Showing papers by "Liqing Zhang published in 2005"

ECG Feature Extraction and Classification Using Wavelet Transform and Support Vector Machines

Abstract: This paper presents a new approach to the feature extraction for reliable heart rhythm recognition This system of classification is comprised of three components including data preprocessing, feature extraction and classification of ECG signals Two different feature extraction methods are applied together to obtain the feature vector of ECG data The wavelet transform is used to extract the coefficients of the transform as the features of each ECG segment Simultaneously, autoregressive modelling (AR) is also applied to obtain the temporal structures of ECG waveforms Then the support vector machine (SVM) with Gaussian kernel is used to classify different ECG heart rhythm Computer simulations are provided to verify the performance of the proposed method From computer simulations, the overall accuracy of classification for recognition of 6 heart rhythm types reaches 9968%

Determining the radiometric response function from a single grayscale image

Abstract: A method is presented for computing the radiometric response function of a camera from a single grayscale image. While most previous techniques require a set of registered images with different exposures to obtain response data, our approach capitalizes on a statistical feature of graylevel histograms at edge regions to gain information for radio-metric calibration. Appropriate edge regions are automatically determined by our technique, and a prior model of radiometric response functions is employed to deal with incomplete data. With this single-image method, radiometric calibration becomes possible to perform in many instances where the camera is unknown.

ECG analysis based on PCA and Support Vector Machines

Abstract: Cardiovascular diseases is one of the main courses of death around the world. Electrocardiogram (ECG) supervising is the most important and efficient way of preventing heart attacks. Machine monitoring and analysis of ECG is becoming a major topic of the modern medical research. In this paper, we propose a system to detect cardiac arrhythmia using the ECG data form MIT-BIH database as a reference. The purpose of this paper is to develop an algorithm for recognizing and classifying normal beat, left bundle branch block beat, right bundle branch block beat and premature ventricular contraction (PVC). In order to do so, we extract more than 6000 signals from the original database, each signal representing a single and complete heart beat. We extract the principal characteristics of the signal by means of the principal component analysis (PCA) technique. Support vector machine (SVM) has a major predominance over other classification methods in complicated problems. SVM method is applied to classify the ECG data into the 4 categories of heart diseases. Base on this idea, we achieved better results in comparison with other pattern classification method from our computer simulations

Blind clustering of DNA fragments based on kullback-leibler divergence

Abstract: In whole genome shotgun sequencing when DNA fragments are derived from thousands of microorganisms in the environment sample, traditional alignment methods are impractical to use because of their high computation complexity. In this paper, we take the divergence vector which is consist of Kullback-Leibler divergences of different word lengths as the feature vector. Based on this, we use BP neural network to identify whether two fragments are from the same microorganism and obtain the similarity between fragments. Finally, we develop a new novel method to cluster DNA fragments from different microorganisms into different groups. Experiments show that it performs well.

Similarity analysis of DNA sequences based on the relative entropy

Abstract: This paper investigates the similarity of two sequences, one of the main issues for fragments clustering and classification when sequencing the genomes of microbial communities directly sampled from natural environment. In this paper, we use the relative entropy as a criterion of similarity of two sequences and discuss its characteristics in DNA sequences. A method for evaluating the relative entropy is presented and applied to the comparison between two sequences. With combination of the relative entropy and the length of variables defined in this paper, the similarity of sequences is easily obtained. The SOM and PCA are applied to cluster subsequences from different genomes. Computer simulations verify that the method works well.

Microbial Communities DNA Sequences Classification Based on Transition Matrix

Abstract: The paper investigates DNA reads classification based on transition matrix before assembling reads into configs in sequencing microbial communities directly sampled from natural environment. Traditional methods use dynamic programming algorithm to directly detect overlaps between reads in order to find out whether or not two reads can be assembled into one config. However, in microbial communities, there are many species with unknown quantity and distribution so that majority of reads can be impossibly assembled together. To save such unnecessary computation detecting overlaps, we present a new method of feature extraction that can be used to group highly similar reads likely to be assembled into a genome. The basic idea is to define the dinucleotide as a state in a sequence which is viewed as Markov chain, and the transition probabilities of the neighboring states constitutes sixteen-by-sixteen transition matrix. The matrix is used as the feature of a sequence and its reshaped 1-by-256 matrix is viewed as the input pattern of a classifier. Computer simulations verify that the method works well

Sparse Representation and Synaptic Adaptation of the Visual Sensory System

Abstract: This paper investigates the sparse representation of visual neural information and its learning algorithm. First we introduce a generative statistical model for internal representation of visual neural information. Then the neural computing mechanism for representing sensory information in the generative model is discussed, and learning algorithm is developed for training the parameters in the generative model. Finally computer simulations are provided to illustrate the sparseness of the internal representation of the visual information

Stability analysis of multichannel blind deconvolution

Abstract: In our previous work [1], we studied the geometrical structures of finite impulse response (FIR) filters and present a permutable cascade decovolution structure based on the filter decomposition [2]. The stability of algorithms was not discussed in [1]. In this paper, we further analyze the cascade structures of filter decomposition and obtain the stability conditions of proposed algorithms. Finally, we give some examples to illustrate the stability conditions are sufficiency.

Stability analysis of multichannel blind deconvolution

Abstract: In our previous work [1], we studied the geometrical structures of finite impulse response (FIR) filters and present a permutable cascade decovolution structure based on the filter decomposition [2]. The stability of algorithms was not discussed in [1]. In this paper, we further analyze the cascade structures of filter decomposition and obtain the stability conditions of proposed algorithms. Finally, we give some examples to illustrate the stability conditions are sufficiency.

Sparse Representation and Associative Learning in Multisensory Integration

Abstract: In this paper, we discuss multisensory information representation and associative learning in multimodal integration. First, we provide a brief overview of anatomical structure and neural information pathways of multimodal cortexes. Then we formulate the multimodal integration problem into a framework of statistical learning. The associative learning algorithm is proposed to adapt the representation matrix and the sparseness of representation is also adapted in the statistical learning model