Complex Preprocessing Research Articles

State of health estimation of lithium-ion batteries with a temporal convolutional neural network using partial load profiles

An accurate aging forecasting and state of health estimation is essential for a safe and economically valuable usage of lithium-ion batteries. However, the non-linear aging of lithium-ion batteries is dependent on various operating and environmental conditions wherefore the degradation estimation is a complex challenge. Moreover, for on-board estimations where only limited memory and computing power are available, a state of health estimation algorithm is needed that is able to process raw sensor data without complex preprocessing. This paper presents a data-driven state of health estimation algorithm for lithium-ion batteries using different segments of partial discharge profiles. Raw sensor data is directly input to a temporal convolutional neural network without the need of executing feature engineering steps. The neural network is able to process raw sensor data and estimate the state of health of battery cells for different aging and degradation scenarios. After executing Bayesian hyperparameter tuning together with a stratified cross validation approach for splitting the training and test data, the achieved generalized aging model estimates the state of health with an overall root mean squared error of 1.0%.

Detection of Malicious Network Flows with Low Preprocessing Overhead

Machine learning (ML) is frequently used to identify malicious traffic flows on a network. However, the requirement of complex preprocessing of network data to extract features or attributes of interest before applying the ML models restricts their use to offline analysis of previously captured network traffic to identify attacks that have already occurred. This paper applies machine learning analysis for network security with low preprocessing overhead. Raw network data are converted directly into bitmap files and processed through a Two-Dimensional Convolutional Neural Network (2D-CNN) model to identify malicious traffic. The model has high accuracy in detecting various malicious traffic flows, even zero-day attacks, based on testing with three open-source network traffic datasets. The overhead of preprocessing the network data before applying the 2D-CNN model is very low, making it suitable for on-the-fly network traffic analysis for malicious traffic flows.

Identification of gas-liquid two-phase flow regime in pipelines with low liquid holdup based on ResNet1D-34

Flow pattern identification is an important topic in multiphase flow research. To overcome the subjectivity of manual identification, intelligent identification of flow patterns has attracted much attention in recent years. Both traditional machine learning methods and deep learning methods have been utilized in this field. However, traditional machine learning methods lack accuracy, and existing deep learning methods mostly rely on artificial feature extraction or complex preprocessing. In this paper, we propose a new method with high accuracy and low preprocessing dependency to solve these issues. We modify ResNet, which has proven high performance in computer vision, to fit the data collected by the wire-mesh sensor system (WMS). Due to its outstanding feature extraction ability, the new model can reach high accuracy with simple normalization as the preprocessing step. Additionally, the model can directly process data at various scales without retraining or rebuilding, which gives it high usability and economic value. The experimental results show that the accuracy of this method can reach 99.58% on our dataset.

A lightweight and accurate double-branch neural network for four-class motor imagery classification

Deep learning is an important pathway for investigation of motor imagery signal classification. Nevertheless, maintaining a good compromise between performance and computational cost has been a major challenge in developing deep models for decoding motor imagery EEG. In this paper, a novel shallow double-branch convolutional neural network (DSCNN) is proposed for four-class motor imagery classification. The proposed CNN adopts parallel extraction of two branches to improve classification accuracy. Meanwhile, in order to constrain the depth of the whole network, the left branch only contained two single temporal and spatial convolutional layers to extract common EEG features. Similarly, the right branch first introduced 1D convolution to exploit the channel dependency and temporal features across multiple time-scales, secondly a depth-wise separable convolutional layer was applied for optimizing EEG signal series. Then the feature representation for final classification was obtained by merging intermediate features extracted from the two branches. Also, the DSCNN is an end-to-end decoder, as it employs the raw EEG data as inputs and does not require additional complex preprocessing. The proposed model was evaluated on public benchmark BCI competition IV dataset 2a and achieved in terms of accuracy is 85% and kappa value is 0.79. Compared with other state-of-the-art algorithms, the experiment results reveal that the DSCNN has higher decoding accuracy and robustness, as well as a 10% improvement in accuracy than the single general shallow model. Furthermore, as a lightweight architecture, the DSCNN relies on a lower computational power than similar mainstream models, which is more in line with the requirements of low delay and real-time performance in practical BCI applications.

Tool Wear Monitoring of a Tree Log Bandsaw using a Deep Convolutional Neural Network on challenging data

In this study, we describe the development of a Deep Convolutional Neural Network (DCNN)-based system that gives warnings about soon to be necessary blade changes of log bandsaw machines. The major difficulties in monitoring the blade wear arise from the complex domain of primary wood machining and a challenging small dataset. The developed concatenated DCNN can accurately indicate necessary saw blade changes, whereas purely feature-based and Multi-Layer-Perceptron-based methods fail. The system can be applied directly without complex preprocessing. To understand the effects leading to the DCNN’s success, we also present an analysis of the prediction behavior in comparison to classical approaches.

Forest Fire Recognition Based on Feature Extraction from Multi-View Images

Forest fire recognition is important to the protection of forest resources. To effectively monitor forest fires, it is necessary to deploy multiple monitors from different angles. However, most of the traditional recognition models can only recognize single-source images. The neglection of multi-view images leads to a high false positive/negative rate. To improve the accuracy of forest fire recognition, this paper proposes a graph neural network (GNN) model based on the feature similarity of multi-view images. Specifically, the correlations (nodes) between multi-view images and library images were established to convert the input features of graph nodes into the correlation features between different images. Based on feature relationships, the image features in the library were updated to estimate the node similarity in the GNN model, improving the image recognition rate of our model. Furthermore, a fire area feature extraction method was designed based on image segmentation, aiming to simplify the complex preprocessing of images, and effectively extract the key features from images. By setting the threshold in the hue-saturation-value (HSV) color space, the fire area was extracted from the images, and the dynamic features were extracted from the continuous frames of the fire area. Experimental results show that our method recognized forest fires more effectively than the baselines, improving the recognition accuracy by 4%. In addition, the multi-source forest fire data experiment also confirms that our method could adapt to different forest fire scenes, and boast a strong generalization ability and anti-interference ability.

Combined deep CNN\u2013LSTM network-based multitasking learning architecture for noninvasive continuous blood pressure estimation using difference in ECG-PPG features

The pulse arrival time (PAT), the difference between the R-peak time of electrocardiogram (ECG) signal and the systolic peak of photoplethysmography (PPG) signal, is an indicator that enables noninvasive and continuous blood pressure estimation. However, it is difficult to accurately measure PAT from ECG and PPG signals because they have inconsistent shapes owing to patient-specific physical characteristics, pathological conditions, and movements. Accordingly, complex preprocessing is required to estimate blood pressure based on PAT. In this paper, as an alternative solution, we propose a noninvasive continuous algorithm using the difference between ECG and PPG as a new feature that can include PAT information. The proposed algorithm is a deep CNN–LSTM-based multitasking machine learning model that outputs simultaneous prediction results of systolic (SBP) and diastolic blood pressures (DBP). We used a total of 48 patients on the PhysioNet website by splitting them into 38 patients for training and 10 patients for testing. The prediction accuracies of SBP and DBP were 0.0 ± 1.6 mmHg and 0.2 ± 1.3 mmHg, respectively. Even though the proposed model was assessed with only 10 patients, this result was satisfied with three guidelines, which are the BHS, AAMI, and IEEE standards for blood pressure measurement devices.

Convolutional neural network based automatic screening tool for cardiovascular diseases using different intervals of ECG signals

Background and Objective: Automatic screening tools can be applied to detect cardiovascular diseases (CVDs), which are the leading cause of death worldwide. As an effective and non-invasive method, electrocardiogram (ECG) based approaches are widely used to identify CVDs. Hence, this paper proposes a deep convolutional neural network (CNN) to classify five CVDs using standard 12-lead ECG signals.Methods: The Physiobank (PTB) ECG database is used in this study. Firstly, ECG signals are segmented into different intervals (one-second, two-seconds and three-seconds), without any wave detection, and three datasets are obtained. Secondly, as an alternative to any complex preprocessing, durations of raw ECG signals have been considered as input with simple min-max normalization. Lastly, a ten-fold cross-validation method is employed for one-second ECG signals and also tested on other two datasets (two-seconds and three-seconds).Results: Comparing to the competing approaches, the proposed CNN acquires the highest performance, having an accuracy, sensitivity, and specificity of 99.59%, 99.04%, and 99.87%, respectively, with one-second ECG signals. The overall accuracy, sensitivity, and specificity obtained are 99.80%, 99.48%, and 99.93%, respectively, using two-seconds of signals with pre-trained proposed models. The accuracy, sensitivity, and specificity of segmented ECG tested by three-seconds signals are 99.84%, 99.52%, and 99.95%, respectively.Conclusion: The results of this study indicate that the proposed system accomplishes high performance and keeps the characterizations in brief with flexibility at the same time, which means that it has the potential for implementation in a practical, real-time medical environment.

A Dynamic Target Recognition Method Based on Correlation Tracking Algorithm

Target recognition is a very important part in the field of image processing. The effect of target recognition directly affects the success or failure of subsequent processing. This paper proposed a dynamic target recognition method based on correlation tracking algorithm. The correlation tracking algorithm only needs matching the original data of grey level of related images, no feature extraction or division is required for images. In this case, all information of this series of images can be kept, and the correlation tracking algorithm also has many advantages, no complex pre-processing of the original images is required, and the grey level of images can be used to directly matching key points, so the possibility of implementation is greatly increased. Moreover, under the condition of low noise, matching also can be carried out, and there is no high demand for the quality of sequential images. In such an approach, its local anti-interference ability is very strong. The correlation tracking algorithm can be used to predict the trajectory of a moving object in real life. The experimental results show that the proposed method can detect the objects in an image by providing the region of objects and the background with different colors. The conclusions can be drawn that the proposed method can successfully tackle the object recognition task in most cases.

Method for Fine Pattern Recognition of Space Targets Using the Entropy Weight Fuzzy-Rough Nearest Neighbor Algorithm

In space target recognition using spectral analysis technology, there is the problem that the composition or chemical properties of surface materials of the space target are similar. This problem leads to the high similarity of spectral curves and low accuracy of space target recognition. Similar object recognition is important in the study of actual space target observation. In this paper, an entropy weight fuzzy-rough nearest neighbor (EFRNN) algorithm is proposed to enhance the recognition accuracy of similar space targets, which is an improvement of the fuzzy-rough nearest neighbor algorithm. By introducing the feature weight determined using information entropy, the features of all the training samples are considered and quantified. Moreover, the proposed algorithm combined with fuzzy rough set theory can overcome the fuzzy uncertainty caused by overlapping classes and the rough uncertainty caused by insufficient features, to a certain extent. The simulation results show that the proposed algorithm achieves very promising performance compared with existing algorithms. The EFRNN classifier yields an overall classification accuracy of 95.83%. The proposed algorithm is simple and efficient for similar space target recognition. Furthermore, the EFRNN algorithm does not require preset parameters and complex preprocessing.

GPU ACCELERATION OF DOPPLER BROADENING FOR NEUTRON TRANSPORT CALCULATIONS1

In radiation transport calculations, the effects of material temperature on neutron/nucleus interactions must be taken into account through Doppler broadening adjustments to the microscopic cross section data. Historically, Monte Carlo transport simulations have accounted for this temperature dependence by interpolating among precalculated Doppler broadened cross sections at a variety of temperatures. More recently, there has been much interest in on-the-fly Doppler broadening methods, where reference data is broadened on-demand during particle transport to any temperature. Unfortunately, Doppler broadening operations are expensive on traditional central processing unit (CPU) architectures, making on-the-fly Doppler broadening unaffordable without approximations or complex data preprocessing. This work considers the use of graphics processing unit (GPU)s, which excel at parallel data processing, for on-the-fly Doppler broadening in continuous-energy Monte Carlo simulations. Two methods are considered for the broadening operations – a GPU implementation of the standard SIGMA1 algorithm and a novel vectorized algorithm that leverages the convolution properties of the broadening operation in an attempt to expose additional parallelism. Numerical results demonstrate that similar cross section lookup throughput is obtained for on-the-fly broadening on a GPU as cross section lookup throughput with precomputed data on a CPU, implying that offloading Doppler broadening operations to a GPU may enable on-the-fly temperature treatment of cross sections without a noticeable reduction in cross section processing performance in Monte Carlo transport codes.

Squeeze excitation densely connected residual convolutional networks for specific emitter identification based on measured signals

Specific emitter identification (SEI) refers to the process of distinguishing emitter individuals, which is important for electric support measure systems. Traditional SEI methods are based on hand-crafted features and have the problem of low accuracy. A novel convolutional neural network (CNN) approach, named squeeze excitation densely connected residual convolutional network (SEDCRN) approach, is therefore proposed. SEDCRN can directly recognize signals without complex preprocessing. It obtains the beneficial advantage of rediscovering new, more useful new features from low-level information and reducing feature redundancy by embedding the squeeze excitation module and residual connections into the densely connected convolutional structure. BesideIn addition, SEDCRN adopts center loss as an auxiliary loss function, which can further enhance the ability of feature learning and expression, especially for SEI tasks. The proposed approach is evaluated using a real automatic, dependent surveillance-broadcast signal dataset. Experiments show that SEDCRN outperforms traditional methods and other CNN-based methods in terms of accuracy and parameter efficiency.

Brain Age Prediction of Children Using Routine Brain MR Images via Deep Learning.

Predicting brain age of children accurately and quantitatively can give help in brain development analysis and brain disease diagnosis. Traditional methods to estimate brain age based on 3D magnetic resonance (MR), T1 weighted imaging (T1WI), and diffusion tensor imaging (DTI) need complex preprocessing and extra scanning time, decreasing clinical practice, especially in children. This research aims at proposing an end-to-end AI system based on deep learning to predict the brain age based on routine brain MR imaging. We spent over 5 years enrolling 220 stacked 2D routine clinical brain MR T1-weighted images of healthy children aged 0 to 5 years old and randomly divided those images into training data including 176 subjects and test data including 44 subjects. Data augmentation technology, which includes scaling, image rotation, translation, and gamma correction, was employed to extend the training data. A 10-layer 3D convolutional neural network (CNN) was designed for predicting the brain age of children and it achieved reliable and accurate results on test data with a mean absolute deviation (MAE) of 67.6 days, a root mean squared error (RMSE) of 96.1 days, a mean relative error (MRE) of 8.2%, a correlation coefficient (R) of 0.985, and a coefficient of determination (R2) of 0.971. Specially, the performance on predicting the age of children under 2 years old with a MAE of 28.9 days, a RMSE of 37.0 days, a MRE of 7.8%, a R of 0.983, and a R2 of 0.967 is much better than that over 2 with a MAE of 110.0 days, a RMSE of 133.5 days, a MRE of 8.2%, a R of 0.883, and a R2 of 0.780.

Detection of Moderate Traumatic Brain Injury from Resting-State Eye-Closed Electroencephalography.

Traumatic brain injury (TBI) is one of the injuries that can bring serious consequences if medical attention has been delayed. Commonly, analysis of computed tomography (CT) or magnetic resonance imaging (MRI) is required to determine the severity of a moderate TBI patient. However, due to the rising number of TBI patients these days, employing the CT scan or MRI scan to every potential patient is not only expensive, but also time consuming. Therefore, in this paper, we investigate the possibility of using electroencephalography (EEG) with computational intelligence as an alternative approach to detect the severity of moderate TBI patients. EEG procedure is much cheaper than CT or MRI. Although EEG does not have high spatial resolutions as compared with CT and MRI, it has high temporal resolutions. The analysis and prediction of moderate TBI from EEG using conventional computational intelligence approaches are tedious as they normally involve complex preprocessing, feature extraction, or feature selection of the signal. Thus, we propose an approach that uses convolutional neural network (CNN) to automatically classify healthy subjects and moderate TBI patients. The input to this computational intelligence system is the resting-state eye-closed EEG, without undergoing preprocessing and feature selection. The EEG dataset used includes 15 healthy volunteers and 15 moderate TBI patients, which is acquired at the Hospital Universiti Sains Malaysia, Kelantan, Malaysia. The performance of the proposed method has been compared with four other existing methods. With the average classification accuracy of 72.46%, the proposed method outperforms the other four methods. This result indicates that the proposed method has the potential to be used as a preliminary screening for moderate TBI, for selection of the patients for further diagnosis and treatment planning.

Improving Description-based Person Re-identification by Multi-granularity Image-text Alignments.

Description-based person re-identification (Re-id) is an important task in video surveillance that requires discriminative cross-modal representations to distinguish different people. It is difficult to directly measure the similarity between images and descriptions due to the modality heterogeneity (the crossmodal problem). And all samples belonging to a single category (the fine-grained problem) makes this task even harder than the conventional image-description matching task. In this paper, we propose a Multi-granularity Image-text Alignments (MIA) model to alleviate the cross-modal fine-grained problem for better similarity evaluation in description-based person Re-id. Specifically, three different granularities, i.e., global-global, global-local and local-local alignments are carried out hierarchically. Firstly, the global-global alignment in the Global Contrast (GC) module is for matching the global contexts of images and descriptions. Secondly, the global-local alignment employs the potential relations between local components and global contexts to highlight the distinguishable components while eliminating the uninvolved ones adaptively in the Relation-guided Global-local Alignment (RGA) module. Thirdly, as for the local-local alignment, we match visual human parts with noun phrases in the Bi-directional Fine-grained Matching (BFM) module. The whole network combining multiple granularities can be end-to-end trained without complex preprocessing. To address the difficulties in training the combination of multiple granularities, an effective step training strategy is proposed to train these granularities step-by-step. Extensive experiments and analysis have shown that our method obtains the state-of-the-art performance on the CUHK-PEDES dataset and outperforms the previous methods by a significant margin.

Specific Emitter Identification Against Unreliable Features Interference Based on Time-Series Classification Network Structure

Specific emitter identification (SEI) aims to distinguish different emitter individuals based on the subtle differences in the signals. The technology can be widely applied to various wireless communication systems. Existing individual identification methods often ignore the radio frequency (RF) fingerprint information carried by the waveforms and thus can be interfered by unreliable RF features. To alleviate these issues, we devise the deep bidirectional long short-term memory (DBi-LSTM) and the one-dimensional residual convolution network with dilated convolution and squeeze-and-excitation block (Conv-OrdsNet). We exploit the combination of DBi-LSTM and Conv-OrdsNet (CoBONet) to extract temporal structure features directly from baseband in-phase and quadrature (I/Q) samples. The proposed network is able to capture the fine-grained details of signals and combine different information extracted by two networks. Moreover, we propose a data augmentation method to solve the interference of unreliable features by randomly changing the values of noise, power, frequency offset (FO), and phase offset (PO). It is worth noting that our method only needs a short slice of a steady-state signal without complex preprocessing, which reduces the cost of acquisition and calculation. Extensive experiments show that our method can effectively extract reliable RF fingerprinting features from I/Q samples and the classification results are better than most existing methods.

Fast-Moving Sound Source Tracking With Relative Doppler Stretch

Tracking a fast-moving sound source in the air in an acoustic way has rarely been seen in the current literature. The speed of the source in the air is often not negligible compared to the speed of sound in the air. Time-varying propagation distance and noticeable Doppler effect also bring difficulties to traditional tracking methods with the time difference of arrival (TDOA) and direction of arrival (DOA). In this paper, we propose a particle filtering framework with relative Doppler stretch. We utilize the propagation delayed measurement (PDM) model and correct the posterior probability in traditional particle filter with propagation delayed state being a bridge. This method avoids the complex pre-processing of the raw acoustic signals. The simulation results show that the algorithm has an expected performance and is superior to the existing methods. The whole tracking process which starts from raw signals is shown for the first time, and the key factors affecting the pre-processing of the raw signals that are not mentioned in other articles are also discussed in this paper.

Automated quality classification of colour fundus images based on a modified residual dense block network

Fundus image-assisted diagnosis has become an intuitive and standard diagnostic technique in recent years. In hospital treatment and telemedicine, experts conduct pathological analysis and treat patients with fundus diseases using retinal images that are captured by fundus cameras. The quality of fundus images is therefore crucial for doctors to provide timely and accurate diagnoses of the diseases. However, the different experiences of fundus photographers result in different image qualities, and some images are marked as unreadable by the diagnostician, seriously delaying the treatment of patients. To solve the quality classification problem, a modified residual dense block convolution neural network (MRDB-CNN) is proposed in this paper. The two categories of images, “good quality” and “poor quality”, are used as the training data set and testing data set to train the CNNs and verify the classification results. The experimental results show that, compared with other existing network structures, the network proposed in this paper can classify the “good-quality” and “poor-quality” fundus images more accurately with an accuracy rate of 99.90%. In addition, the tests for the “ambiguous-quality” category further demonstrate that our method can obtain more detailed features of the fundus images and give an objective score to the image quality. In short, the MRDB-CNN can accurately classify the quality of fundus images while avoiding the complex preprocessing of traditional algorithms, and it satisfies the needs of hospital treatment and telemedicine for the real-time quality judgement of fundus images.

Making Sense of Spatio-Temporal Preserving Representations for EEG-Based Human Intention Recognition.

Brain-computer interface (BCI) is a system empowering humans to communicate with or control the outside world with exclusively brain intentions. Electroencephalography (EEG)-based BCI is one of the promising solutions due to its convenient and portable instruments. Despite the extensive research of EEG in recent years, it is still challenging to interpret EEG signals effectively due to its nature of noise and difficulties in capturing the inconspicuous relations between EEG signals and specific brain activities. Most existing works either only consider EEG as chain-like sequences while neglecting complex dependencies between adjacent signals or requiring complex preprocessing. In this paper, we introduce two deep learning-based frameworks with novel spatio-temporal preserving representations of raw EEG streams to precisely identify human intentions. The two frameworks consist of both convolutional and recurrent neural networks effectively exploring the preserved spatial and temporal information in either a cascade or a parallel manner. Extensive experiments on a large scale movement intention EEG dataset (108 subjects, 3 145 160 EEG records) have demonstrated that the proposed frameworks achieve high accuracy of 98.3% and outperform a set of state-of-the-art and baseline models. The developed models are further evaluated with a real-world brain typing BCI and achieve a recognition accuracy of 93% over five instruction intentions suggesting good generalization over different kinds of intentions and BCI systems.

Photoelasticity-based evaluation of cellular contractile force for phenotypic discrimination of vascular smooth muscle cells

Vascular smooth muscle cells (VSMCs) have two distinct phenotypes: contractile and synthetic. The major difference between these phenotypes lies in the magnitude of the contractile force produced by the cell. Although traction force microscopy (TFM) is often used to evaluate cellular contractile force, this method requires complex preprocessing and a sufficiently compliant substrate. To evaluate the contractile force and the phenotype of living VSMCs with minimal effort and in a manner independent of the substrate stiffness, we propose a photoelasticity-based method using retardation, which is related to the difference between the first and second principal stresses and their orientation. The results demonstrate that actin filaments co-localize with areas of high retardation in cells, indicating that the retardation of VSMCs is promoted by actin filaments. The retardation of cells treated with calyculin A and Y-27632 tended to be larger and smaller, respectively, than that of control cells. Cell traction force significantly correlates with total cell retardation (r2 = 0.38). The retardation of contractile VSMCs (passage 2) was significantly higher than that of synthetic VSMCs (passage 12). These results indicate that cell retardation can be used to assess cell contractile force and, thus, determine the phenotype of VSMCs.