Proper Window Size Research Articles

Minimizing the total strain error in point-wise least squares using rotated gaussian weight strain filter (RGW-SF) in digital image correlation

The point-wise least squares (PLS) has been widely accepted as a simple and trustworthy strain computation method that smooths and differentiates the noisy displacement field calculated directly by digital image correlation (DIC). Smooth window size and kernel function need to be determined before using PLS. The selection of the proper smooth window size has long been a tricky issue for practitioners, which has an obvious impact on the final total strain error of complex deformation. A small smooth window may reduce the system/bias error but increase the random error, and vice versa. In this paper, a rotated Gaussian weight strain filter (RGW-SF) is proposed for heterogeneous strain field measurement. The motivation of RGW-SF is to adapt the weights of each point in tradition PLS smooth window to obtain lower total strain error, which is derived and then minimized to select an optimum strain filter with three weight parameters (i.e., σx,σy, and θ). The effectiveness of the proposed RGW-SF is verified using simulated sinusoidal deformation images and the Star 6 image set from DIC-challenge 2.0. Experiments show that the RGW-SF can find a trade-off between the system error and the random error. The displacement and strain field accuracy, especially for the metrological efficiency indicator, by RGW-SF is much better than that of the PLS and the self-adaptive strain algorithm (SSA). The proposed RGW-SF is strongly suggested for unknown severely heterogeneous strain measurement.

Overview of the EEG-Based Classification of Motor Imagery Activities Using Machine Learning Methods and Inference Acceleration with FPGA-Based Cards

In this article, we provide a brief overview of the EEG-based classification of motor imagery activities using machine learning methods. We examined the effect of data segmentation and different neural network structures. By applying proper window size and using a purely convolutional neural network, we achieved 97.7% recognition accuracy on data from twenty subjects in three classes. The proposed architecture outperforms several networks used in previous research and makes the motor imagery-based BCI more efficient in some applications. In addition, we examined the performance of the neural network on a FPGA-based card and compared it with the inference speed and accuracy provided by a general-purpose processor.

MGRank: A keyword extraction system based on multigraph GoW model and novel edge weighting procedure

Keyword extraction is the process of extracting the most descriptive words from a textual document. State-of-the-art graph-based keyword extraction systems generally represent text documents using a simple graph called a graph-of-words (GoW), based on the sliding window concept. This representation of a text document requires determining a proper window size, models the document on a local scale, and allows the establishment of a single relation between two candidate keywords. In this study, we address these problems and propose a keyword extraction system called MGRank which uses a complete multigraph structure to build a GoW model to represent a text document. The completeness property of the proposed GoW model provides a means to represent a document globally and eliminates the need to determine the window-size parameter. Parallel edges allow the establishment of multiple relations between candidate keywords. In this study, we also propose a new edge-weighting procedure based on the positional distance of candidate keywords. To evaluate the performance of MGRank, we performed experiments on seven benchmark datasets and compared the results with those of six baseline methods. The experimental results show that MGRank outperforms the baseline methods statistically in precision, recall, and F1-score in almost all cases. In terms of mean average precision and mean reciprocal rank, MGRank performs statistically better than node ranking-based and statistical baseline methods and achieves on-par results with topic-based baseline methods. Furthermore, the experimental results showed that MGRank extracted the most relevant keywords.

Determining the Rolling Window Size of Deep Neural Network Based Models on Time Series Forecasting

Time series forecasting has always been a significant task in various domains. In this paper, we propose DeepARMA, a LSTM-based recurrent neural network to tackle this problem. DeepARMA is derived from an existing time series forecasting baseline, DeepAR, overcoming two of its weaknesses: (1) rolling window size determination: the way DeepAR determines rolling window size is casual and vulnerable, which may lead to the unnecessary computation and inefficiency of the model;(2) neglect of the noise: pure autoregressive model cannot deal with the condition where data are composed of various kinds of noise, neither do most of time series models including DeepAR. In order to solve these two problems, we first combine a classic information theoretic criterion, AIC, with the network to determine the proper rolling window size. Then, we propose a jointly-learned neural network fusing white Gaussian noise series given by ARIMA models to DeepAR’s input. That is exactly why we name the network ‘DeepARMA’. Our experiments on a real-world dataset demonstrate that our improvement settles those two problems put forward above.

A Sliding-Window Principal Component thermography reconstruction approach for enhancement and identification of electronic components internal structure

Electronic supply chain vulnerability has been threatened by counterfeits for decades and as a result, the authentication and quality of Electronic Components (ECs) are highly valued by stakeholders for safety-critical industrial systems. Thermography Nondestructive Testing (NDT), offering a rapid inspection while covering a large area within a short time frame, is a promising technique to inspect the integrity of ECs. However, recognising the internal structural layout directly through a packaged EC from thermography images is still a challenge due to its sealed and cramped structure. In this research, we propose a Pulsed Thermography (PT) based data analysis method for enhancement and identification of the layout of die and lead frame in ECs. The proposed solution combines the Sliding Window (SW) concept, Thermographic Signal Reconstruction (TSR) and Principal Component Thermographic (PCT) technique, named SWPCT, to enhance the contrast of die, die substrate and lead frame. Compared to the conventional full-scale PT image analysis, the proposed method presents its superior sensitivity in small thermal features by reducing the spatial thermal information redundancy benefiting from the sliding window PCT analysis. It overcomes the structural thermograms obstruction from the thick packaging, reveals the hidden die and lead frame layout and suppresses the noise simultaneously. By selecting a proper window size, this paper reveals that SWPCT can reconstruct the position and dimension of die and die substrate in ECs, and spatially distinguish the lead frame layout. Two groups of typical OpAmps chip samples with internal structure diversity (X-ray observed) were tested using PT to demonstrate the effectiveness of the proposed method. Simulation data were also used to verify this approach. This research could unlock the challenge of thermography-based inspection for key internal structure in the encapsulated integrated components.

Model‐agnostic online forecasting for PV power output

A reliable forecasting model is required for photovoltaic (PV) power output because solar energy is highly volatile. Another driver for the need of a reliable forecasting model is concept drift, which means that the statistical properties of the data change over time. In this paper, an online forecasting method to handle concept drift is proposed. First, the problem of forecasting in batch learning is transformed into a forecasting in online learning setting. Then, an online learning algorithm is applied, which is good for handling concept drift. Through experiments using the real-world data, it is shown that the method noticeably improves performance compared to the case where a trained model is used. Under various concept drift scenarios, the method improves performance by up to 87.3%. It is also shown that the re-training method (a representative existing method) has several limitations. This method requires several issues to be solved, such as selection of a proper window size, and this is evident through results showing different performance under different settings. In contrast, the method shows a reliable and desirable performance under various concept drift scenarios and thus outperforms the re-training method. The method improves performance by up to 79%.

DeepDensity: Convolutional neural network based estimation of local fringe pattern density

Fringe pattern based measurement techniques are crucial both in macroscale, e.g., fringe projection profilometry, and microscale, e.g., label-free quantitative phase microscopy. Accurate estimation of the local fringe density map can significantly facilitate almost all stages of fringe pattern analysis process. Example includes: (1) using density map as a determinant for the selection of the proper window size in windowed Fourier transform method, (2) guiding the decomposition process in empirical mode decomposition, (3) improving the phase unwrapping accuracy by providing additional reliability indicators, (4) guiding phase estimation process in regularized phase tracking. For these reasons, the accurate and robust estimation of local fringe density map is of high importance and can boost fringe pattern analysis on different stages of processing path, resulting in increased capacity of the full-field noncontact/noninvasive optical measurement system. In this paper, we propose a new, accurate, robust, and fast numerical solution for local fringe density map estimation called DeepDensity. DeepDensity is based on the convolutional neural network and deep learning, making it significantly outperform other conventional solutions to this problem. Numerical simulations and experimental results corroborate the effectiveness of the proposed DeepDensity.

A high-performance approach for predicting donor splice sites based on short window size and imbalanced large samples

BackgroundSplice sites prediction has been a long-standing problem in bioinformatics. Although many computational approaches developed for splice site prediction have achieved satisfactory accuracy, further improvement in predictive accuracy is significant, for it is contributing to predict gene structure more accurately. Determining a proper window size before prediction is necessary. Overly long window size may introduce some irrelevant features, which would reduce predictive accuracy, while the use of short window size with maximum information may performs better in terms of predictive accuracy and time cost. Furthermore, the number of false splice sites following the GT–AG rule far exceeds that of true splice sites, accurate and rapid prediction of splice sites using imbalanced large samples has always been a challenge. Therefore, based on the short window size and imbalanced large samples, we developed a new computational method named chi-square decision table (χ2-DT) for donor splice site prediction.ResultsUsing a short window size of 11 bp, χ2-DT extracts the improved positional features and compositional features based on chi-square test, then introduces features one by one based on information gain, and constructs a balanced decision table aimed at implementing imbalanced pattern classification. With a 2000:271,132 (true sites:false sites) training set, χ2-DT achieves the highest independent test accuracy (93.34%) when compared with three classifiers (random forest, artificial neural network, and relaxed variable kernel density estimator) and takes a short computation time (89 s). χ2-DT also exhibits good independent test accuracy (92.40%), when validated with BG-570 mutated sequences with frameshift errors (nucleotide insertions and deletions). Moreover, χ2-DT is compared with the long-window size-based methods and the short-window size-based methods, and is found to perform better than all of them in terms of predictive accuracy.ConclusionsBased on short window size and imbalanced large samples, the proposed method not only achieves higher predictive accuracy than some existing methods, but also has high computational speed and good robustness against nucleotide insertions and deletions.ReviewersThis article was reviewed by Ryan McGinty, Ph.D. and Dirk Walther.

Nonlocal Similarity Based Nonnegative Tucker Decomposition for Hyperspectral Image Denoising

Compared with color or grayscale images, hyperspectral images deliver more informative representation of ground objects and enhance the performance of many recognition and classification applications. However, hyperspectral images are normally corrupted by various types of noises, which have a serious impact on the subsequent image processing tasks. In this paper, we propose a novel hyperspectral image denoising method based on tucker decomposition to model the nonlocal similarity across the spatial domain and global similarity along the spectral domain. In this method, 3-D full band patches extracted from a hyperspectral image are grouped to form a third-order tensor by utilizing the nonlocal similarity in a proper window size. In this way, the task of image denoising is transformed into a high-order tensor approximation problem, which can be solved by nonnegative tucker decomposition. Instead of a traditional alternative least square based tucker decomposition, we propose a hierarchical least square based nonnegative tucker decomposition method to reduce the computational cost and improve the denoising effect. In addition, an iterative denoising strategy is adopted to achieve better denoising outcome in practice. Experimental results on three datasets show that the proposed method outperforms several state-of-the-art methods and significantly enhances the quality of the corrupted hyperspectral image.

Low-Power LDPC-CC Decoding Architecture Based on the Integration of Memory Banks

This brief proposes a low-power low-density parity check convolutional code (LDPC-CC) decoder that is fully compatible with the IEEE 1901 standard. The proposed architecture merges multiple memory banks into one to make it consume much less power than the conventional architecture. Memory operations conducted by all the unit processors are synchronized in the proposed decoder to merge the memory and avoid any possible data hazard. The data hazard happens when a unit processor tries to read a log-likelihood ratio before a different processor updates it, degrading the error-correcting performance. Memory-access patterns appearing in a memory-based LDPC-CC decoder are formulated to determine the size of a sliding window adequate for decoding. Experimental results show that the decoding architecture employing the merged memory and the proper window size reduces the power consumption by up to 40% compared to the conventional architecture that employs multiple memory banks.

Shape measurement with modified phase-shift lateral shearing interferometry illumination and radial basis function.

Three-dimensional shapes of objects were evaluated with modified phase-shift lateral shearing interferometry illumination and radial basis function. A simple optical system was developed to create the fringe pattern based on the Murty interferometer. The phase shift was generated only by moving a plane-parallel plate along an in-plane parallel direction. A novel moving radial basis function method was presented to improve the quality of fringe patterns. And the proper calculation window size was given based on numerical simulation. Three-dimensional shapes of two kinds of objects were determined to verify the feasibility and effectiveness of the proposed method, and the reconstructed height distributions were in good accordance with the referenced data.

Localized evaluation of actuator tracking for real-time hybrid simulation using frequency-domain indices

Accurate actuator tracking plays an important role in real-time hybrid simulation (RTHS) to ensure accurate and reliable experimental results. Frequency-domain evaluation index (FEI) interprets actuator tracking into amplitude and phase errors thus providing a promising tool for quantitative assessment of real-time hybrid simulation results. Previous applications of FEI successfully evaluated actuator tracking over the entire duration of the tests. In this study, FEI with moving window technique is explored to provide post-experiment localized actuator tracking assessment. Both moving window with and without overlap are investigated through computational simulations. The challenge is discussed for Fourier Transform to satisfy both time domain and frequency resolution for selected length of moving window. The required data window length for accuracy is shown to depend on the natural frequency and structural nonlinearity as well as the ground motion input for both moving windows with and without overlap. Moving window without overlap shows better computational efficiency and has potential for future online evaluation. Moving window with overlap however requires much more computational efforts and is more suitable for post-experiment evaluation. Existing RTHS data from Network Earthquake Engineering Simulation (NEES) is utilized to further demonstrate the effectiveness of the proposed approaches. It is demonstrated that with proper window size, FEI with moving window techniques enable accurate localized evaluation of actuator tracking for real-time hybrid simulation.

Detecting defective electrical components in heterogeneous infra-red images by spatial control charts

Distribution network components connect machines and other loads to electrical sources. If resistance or current of any component is more than specified range, its temperature may exceed the operational limit which can cause major problems. Therefore, these defects should be found and eliminated according to their severity. Although infra-red cameras have been used for inspection of electrical components, maintenance prioritization of distribution cubicles is mostly based on personal perception and lack of training data prevents engineers from developing image processing methods. New research on the spatial control chart encouraged us to use statistical approaches instead of the pattern recognition for the image processing. In the present study, a new scanning pattern which can tolerate heavy autocorrelation among adjacent pixels within infra-red image was developed and for the first time combination of kernel smoothing, spatial control charts and local robust regression were used for finding defects within heterogeneous infra-red images of old distribution cubicles. This method does not need training data and this advantage is crucially important when the training data is not available. AimsDeveloping a new method to detect defective electrical components in the power distribution cubicles. Place and duration of studyTehran province, Iran, 2011–2013. MethodologyCombination of kernel smoothing, spatial control charts and local robust regression used for finding defects within heterogeneous infra-red image of old distribution cubicles. ResultsThis study showed that the IM-R control chart that plots forecasting residual of local robust regression and EWMA control chart with proper λ parameter with proper scan window size are powerful control charts which can be used to finding defected components in the power distribution cubicles. ConclusionIn some applications like analyzing thermal images of the old power distribution cubicles, it is not possible to train a sophisticated model like artificial neural network to identified defects. Therefore, spatial control chart that does not need training data is a valuable tool for these applications.

On-Body Smartphone Localization with an Accelerometer

A user of a smartphone may feel convenient, happy, safe, etc., if his/her smartphone works smartly based on his/her context or the context of the device. In this article, we deal with the position of a smartphone on the body and carrying items like bags as the context of a device. The storing position of a smartphone impacts the performance of the notification to a user, as well as the measurement of embedded sensors, which plays an important role in a device’s functionality control, accurate activity recognition and reliable environmental sensing. In this article, nine storing positions, including four types of bags, are subject to recognition using an accelerometer on a smartphone. In total, 63 features are selected as a set of features among 182 systematically-defined features, which can characterize and discriminate the motion of a smartphone terminal during walking. As a result of leave-one-subject-out cross-validation, an accuracy of 0.801 for the nine-class classification is shown, while an accuracy of 0.859 is obtained against five classes, which merges the subclasses of trouser pockets and bags. We also show the basic performance evaluation to select the proper window size and classifier. Furthermore, the analysis of the contributive features is presented.

규칙기반 및 상관분석 방법을 이용한 시계열 계측 데이터의 이상치 판정

In this study, detection methods of outlier in various monitoring data that fit into big data category were developed and outlier detections were conducted for both artificial data and real field monitoring data. Rule-based methods applied rate of change and probability of error for monitoring data are effective to detect a large-scale short faults and constant faults having no change within a certain period. There are however, problems with misjudgement that consider the normal data with a large scale variation as outlier caused by using independent single dataset. Rule-based methods for noise faults detection have a limit to application of real monitoring data due to the problem with a choice of proper window size of data and finding of threshold for outlier judgment. A correlation analysis among different two datasets were very effective to detect localized outlier and abnormal variation for short and long-term monitoring dataset if reasonable range of training data could be selected.

Robust alignment of chromatograms by statistically analyzing the shifts matrix generated by moving window fast Fourier transform cross-correlation.

Retention time shift is one of the most challenging problems during the preprocessing of massive chromatographic datasets. Here, an improved version of the moving window fast Fourier transform cross-correlation algorithm is presented to perform nonlinear and robust alignment of chromatograms by analyzing the shifts matrix generated by moving window procedure. The shifts matrix in retention time can be estimated by fast Fourier transform cross-correlation with a moving window procedure. The refined shift of each scan point can be obtained by calculating the mode of corresponding column of the shifts matrix. This version is simple, but more effective and robust than the previously published moving window fast Fourier transform cross-correlation method. It can handle nonlinear retention time shift robustly if proper window size has been selected. The window size is the only one parameter needed to adjust and optimize. The properties of the proposed method are investigated by comparison with the previous moving window fast Fourier transform cross-correlation and recursive alignment by fast Fourier transform using chromatographic datasets. The pattern recognition results of a gas chromatography mass spectrometry dataset of metabolic syndrome can be improved significantly after preprocessing by this method. Furthermore, the proposed method is available as an open source package at https://github.com/zmzhang/MWFFT2.

A Time Fairness-Based MAC Algorithm for Throughput Maximization in 802.11 Networks

This paper focuses on designing a distributed medium access control algorithm for fairly sharing network resources among contending stations in an 802.11 wireless network. Because the notion of fairness is not universal and there lacks a rigorous analysis on the relationships among the four types of most popular fairness criteria, we first mathematically prove that there exist certain connections between these types of fairness criteria. We then propose an efficient medium access algorithm that aims at achieving time fairness and throughput enhancement in a fully distributed manner. The core idea of our proposed algorithm lies in that each station needs to select an appropriate contention window size so as to fairly share the channel occupancy time and maximize the throughput under the time fairness constraint. The derivation of the proper contention window size is addressed rigorously. We evaluate the performance of our proposed algorithm through an extensive simulation study, and the evaluation results demonstrate that our proposed algorithm leads to nearly perfect time fairness, high throughput, and low collision overhead.

Nonlinear Dimensionality Reduction via the ENH-LTSA Method for Hyperspectral Image Classification

The problems of neglecting spatial features in hyperspectral imagery (HSI) and the high complexity of Local Tangent Space Alignment (LTSA) still exist in the nonlinear dimensionality reduction with LTSA for classification. Therefore, this paper proposes an innovative ENH-LTSA (Enhanced-Local Tangent Space Alignment) method to solve the two problems. First, random projection is introduced to preliminarily reduce the dimension of HSI data. It aims to improve the speed of neighbor searching and the local tangent space construction. Then, the new method presents the similarity measure via the adaptive weighted summation kernel (AWSK) distance. The AWSK distance considers both spectral and spatial features in HSI data, and attempts to ameliorate the k-nearest neighbors (KNNs) of each pixel. Furthermore, the adaptive spatial window is proposed to automatically estimate the proper window size for the description of spatial features. After that, fast approximate KNNs graph construction via Recursive Lanczos Bisection is incorporated into the new method to reduce the complexity of KNNs searching. When finishing constructing each local tangent space, the new method uses a fast low-rank approximate singular value decomposition to speed up eigenvalue decomposition of the global alignment matrix that is constituted with local manifold coordinates. Five groups of experiments with two different HSI datasets are designed to completely analyze and testify the ENH-LTSA method. Experimental results show that ENH-LTSA outperforms LTSA, both in classification results and in computational speed.

Parameter-free based two-stage method for binarizing degraded document images

Binarization plays an important role in document image processing, especially in degraded documents. For degraded document images, adaptive binarization methods often incorporate local information to determine the binarization threshold for each individual pixel in the document image. We propose a two-stage parameter-free window-based method to binarize the degraded document images. In the first stage, an incremental scheme is used to determine a proper window size beyond which no substantial increase in the local variation of pixel intensities is observed. In the second stage, based on the determined window size, a noise-suppressing scheme delivers the final binarized image by contrasting two binarized images which are produced by two adaptive thresholding schemes which incorporate the local mean gray and gradient values. Empirical results demonstrate that the proposed method is competitive when compared to the existing adaptive binarization methods and achieves better performance in precision, accuracy, and F-measure.

Active Detection With a Barrier Sensor Network Using a Scan Statistic

Abstract-The cooperation of an active acoustic source and a large number of distributed passive sensors offers an opportunity for active sonar detection. The system works as follows: first, each sensor compares its matched filter output with a given threshold to obtain a binary local decision-"0" or "1"; then, a fusion center (FC) collects them to make a system-level inference. How effectively to combine these local results-distributed detection fusion-is the concentration of this paper. Suppose that the sensor network is unaware of the target's reflection model. Then, the local detection probabilities cannot be obtained; therefore, the optimal fusion rule is unavailable. The obvious detection strategy is a counting rule test (CRT), which simply counts the total number of 1's and compares it to a threshold. This approach does not require knowledge of sensor locations, and equally considers all network subareas. However, the reflected signal from some targets, such as a submarine, can be highly aspect dependent, and in many instances only sensors in a particular zone can receive its echoes. This paper focuses on the scan statistic, which slides a window across the sensor field, and selects the subarea with the largest number of 1's to make a decision. The scan statistic integrates the aspect-dependence characteristic of the target into detection fusion. With a proper window size, it may suppress the subarea interference, and improve the system-level performance.