image segmentation is almost always a necessary step in imageprocessing. The employed threshold algorithms are based on the detection of local minima in the gray level histograms of the entire image. In automatic cel...
详细信息
ISBN:
(纸本)0819419230
image segmentation is almost always a necessary step in imageprocessing. The employed threshold algorithms are based on the detection of local minima in the gray level histograms of the entire image. In automatic cell recognition equipment, like chromosome analysis or micronuclei counting systems, flexible and adaptive thresholds are required to consider variation in gray level intensities of the background and of the specimen. We have studied three different methods of threshold determination: 1) a statistical procedure, which uses the interclass entropy maximization of the gray level histogram. The iterative algorithm can be used for multithreshold segmentation. The contribution of iteration step 'i' is 2i-1 number of thresholds;2) a numerical approach, which detects local minima in the gray level histogram. The algorithm must be tailored and optimized for specific applications like cell recognition with two different thresholds for cell nuclei and cell cytoplasm segmentation;3) an artificialneural network, which is trained with learning sets of image histograms and the corresponding interactively determined thresholds. We have investigated feed forward networks with one and two layers, respectively. The gray level frequencies are used as inputs for the net. The number of different thresholds per image determines the output channels. We have tested and compared these different threshold algorithms for practical use in fluorescence microscopy as well as in bright field microscopy. The implementation and the results are presented and discussed.
To fully exploit the real-time computational capabilities of neuralnetworks (NN) - as applied to imageprocessingapplications - a high performance VMEbus based analog neurocomputing architecture (VMENA) is developed...
详细信息
In this paper an overview of Cellular neuralnetworks (CNNs) and their applications is reported. CNNs are nonlinear dynamical systems with a large number of state variables. Moreover, these artificial systems have bee...
详细信息
In this paper an overview of Cellular neuralnetworks (CNNs) and their applications is reported. CNNs are nonlinear dynamical systems with a large number of state variables. Moreover, these artificial systems have been often applied to the modelling and simulation of other large scale systems in physics, biology and a lot of other different areas because of their powerful real-time processing capabilities. The CNNs basics and their main applications reported in literature are dealt with.
neuralnetworks have been used to classify high resolution remote-sensed data. Experiments have demonstrated the potential of neuralnetworks for clustering a large number of ground cover instances using supervised me...
详细信息
neuralnetworks have been used to classify high resolution remote-sensed data. Experiments have demonstrated the potential of neuralnetworks for clustering a large number of ground cover instances using supervised methods. The paper describes a new algorithm of unsupervised learning, based on artificialneuralnetworks. Its performance has been compared with the competitive learning algorithm. The efficiency of this approach has been demonstrated through experimental results obtained on the real-world of multispectral remote sensing data.< >
The multi layer perceptron is one of the most popular artificialneuralnetworks with applications to e.g. signal and imageprocessing as well as pattern recognition and classification. In order to implement this netw...
详细信息
The multi layer perceptron is one of the most popular artificialneuralnetworks with applications to e.g. signal and imageprocessing as well as pattern recognition and classification. In order to implement this network and the corresponding learning algorithm in electronic or optoelectronic hardware, the synaptic weights have to be discretized. This discretization, however, poses difficulties for the successful training of a multi layer perceptron. Variants of the FLETCHER-REEVES algorithm are compared to the genetic algorithm as learning strategies for the multi layer perceptron with discretized synaptic weights. Simulation results with the XOR problem as benchmark are given.
This paper presents a study of extended learning mechanisms for Kosko's bidirectional associative memory-a classical artificialneural network with applications to, for example, pattern recognition and image proce...
详细信息
This paper presents a study of extended learning mechanisms for Kosko's bidirectional associative memory-a classical artificialneural network with applications to, for example, pattern recognition and imageprocessing. The incorporation of sigma-pi (/spl Sigma//spl Pi/)-units instead of the conventional /spl Sigma/-units in the BAM is considered leading to the /spl Sigma//spl Pi/-BAM. A Hebbian learning algorithm for /spl Sigma//spl Pi/-units is proposed and simulation results are given, indicating the increased performance of the /spl Sigma//spl Pi/-BAM as a pattern association device.
The Airborne Visible InfraRed Imaging Spectrometer (AVIRIS), presently being flown by the Jet Propulsion Laboratory, acquires images of the earth in the visible and reflected infrared. The wavelengths of the measured ...
详细信息
ISBN:
(纸本)0819419281
The Airborne Visible InfraRed Imaging Spectrometer (AVIRIS), presently being flown by the Jet Propulsion Laboratory, acquires images of the earth in the visible and reflected infrared. The wavelengths of the measured radiation range from about 400 nm to 2400 nm and are divided into 224 contiguous channels having a nominal spectral bandwidth of 10 nm. This means a high resolution radiance spectrum is acquired for each 20 m × 20 m ground cell in the AVIRIS scene. Geologic mapping from such data is possible by classifying each pixel based on the distinctive spectral signatures recorded in the channels. artificialneuralnetworks (ANN) have used these spectra successfully to classify an AVIRIS subscene of the Lunar Craters Volcanic Field (LCVF) in Nye County, Nevada. The size and number of spectra in an AVIRIS scene makes classifying these images a computationally intensive task. By classifying the data in a compressed format, savings in computer time may be realized. The wavelet bases have the desirable property of rendering signals similar to the AVIRIS spectra sparse in the wavelet domain. In this investigation, the discrete wavelet transform was applied to the spectra. This produced a set of wavelet coefficients for the spectra that could be made sparse with seemingly little loss of accuracy. Small subsets of the wavelet coefficients were used to classify the LCVF scene by ANN. The degree to which information was lost in the wavelet transform and the elimination of wavelet coefficients from the classification was assessed by making comparisons between the different ANN classifications. The ANN was chosen over more conventional classifiers because of its proven sensitivity in distinguishing subtle but geologically relevant features in these spectra.
Singular value decomposition, which has previously been applied to the problem of signal extraction from marine data is currently being implemented for land use classification. neuralnetworks are becoming increasingl...
详细信息
Singular value decomposition, which has previously been applied to the problem of signal extraction from marine data is currently being implemented for land use classification. neuralnetworks are becoming increasingly popular for the characterisation of multispectral remote sensing data. There are a number of significant problems with using ANN as classifiers for this type of data. A comparison of these two procedures is performed and there merits and difficulties discussed. The authors introduce the Levenberg Marquardt technique as an advanced method for finding the global minima during a backpropagation training scenario. These techniques are applied to simulated data generated from Landsat TM and SPOT satellite data of the County Wicklow area of Ireland. This data comprises five classes including Sitka spruce and Scots pine.< >
The task of finding whether vertebral levels of the human spine can be mathematically differentiated is posed, using a morphometric database of the vertebral quantitative morphology as a training set. There is no cert...
详细信息
The task of finding whether vertebral levels of the human spine can be mathematically differentiated is posed, using a morphometric database of the vertebral quantitative morphology as a training set. There is no certainty in the medical community about this possibility. The axial projections of a number of vertebrae were digitized and their features quantified by automatic image analysis algorithms. These measurements build a spinal morphometric database. After the application of feature selection procedures, the selected measurements are used as inputs to different pattern recognition algorithms try to find whether vertebral levels can be distinguished. The authors have proved the capability to fulfil this classification with a small degree of uncertainty. artificialneuralnetworks, in particular, have shown their capability to perform well in this difficult pattern recognition task.< >
Both genetic algorithms (GAs) and artificialneuralnetworks (ANNs) (connectionist learning models) are effective generalisations of successful biological techniques to the artificial realm. Both techniques are inhere...
详细信息
ISBN:
(纸本)9780897918169
Both genetic algorithms (GAs) and artificialneuralnetworks (ANNs) (connectionist learning models) are effective generalisations of successful biological techniques to the artificial realm. Both techniques are inherently parallel and seem ideal for implementation on the current generation of parallel supercomputers. We consider how the two techniques complement each other and how combining them (i.e. evolving artificialneuralnetworks with a genetic algorithm), may give insights into the evolution of structure and modularity in biological brains. The incorporation of evolutionary and modularity concepts into artificial systems has the potential to decrease the development time of ANNs for specific image and information processingapplications. General considerations when genetically encoding ANNs are discussed, and a new encoding method developed, which has the potential to simplify the generation of complex modular networks. The implementation of this technique on a CM-5 parallel supercomputer raises many practical and theoretical questions in the application and use of evolutionary models with artificialneuralnetworks.
暂无评论