Orthogonal Windows Research Articles

Research on bark-frequency spectral coefficients heart sound classification algorithm based on multiple window time-frequency reassignment

The multi-window time-frequency reassignment helps to improve the time-frequency resolution of bark-frequency spectral coefficient (BFSC) analysis of heart sounds. For this purpose, a new heart sound classification algorithm combining feature extraction based on multi-window time-frequency reassignment BFSC with deep learning was proposed in this paper. Firstly, the randomly intercepted heart sound segments are preprocessed with amplitude normalization, the heart sounds were framed and time-frequency rearrangement based on short-time Fourier transforms were computed using multiple orthogonal windows. A smooth spectrum estimate is calculated by arithmetic averaging each of the obtained independent spectra. Finally, the BFSC of reassignment spectrum is extracted as a feature by the Bark filter bank. In this paper, convolutional network and recurrent neural network are used as classifiers for model comparison and performance evaluation of the extracted features. Eventually, the multi-window time-frequency rearrangement improved BFSC method extracts more discriminative features, with a binary classification accuracy of 0.936, a sensitivity of 0.946, and a specificity of 0.922. These results present that the algorithm proposed in this paper does not need to segment the heart sounds and randomly intercepts the heart sound segments, which greatly simplifies the computational process and is expected to be used for screening of congenital heart disease.

Time–frequency domain local spectral analysis of seismic signals with multiple windows

For a non-stationary seismic signal, time–frequency analysis methods often include a time window function that serves as a weighting function and by which the signal is multiplied to form a segment. The time window function often has the highest weighting coefficient for the central sample of the signal segment. For the rest of the segment, there is no adequate representation in the frequency spectrum. Here, I propose to use multiple orthogonal window functions to properly represent the local spectral property in the time–frequency plane and recover the information lost due to time windowing before applying the Fourier transform. First, I propose to construct multiple window functions directly using a stack of Gaussian functions. The weighted average spectrum of the multiple window functions has a flat passband, which is better than the conventional multiple windows. Taking advantage of the linearity of the Fourier transform, we can apply each window to the analytic signal to generate the instantaneous autocorrelation accordingly and form an averaged instantaneous autocorrelation by a weighted sum before performing the Fourier transform to generate the Wigner–Ville distribution (WVD). This multi-window WVD method successfully represents the local spectrum of the non-stationary seismic signal in the time–frequency plane.

Executing convex polytope queries on nD point clouds

Efficient spatial queries are frequently needed to extract useful information from massive nD point clouds. Most previous studies focus on developing solutions for orthogonal window queries, while rarely considering the polytope query. The latter query, which includes the widely adopted polygonal query in 2D, also plays a critical role in many nD spatial applications such as the perspective view selection. Aiming for an nD solution, this paper first formulates a convex nD-polytope for querying. Then, the paper integrates three approximate geometric algorithms – SWEEP, SPHERE, VERTEX, and a linear programming method CPLEX, developing a solution based on an Index-Organized Table (IOT) approach. IOT is applied with space filling curve based clustering and advanced querying mechanism which recursively refines hypercubic nD spaces to approach the query geometry for primary filtering. Results from experiments based on both synthetic and real data have confirmed the superior performance of SWEEP. However, the algorithm may lag behind CPLEX due to pessimistic intersection computation in high dimensional spaces. In a real application, by properly transforming a perspective view selection into a polytope query, the solution achieves a sub-second querying performance using SWEEP. In another flood risk query, SWEEP also leads the others. In general, the robust and efficient solution can be immediately used to address different polytope queries, including those abstract ones whose constraints on combinations of different dimensions are formed into a polytope model. Besides, the knowledge of high-dimensional computations acquired also provides significant guidance for handling more nD GIS issues.

A new registration algorithm of electronic portal imaging devices images based on the automatic detection of bone edges during radiotherapy

The precision and efficiency of the registration of megavolt-level electronic portal imaging devices (EPID) images with the naked eye in the orthogonal window are reduced. This study aims to develop a new registration algorithm with enhanced accuracy and efficiency. Ten setup errors with different translation and rotation were simulated with the phantom. For each error, one set of simulated computer tomography images and EPID images were acquired and registered with the traditional and the new method. The traditional method was performed by two senior physicists with the Varian Offline Review software. The new method is basing on the comparison of the precise contours of the same bone structure in the digital reconstruction radiography images and the EPID images, and the contours were fitted with an automatic edge detection algorithm based on gradient images. The average error of the new method was decreased by 44.44%, 28.33%, 49.09% in the translation of X, Y, and Z axes (The traditional vs. the new: X axes, 0.45 mm vs. 0.25 mm; Y axes, 0.75 mm vs. 0.35 mm; Z axes, 0.55 mm vs. 0.28 mm), 42.86% and 40.48% in the rotation of X and Z axes (The traditional vs. the new: X axes, 0.49° vs. 0.28°; Z axes, 0.42° vs. 0.25°), respectively. The average elapsed time in the new method was reduced by 11.14% (The traditional vs. the new: 44 s vs. 39.1 s). The new registration method has significant advantages of accuracy and efficiency compared with the traditional method.

Development of a High-Pressure, High-Temperature, Optically Accessible Continuous-Flow Vessel for Fuel-Injection Experiments

A vessel has been designed for nonreacting fuel-injection experiments with continuous flow of sweep gas at pressures up to 1380 kPa and temperatures up to 200 °C. Four orthogonal windows provide optical access for high-speed spray-visualization using a fast-pulsed light emitting diode (LED) and a high-speed camera. Initial experiments have been conducted to determine spray characteristics of n-heptane. At room conditions, liquid length and cone angle were 170 mm and 14.5 deg, respectively. With air flow in the chamber at 690 kPa and 100 °C, liquid length was considerably shorter at 92 mm and cone angle was wider at 16.5 deg.

基于周期图的频谱检测改进算法

传统的周期图检测法是一种近似无偏的估计方法，但存在估计偏差和估计方差相互制约的问题。针对此问题，本文采用多窗谱估计，该方法是基于滤波器组理论改进的周期图检测法，利用多个正交窗进行谱估计，有极佳的能量集中性，能解决谱估计存在较大方差的问题。理论推导及仿真结果证明，基于多窗谱的频谱检测算法是一种低方差、高分辨率的频谱检测方法，能有效实现低信噪比条件下的信号检测，且相比于其他检测算法能达到更好的检测性能。 The traditional periodogram spectral estimation method is an approximately unbiased spectral estimation method, however, with the difficulties of estimated bias and estimated variance dilemmatic. To solve this problem, a kind of accurate non-parametric multitaper method was introduced in this paper which is actually regarded as the improved periodogram spectral estimation method based on filter banks. The multitaper method makes use of multiple orthogonal windows for spectral estimation. By this means, the spectrum is estimated with excellent energy concentration. It can solve the problem of spectral estimation variance. The theoretical derivations and simulations show that the multitaper spectral spectrum detecting algorithm is a low variance and high resolution spectrum detection method, especially for the weak signal detection under strong noise background. It has effective performance and outperforms other spectrum detecting methods in low SNR.

Minimum-Variance Multitaper Spectral Estimation on the Sphere

We develop a method to estimate the power spectrum of a stochastic process on the sphere from data of limited geographical coverage. Our approach can be interpreted either as estimating the global power spectrum of a stationary process when only a portion of the data are available for analysis, or estimating the power spectrum from local data under the assumption that the data are locally stationary in a specified region. Restricting a global function to a spatial subdomain -- whether by necessity or by design -- is a windowing operation, and an equation like a convolution in the spectral domain relates the expected value of the windowed power spectrum to the underlying global power spectrum and the known power spectrum of the localization window. The best windows for the purpose of localized spectral analysis have their energy concentrated in the region of interest while possessing the smallest effective bandwidth as possible. Solving an optimization problem in the sense of Slepian (1960) yields a family of orthogonal windows of diminishing spatiospectral localization, the best concentrated of which we propose to use to form a weighted multitaper spectrum estimate in the sense of Thomson (1982). Such an estimate is both more representative of the target region and reduces the estimation variance when compared to estimates formed by any single bandlimited window. We describe how the weights applied to the individual spectral estimates in forming the multitaper estimate can be chosen such that the variance of the estimate is minimized.

Estimation of Modal Damping in Power Networks

This paper describes a new Fourier-based sliding window method for estimating the damping of exponentially decaying modes that occur in power networks as a result of electric disturbances. The key innovation in the new method is the use of multiple orthogonal sliding windows rather than just a single pair of sliding windows. The use of these multiple orthogonal windows allows least-squares averaging strategies to be used, enabling lower variance estimates to be obtained as a result. A statistical analysis is provided, and simulations are presented to illustrate the effectiveness of the new method. The technique is also applied to real power system data.

Application of multitaper methods to passive sonar

While the multitaper spectral estimation method [Thomson, Proc. IEEE 70, 1055–1096 (1982)] is used extensively in time series analysis, it has received much less attention in the array processing literature. The multitaper approach reduces the variance of power estimates by averaging windowed Fourier transforms of the data obtained using a set of orthogonal window functions. One advantage of the multitaper method is that it is designed to work with low sample support. Since multitaper estimates are often formed using a single data snapshot, they are suitable for use in highly nonstationary environments where it is difficult to obtain reliable estimates of the noise and interference statistics. This talk focuses on the application of the multitaper method to the passive sonar array processing problem. Standard techniques exist for detecting the presence of line components using multitaper estimates. This talk describes how to extend these techniques to detect plane waves in sonar data, where the signals are typically modeled as complex Gaussian rather than deterministic. Processing of non‐plane‐wave signals, e.g., normal modes, will also be briefly discussed. [Work supported by an ONR Young Investigator Award.]

Orthogonal queries in segments

We consider theorthgonal clipping problem in a set of segments: Given a set ofn segments ind-dimensional space, we preprocess them into a data structure such that given an orthogonal query window, the segments intersecting it can be counted/reported efficiently. We show that the efficiency of the data structure significantly depends on a geometric discrete parameterK named theProjected-image complexity, which becomes ź(n2) in the worst case but practically much smaller. If we useO(m) space, whereK log4dź7nźmźn log4dź7n, the query time isO((K/m)1/2 logmax{4, 4dź5}n). This is near to an Ω((K/m)1/2) lower bound.

Diagnostics of direct-current-magnetron discharges by the emission-selected computer-tomography technique

The dc magnetron discharges in Ar are commonly used for thin film deposition. It is essential to investigate the two-dimensional emission profile Φλ(z,r) in order to analyze the geometrical structure of the ring-shaped magnetron discharge. In the present work, an emission-selected computer tomography (ESCT) system was constructed to investigate the three-dimensional structure of the magnetron discharge in Ar. Photons from the discharge are detected by two identical scanning mirror systems, movable both parallel and perpendicular to the electrode, positioned outside the two small orthogonal windows on the chamber. Experiments have been performed in a planar magnetron with an aluminum target having a 100 mm diam and a 20 mm spacing in Ar at a pressure of 0.3 Pa, a current of 40 mA, and a voltage of 290 V. Three lines, Ari(419.8 nm), Arii(434.8 nm), and Ali(396.2 nm) are selected for the ESCT analysis and discussion of the structure of the discharge. Results from the tomographic imaging showed that the peak in photon production were found at the radius where the magnetic field became parallel to the electrode. The eroded ringlike pattern on the target surface coincided with the radial distribution of photon production in front of the target, and the radial position of the maximum depth of the erosion exactly corresponded to that of the peak emission. The overall features of dc magnetron sputtering were also shown in detail by the present ESCT technology.