This paper describes a system capable of classifying stochastic, self-affine, nonstationary signals produced by nonlinear systems. The classification and analysis of these signals is important because they are generat...
详细信息
ISBN:
(纸本)0769519865
This paper describes a system capable of classifying stochastic, self-affine, nonstationary signals produced by nonlinear systems. The classification and analysis of these signals is important because they are generated by many real-world processes. The first stage of the signal classification process entails the transformation of the signal into the multifractal dimension domain, through the computation of the variance fractal dimension trajectory (VFDT). Features can then be extracted from the VFDT using a Kohonen self-organizing feature map. The second stage involves the use of a complex domain neural network and a probabilistic neural network to determine the class of a signal based on these extracted features. The results of this paper show that these techniques can be successful in creating a classification system which can obtain correct classification rates of about 87% when performing classification of such signals without knowing the number of classes.
This paper describes a system capable of classifying stochastic, self-affine, nonstationary signals produced by nonlinear systems. The classification and analysis of these signals is important because they are generat...
详细信息
This paper describes a system capable of classifying stochastic, self-affine, nonstationary signals produced by nonlinear systems. The classification and analysis of these signals is important because they are generated by many real-world processes. The first stage of the signal classification process entails the transformation of the signal into the multifractal dimension domain, through the computation of the variance fractal dimension trajectory (VFDT). Features can then be extracted from the VFDT using a Kohonen self-organizing feature map. The second stage involves the use of a complex domain neural network and a probabilistic neural network to determine the class of a signal based on these extracted features. The results of this paper show that these techniques can be successful in creating a classification system which can obtain correct classification rates of about 87% when performing classification of such signals with an unknown number of classes.
In this paper, a novel multifractal approach to the classification of unknown self-affine signals is presented as an improvement over traditional traffic signal classifiers. The fundamental advantages of using multifr...
详细信息
In this paper, a novel multifractal approach to the classification of unknown self-affine signals is presented as an improvement over traditional traffic signal classifiers. The fundamental advantages of using multifractal measures include normalization and a very high compression ratio of a signature of the traffic, thereby leading to faster implementations, and the ability to add new traffic classes without redesigning the traffic classifier. The variance fractal dimension trajectory is used to provide a multifractal "signature" for each type of traffic over its duration, and the modelling of its statistical histograms provides further compression and generalization. Principal component analysis is used to reduce the dimensionality of the data, and the k-means clustering algorithm is used to determine the number of classes in the multifractal signatures. A probabilistic neural network is then trained with these signatures, and its performance on classifying unknown traffic is used to indicate the most likely number of classes in the data.
This paper presents a method of modelling of power transients and their classification. A discrete wavelet transform and multifractal analysis based on a variance fractal dimension trajectory technique are used as too...
详细信息
This paper presents a method of modelling of power transients and their classification. A discrete wavelet transform and multifractal analysis based on a variance fractal dimension trajectory technique are used as tools to analyze the transients for feature extraction. A probabilistic neural network is used as a classifier for classification of transients associated with power system faults and switching. Experiments show that the classification system can achieve classification rate of 99% for power transients, and is robust in noisy environments.
This paper presents an electrocardiogram (ECG) frame (beat) compression scheme using block encoding and windowed-variance techniques, where the ECG frame has already been classified. compression of the complicated ECG...
详细信息
This paper presents an electrocardiogram (ECG) frame (beat) compression scheme using block encoding and windowed-variance techniques, where the ECG frame has already been classified. compression of the complicated ECG frame is converted into a linear time registration problem of segments in this scheme. The segment in the frame is detected and partitioned by the windowed-variance technique. It is computationally inexpensive. A high compression ratio of nearly 50:1 is achieved and not related to the reconstruction error. The normalized percent root-mean-square difference is about 2.53% for a 10-minute ECG recording.
Evaluates the performance of the extended-infomax independent component analysis (ICA) algorithm in a simulated biomedical blind source separation problem. Independent signals representing an alphawave and a heartbeat...
详细信息
Evaluates the performance of the extended-infomax independent component analysis (ICA) algorithm in a simulated biomedical blind source separation problem. Independent signals representing an alphawave and a heartbeat are generated and then mixed linearly in the presence of white or pink noise to simulate a one-minute recording of an electroencephalogram and electrocardiogram. The selected ICA algorithm separates the white and pink noises equally well. The maximum estimation signal-to-noise ratio of the source estimates is equivalent to the added noise level, so the separation is optimum to second-order. The higher-order demixing performance, as measured by the Amari index, indicates that when the noise contamination exceeds the mixing contamination the ICA separation is reduced. These results represent a lower bound to the performance of extended-infomax ICA in noisy, time-correlated electrophysiological conditions.
Presents an electrocardiogram (ECG) frame classification technique realized by a dynamic time warping (DTW) matching technique, which has been used successfully in speech recognition. We use the DTW to classify ECG fr...
详细信息
Presents an electrocardiogram (ECG) frame classification technique realized by a dynamic time warping (DTW) matching technique, which has been used successfully in speech recognition. We use the DTW to classify ECG frames because ECG and speech signals have similar nonstationary characteristics. The DTW mapping function is obtained by searching the frame from its end to start. A threshold is setup for DWT matching residual either to classify an ECG frame or to add a new class. Classification and establishment of a template set are carried out simultaneously. A frame is classified into a category with a minimal residual and satisfying a threshold requirement. A classification residual of 1.33% is achieved by the DTW for a 10-minute ECG recording.
This paper demonstrates the feasibility of using automated netlist partitioning to improve the performance of a single FPGA run-time reconfigurable system. Run-time reconfiguration (RTR), an important model for comput...
详细信息
This paper demonstrates the feasibility of using automated netlist partitioning to improve the performance of a single FPGA run-time reconfigurable system. Run-time reconfiguration (RTR), an important model for computing, can outperform other computing solutions for some applications and exhibit high architectural flexibility. The performance of RTR is heavily influenced by circuit partitioning. While partitioning procedures to enhance RTR have been previously explored, their nature remains poorly understood. A partitioning system has, therefore, been designed and implemented in software to interface functional circuits to an industry standard graph partitioning heuristic and to study its performance. Experimental results demonstrate that a 103 gate, low-power DCT implementation with 105 nets may be bipartitioned with an interpartition bandwidth of 3 nets.
This paper addresses the problem of approximation of arbitrary boundaries and edges resulting from image segmentation. Using MPEG-7 terminology, this problem can be considered as the search for a descriptor for a boun...
详细信息
This paper addresses the problem of approximation of arbitrary boundaries and edges resulting from image segmentation. Using MPEG-7 terminology, this problem can be considered as the search for a descriptor for a boundary feature of an object. The shape of an object is usually described by its boundaries, to differentiate it from the other regions of the image. Traditionally, Bezier curve fitting has been employed to model the boundaries of objects consisting of smooth and crisp curves. However, natural images contain complex shapes that are difficult to describe by such smooth-boundary modelling techniques. Consequently, we focus on fractals and multifractals to model arbitrary boundaries based on the singularity measure property of fractals. We use wavelets to identify boundary control points and reconstruct the self-similar boundaries with any fractal or multifractal dimension, using the midpoint displacement algorithm.
Multimedia involves a myriad of data and multidimensional signals, including not only plain and formatted text, but also mathematical and other symbols, tables, vector and bitmap graphics, images, sound, animation, vi...
详细信息
ISBN:
(纸本)0769517242
Multimedia involves a myriad of data and multidimensional signals, including not only plain and formatted text, but also mathematical and other symbols, tables, vector and bitmap graphics, images, sound, animation, video, and interactive virtual reality objects. compression of such signals is usually necessary to fit them into the available communications channels and digital storage, or for data mining. This paper provides an overview of important compression methods and techniques, including lossless entropy coding techniques designed to reduce the redundancy in the critical multimedia material, as well as lossy coding techniques designed to preserve the relevancy of the noncritical multimedia material. Modern lossy techniques often employ wavelets, wavelet packets, fractals, and neural networks. Progressive image transmission is also employed to deliver the material quickly. The paper also addresses several approaches to blind separation of signal from noise (denoising) to improve the compression, and to the difficult question of objective and subjective image quality assessment through complexity metrics.
暂无评论