The model driven localization as well as the data driven verification are based on steerable filters, which build a representation comparable with one provided by a receptive field in the human visual system. We apply a DCS based classifier because of its ability to grasp the topological structure of complex input spaces and because it has proved successful in a number of other classification tasks. In our experiments the average error resulting from false positive classifications is less than 1%.

Algorithms↗

PubMed1992

Can artificial neural networks provide an "expert's" view of medical students performances on computer based simulations?

Artificial neural networks were trained to recognize the test selection patterns of students' successful solutions to seven immunology computer based simulations. When new student's test selections were presented to the trained neural network, their problem solutions were correctly classified as successful or non-successful > 90% of the time. Examination of the neural networks output weights after each test selection revealed a progressive increase for the relevant problem suggesting that a successful solution was represented by the neural network as the accumulation of relevant tests. Unsuccessful problem solutions revealed two patterns of students performances. The first pattern was characterized by low neural network output weights for all seven problems reflecting extensive searching and lack of recognition of relevant information. In the second pattern, the output weights from the neural network were biased towards one of the remaining six incorrect problems suggesting that the student mis-represented the current problem as an instance of a previous problem.

Allergy and Immunology↗