Brain tumor detection and division is a difficult tedious undertaking in clinical image preparation.When it comes to the new technology that enables accurate identification of the mysterious tissues of the brain,magne...Brain tumor detection and division is a difficult tedious undertaking in clinical image preparation.When it comes to the new technology that enables accurate identification of the mysterious tissues of the brain,magnetic resonance imaging(MRI)is a great tool.It is possible to alter the tumor’s size and shape at any time for any number of patients by using the Brain picture.Radiologists have a difficult time sorting and classifying tumors from multiple images.Brain tumors may be accurately detected using a new approach called Nonlinear Teager-Kaiser Iterative Infomax Boost Clustering-Based Image Segmentation(NTKFIBC-IS).Teager-Kaiser filtering is used to reduce noise artifacts and improve the quality of images before they are processed.Different clinical characteristics are then retrieved and analyzed statistically to identify brain tumors.The use of a BraTS2015 database enables the proposed approach to be used for both qualitative and quantitative research.This dataset was used to do experimental evaluations on several metrics such as peak signal-to-noise ratios,illness detection accuracy,and false-positive rates as well as disease detection time as a function of a picture count.This segmentation delivers greater accuracy in detecting brain tumors with minimal time consumption and false-positive rates than current stateof-the-art approaches.展开更多
Across the world, we are currently witnessing the deployments of 4 G LTE-Advanced and the 5 G research is reaching its peak point. The 5 G research mainly concentrates on addressing some of the existing OFDM based LTE...Across the world, we are currently witnessing the deployments of 4 G LTE-Advanced and the 5 G research is reaching its peak point. The 5 G research mainly concentrates on addressing some of the existing OFDM based LTE problems along with use of non-contiguous fragmented spectrum. Universal Filtered Multi Carrier(UFMC) has been considered as one of the candidate waveform for the 5 G communications because it provides robustness against the Inter Symbol Interference(ISI), and Inter Carrier Interference(ICI) and is suitable for low latency scenarios. In this paper, a novel approach is proposed to use Kaiser-Bessel filter based pulse shaping instead of standard Dolph-Chebyshev filter for UFMC based waveform to reduce the spectral leakage into nearby sub-bands. In this paper, UFMC system is simulated using MATLAB software, a comparative study for Dolph-Chebyshev and Kaiser-Bessel filters are performed and the results are also presented in terms of power spectrum density(PSD) analysis, Complementary Cumulative Distribution Function(CCDF) analysis, and Adjacent Channel Power Ratio(ACPR) analysis. The simulated results show a better power spectral density and lower sidebands for UFMC(Kaiser Based window), when compared with UFMC(Dolph-Chebyshev) and conventional OFDM.展开更多
针对S变换在电能质量扰动检测中存在计算量过大,时频分辨率低,电能质量扰动数据集常具备类别不平衡的问题,提出一种基于改进Kaiser窗快速S变换(modified Kaiser window fast S-transform,FMKST)和轻梯度提升机(light gradient boosting ...针对S变换在电能质量扰动检测中存在计算量过大,时频分辨率低,电能质量扰动数据集常具备类别不平衡的问题,提出一种基于改进Kaiser窗快速S变换(modified Kaiser window fast S-transform,FMKST)和轻梯度提升机(light gradient boosting machine,LightGBM)的电能质量扰动识别与分类新方法。首先通过快速傅里叶变换得到采样信号频谱;然后利用迭代循环滤波区间定位算法确定扰动频率区间;再根据扰动频率区间所处频段确定窗宽调节因子并对相应区间进行变换;最后从采样信号的FMKST模时频矩阵中提取特征向量并构建改进LightGBM分类器进行分类。仿真与实验结果表明,提出的方法具有更高的识别准确率与更快的诊断速度,适用于海量电能质量扰动数据的快速识别与分类。展开更多
The evaluation of distortion diagnosis using Wavelet function for Electrocardiogram (ECG), Electroencephalogram (EEG) and Phonocardiography (PCG) is not novel. However, some of the technological and economic issues re...The evaluation of distortion diagnosis using Wavelet function for Electrocardiogram (ECG), Electroencephalogram (EEG) and Phonocardiography (PCG) is not novel. However, some of the technological and economic issues remain challenging. The work in this paper is focusing on the reduction of the noise interferences and analyzes different kinds of ECG signals. Furthermore, a physiological monitoring system with a programming model for the filtration of ECG is presented. Kaiser based Finite Impulse Response (FIR) filter is used for noise reduction and identification of R peaks based on Peak Detection Algorithm (PDA). Two approaches are implemented for detecting the R peaks;Amplitude Threshold Value (ATV) and Peak Prediction Technique (PPT). Daubechies wavelet transform is applied to analyze the ECG of driver under stress, arrhythmia and sudden cardiac arrest signals. From the obtained results, it was found that the PPT is an effective and efficient technique in detecting the R peaks compared to ATV.展开更多
文摘Brain tumor detection and division is a difficult tedious undertaking in clinical image preparation.When it comes to the new technology that enables accurate identification of the mysterious tissues of the brain,magnetic resonance imaging(MRI)is a great tool.It is possible to alter the tumor’s size and shape at any time for any number of patients by using the Brain picture.Radiologists have a difficult time sorting and classifying tumors from multiple images.Brain tumors may be accurately detected using a new approach called Nonlinear Teager-Kaiser Iterative Infomax Boost Clustering-Based Image Segmentation(NTKFIBC-IS).Teager-Kaiser filtering is used to reduce noise artifacts and improve the quality of images before they are processed.Different clinical characteristics are then retrieved and analyzed statistically to identify brain tumors.The use of a BraTS2015 database enables the proposed approach to be used for both qualitative and quantitative research.This dataset was used to do experimental evaluations on several metrics such as peak signal-to-noise ratios,illness detection accuracy,and false-positive rates as well as disease detection time as a function of a picture count.This segmentation delivers greater accuracy in detecting brain tumors with minimal time consumption and false-positive rates than current stateof-the-art approaches.
文摘Across the world, we are currently witnessing the deployments of 4 G LTE-Advanced and the 5 G research is reaching its peak point. The 5 G research mainly concentrates on addressing some of the existing OFDM based LTE problems along with use of non-contiguous fragmented spectrum. Universal Filtered Multi Carrier(UFMC) has been considered as one of the candidate waveform for the 5 G communications because it provides robustness against the Inter Symbol Interference(ISI), and Inter Carrier Interference(ICI) and is suitable for low latency scenarios. In this paper, a novel approach is proposed to use Kaiser-Bessel filter based pulse shaping instead of standard Dolph-Chebyshev filter for UFMC based waveform to reduce the spectral leakage into nearby sub-bands. In this paper, UFMC system is simulated using MATLAB software, a comparative study for Dolph-Chebyshev and Kaiser-Bessel filters are performed and the results are also presented in terms of power spectrum density(PSD) analysis, Complementary Cumulative Distribution Function(CCDF) analysis, and Adjacent Channel Power Ratio(ACPR) analysis. The simulated results show a better power spectral density and lower sidebands for UFMC(Kaiser Based window), when compared with UFMC(Dolph-Chebyshev) and conventional OFDM.
文摘针对S变换在电能质量扰动检测中存在计算量过大,时频分辨率低,电能质量扰动数据集常具备类别不平衡的问题,提出一种基于改进Kaiser窗快速S变换(modified Kaiser window fast S-transform,FMKST)和轻梯度提升机(light gradient boosting machine,LightGBM)的电能质量扰动识别与分类新方法。首先通过快速傅里叶变换得到采样信号频谱;然后利用迭代循环滤波区间定位算法确定扰动频率区间;再根据扰动频率区间所处频段确定窗宽调节因子并对相应区间进行变换;最后从采样信号的FMKST模时频矩阵中提取特征向量并构建改进LightGBM分类器进行分类。仿真与实验结果表明,提出的方法具有更高的识别准确率与更快的诊断速度,适用于海量电能质量扰动数据的快速识别与分类。
文摘The evaluation of distortion diagnosis using Wavelet function for Electrocardiogram (ECG), Electroencephalogram (EEG) and Phonocardiography (PCG) is not novel. However, some of the technological and economic issues remain challenging. The work in this paper is focusing on the reduction of the noise interferences and analyzes different kinds of ECG signals. Furthermore, a physiological monitoring system with a programming model for the filtration of ECG is presented. Kaiser based Finite Impulse Response (FIR) filter is used for noise reduction and identification of R peaks based on Peak Detection Algorithm (PDA). Two approaches are implemented for detecting the R peaks;Amplitude Threshold Value (ATV) and Peak Prediction Technique (PPT). Daubechies wavelet transform is applied to analyze the ECG of driver under stress, arrhythmia and sudden cardiac arrest signals. From the obtained results, it was found that the PPT is an effective and efficient technique in detecting the R peaks compared to ATV.
