Clustering algorithms in feature space are important methods in image segmentation. The choice of the effective feature parameters and the construction of the clustering method are key problems encountered with cluste...Clustering algorithms in feature space are important methods in image segmentation. The choice of the effective feature parameters and the construction of the clustering method are key problems encountered with clustering algorithms. In this paper, the multifractal dimensions are chosen as the segmentation feature parameters which are extracted from original image and wavelet-transformed image. SOM (Self-Organizing Map) network is applied to cluster the segmentation feature parameters. The experiment shows that the performance of the presented algorithm is very good.展开更多
A novel neural network based iterated function system (IFS) model is presented in this paper while the precondition to ensure the model is also explored. Applying it to some practical data, the given signal can be app...A novel neural network based iterated function system (IFS) model is presented in this paper while the precondition to ensure the model is also explored. Applying it to some practical data, the given signal can be approximated exactly by the attractor generated by this model, which provides another way to resolve fractal inverse problem.展开更多
Electrochemical techniques and fractal theory were employed to study the corrosion behaviors and pits distribution characteristics on the corroded surfaces of 304 stainless steel exposed in FeCl3 solution. Fractal fea...Electrochemical techniques and fractal theory were employed to study the corrosion behaviors and pits distribution characteristics on the corroded surfaces of 304 stainless steel exposed in FeCl3 solution. Fractal features of pits distribution over the corroded surfaces were observed and described by the fractal dimension. A 5-8-2 back-propagation （BP） artificial neural network model for the diagnoses of the pitting corrosion rate and pits deepness of 304 stainless steel under various conditions was developed by considering the fractal dimension as a key parameter for describing the pitting corrosion characteristics. The predicted results are well in agreement with the experimental data of pitting corrosion rate and pit deepness. The max relative errors between their experimental and simulation data are 6.69% and 4.62%, respectively.展开更多
Grain shape of the hot deforming alloy is an important of material. The fractal theory was applied to analyze index to character the microstructure and performance the recrystallized microstructure of Ti-15-3 alloy af...Grain shape of the hot deforming alloy is an important of material. The fractal theory was applied to analyze index to character the microstructure and performance the recrystallized microstructure of Ti-15-3 alloy after hot deformation and solution treatment. The fractal dimensions of recrystallized grains were calculated by slit island method. The influence of processing parameters on fractal dimension and grain size was studied, It has been shown that the shapes of recrystallized grain boundaries are self-similar, and the fractal dimension varies from 1 to 2. With increasing deformation degree and strain rate or decreasing deformation temperature, the fractal dimension of grain boundaries increased and the grain size decreased. So the fractal dimension could characterize the grain shape and size. A neural network model was trained to predict the fractal dimension of recrystallized microstructure and the result is in excellent agreement with the experimental data.展开更多
The back-propagation neural （BPN） network was proposed to model the relationship between the parameters of the dieless draw- ing process and the microstrecmres of the QSi3-1 silicon bronze alloy. Combined with image...The back-propagation neural （BPN） network was proposed to model the relationship between the parameters of the dieless draw- ing process and the microstrecmres of the QSi3-1 silicon bronze alloy. Combined with image processing techniques, grain sizes and grain-boundary morphologies were respectively determined by the quantitative metallographic method and the flactal theory. The outcomes obtained show that the deformed microstructures exhibit typical fractal features, and the boundaries can be characterized quantitatively by ffactal dimensions. With the temperature of 600-800℃ and the drawing speed of 0.67-1.00 mm-s-1, either a lower temperature or a higher speed will cause a smaller grain size together with an elevated fractal dimension. The developed model can be capable for forecasting the microstructure evolution with a minimum error. The average relative errors between the predicted results and the experimental values of grain size and fractal dimension are 3.9% and 0.9%, respectively.展开更多
This paper proposes the fractal patterns classifier for multiple cardiac arrhythmias on field-programmable gate array (FPGA) device. Fractal dimension transformation (FDT) is employed to adjoin the fractal features of...This paper proposes the fractal patterns classifier for multiple cardiac arrhythmias on field-programmable gate array (FPGA) device. Fractal dimension transformation (FDT) is employed to adjoin the fractal features of QRS-complex, including the supraventricular ectopic beat, bundle branch ectopic beat, and ventricular ectopic beat. FDT with fractal dimension (FD) is addressed for constructing various symptomatic patterns, which can produce family functions and enhance features, making clear differences between normal and unhealthy subjects. The probabilistic neural network (PNN) is proposed for recognizing multiple cardiac arrhythmias. Numerical experiments verify the efficiency and higher accuracy with the software simulation in order to formulate the mathematical model logical circuits. FDT results in data self-similarity for the same arrhythmia category, the number of dataset requirement and PNN architecture can be reduced. Its simplified model can be easily embedded in the FPGA chip. The prototype classifier is tested using the MIT-BIH arrhythmia database, and the tests reveal its practicality for monitoring ECG signals.展开更多
土壤养分具有变异性,其变异性特征能够反映区域生态功能和景观格局的动态变化,是土壤的重要属性之一。本文从土壤养分空间变异的研究意义出发,通过对相关文献的阅读分析,简述了国内外关于土壤养分空间变异研究的发展历程和研究现状,分...土壤养分具有变异性,其变异性特征能够反映区域生态功能和景观格局的动态变化,是土壤的重要属性之一。本文从土壤养分空间变异的研究意义出发,通过对相关文献的阅读分析,简述了国内外关于土壤养分空间变异研究的发展历程和研究现状,分析了土壤养分空间变异的来源,简述了地统计学、地理信息系统(Geographic Information System, GIS)、遥感以及人工神经网络等研究方法在土壤养分空间变异研究的应用现状,并对目前研究所存在的问题进行了剖析,在此基础上对未来的研究方向进行了展望。展开更多
This paper presents a scheme for improving encoding time for fractal image compression. The approachcombines feature extraction with domain classification using a selforganizing neural network. Feature extractionreduc...This paper presents a scheme for improving encoding time for fractal image compression. The approachcombines feature extraction with domain classification using a selforganizing neural network. Feature extractionreduces the dimensionalics of the problem and enables the neural network to be trained on an image separate fromthe test image. The seaorganizing network introduces a neighborhood topology for classytcation, and alsoeliminates the need to specify a prior set of appropriate image classes. The network organizes itself according to thedistribution of the image features observed during the training. The paper presents results showing that thisclassification approach can reduce encoding time by two orders of magnitude while maintaining comparableaccuracy and compression performance.展开更多
文摘Clustering algorithms in feature space are important methods in image segmentation. The choice of the effective feature parameters and the construction of the clustering method are key problems encountered with clustering algorithms. In this paper, the multifractal dimensions are chosen as the segmentation feature parameters which are extracted from original image and wavelet-transformed image. SOM (Self-Organizing Map) network is applied to cluster the segmentation feature parameters. The experiment shows that the performance of the presented algorithm is very good.
文摘A novel neural network based iterated function system (IFS) model is presented in this paper while the precondition to ensure the model is also explored. Applying it to some practical data, the given signal can be approximated exactly by the attractor generated by this model, which provides another way to resolve fractal inverse problem.
基金the Natural Science Foundation of Liaoning Province (No.972210)
文摘Electrochemical techniques and fractal theory were employed to study the corrosion behaviors and pits distribution characteristics on the corroded surfaces of 304 stainless steel exposed in FeCl3 solution. Fractal features of pits distribution over the corroded surfaces were observed and described by the fractal dimension. A 5-8-2 back-propagation （BP） artificial neural network model for the diagnoses of the pitting corrosion rate and pits deepness of 304 stainless steel under various conditions was developed by considering the fractal dimension as a key parameter for describing the pitting corrosion characteristics. The predicted results are well in agreement with the experimental data of pitting corrosion rate and pit deepness. The max relative errors between their experimental and simulation data are 6.69% and 4.62%, respectively.
基金supported by the National Natural Science Foundation of China under grant No.50405020.
文摘Grain shape of the hot deforming alloy is an important of material. The fractal theory was applied to analyze index to character the microstructure and performance the recrystallized microstructure of Ti-15-3 alloy after hot deformation and solution treatment. The fractal dimensions of recrystallized grains were calculated by slit island method. The influence of processing parameters on fractal dimension and grain size was studied, It has been shown that the shapes of recrystallized grain boundaries are self-similar, and the fractal dimension varies from 1 to 2. With increasing deformation degree and strain rate or decreasing deformation temperature, the fractal dimension of grain boundaries increased and the grain size decreased. So the fractal dimension could characterize the grain shape and size. A neural network model was trained to predict the fractal dimension of recrystallized microstructure and the result is in excellent agreement with the experimental data.
基金supported by the National Basic Research Priorities Program of China (No.2006CB605200)the National Natu-ral Science Foundation of China (Nos.50674008 and 50634010)+1 种基金the Program for New Century Excellent Talents in Chinese Universities (No.NCET-06-0083)the Foundation of State Key Laboratory for Advanced Metals and Materials (No.2008Z-15)
文摘The back-propagation neural （BPN） network was proposed to model the relationship between the parameters of the dieless draw- ing process and the microstrecmres of the QSi3-1 silicon bronze alloy. Combined with image processing techniques, grain sizes and grain-boundary morphologies were respectively determined by the quantitative metallographic method and the flactal theory. The outcomes obtained show that the deformed microstructures exhibit typical fractal features, and the boundaries can be characterized quantitatively by ffactal dimensions. With the temperature of 600-800℃ and the drawing speed of 0.67-1.00 mm-s-1, either a lower temperature or a higher speed will cause a smaller grain size together with an elevated fractal dimension. The developed model can be capable for forecasting the microstructure evolution with a minimum error. The average relative errors between the predicted results and the experimental values of grain size and fractal dimension are 3.9% and 0.9%, respectively.
文摘This paper proposes the fractal patterns classifier for multiple cardiac arrhythmias on field-programmable gate array (FPGA) device. Fractal dimension transformation (FDT) is employed to adjoin the fractal features of QRS-complex, including the supraventricular ectopic beat, bundle branch ectopic beat, and ventricular ectopic beat. FDT with fractal dimension (FD) is addressed for constructing various symptomatic patterns, which can produce family functions and enhance features, making clear differences between normal and unhealthy subjects. The probabilistic neural network (PNN) is proposed for recognizing multiple cardiac arrhythmias. Numerical experiments verify the efficiency and higher accuracy with the software simulation in order to formulate the mathematical model logical circuits. FDT results in data self-similarity for the same arrhythmia category, the number of dataset requirement and PNN architecture can be reduced. Its simplified model can be easily embedded in the FPGA chip. The prototype classifier is tested using the MIT-BIH arrhythmia database, and the tests reveal its practicality for monitoring ECG signals.
文摘土壤养分具有变异性,其变异性特征能够反映区域生态功能和景观格局的动态变化,是土壤的重要属性之一。本文从土壤养分空间变异的研究意义出发,通过对相关文献的阅读分析,简述了国内外关于土壤养分空间变异研究的发展历程和研究现状,分析了土壤养分空间变异的来源,简述了地统计学、地理信息系统(Geographic Information System, GIS)、遥感以及人工神经网络等研究方法在土壤养分空间变异研究的应用现状,并对目前研究所存在的问题进行了剖析,在此基础上对未来的研究方向进行了展望。
文摘This paper presents a scheme for improving encoding time for fractal image compression. The approachcombines feature extraction with domain classification using a selforganizing neural network. Feature extractionreduces the dimensionalics of the problem and enables the neural network to be trained on an image separate fromthe test image. The seaorganizing network introduces a neighborhood topology for classytcation, and alsoeliminates the need to specify a prior set of appropriate image classes. The network organizes itself according to thedistribution of the image features observed during the training. The paper presents results showing that thisclassification approach can reduce encoding time by two orders of magnitude while maintaining comparableaccuracy and compression performance.
