Two-stage Processing Research Articles

A new graph-based clustering method with dual-feature regularization and Laplacian rank constraint

The performance of graph-based clustering is commonly limited by two-stage processing (Constructing and dividing similarity graph) and the quality of similar graphs. To this end, we propose a new graph-based clustering method with dual-feature regularization and Laplacian rank constraint. Specifically, our method reveals the clustering structure and unifies the two-stage process. It imposes a Laplacian rank constraint on the similarity graph to ensure that it has C connected components. In addition, a method based on dual-feature regularization is designed to capture local data feature information from both feature extraction and adaptive regression, and is applied to an accurate distance metric learning. A reweighting optimization is integrated to learn a high-quality robust similarity graph. Comprehensive experiments on Ecoli, Yale and Yeast datasets show that our method outperforms the existing graph-based clustering methods with an average improvement of about 4%, 5% and 7% on the evaluation metrics ACC, NMI and RI, respectively.

The two-stage processing of judgment of confidence: evidence from ERP

BackgroundThe judgment of confidence (JOC) refers to the confidence in the accuracy of the target item individuals have just retrieved and is a typical retrospective metacognitive monitoring process. In the classical paradigm of JOC, JOC occurs after the recognition or recall task. While initially viewed as a single-stage monitoring process, recent research on JOC suggests its internal mechanisms may be more complex, potentially encompassing both retrieval and monitoring processes. This study aims to delve into these mechanisms concerning neural temporal processes.MethodsIn this study, event-related potential (ERP) was used to compare N400 and slow-wave ERPs of high and low JOCs at different time windows using a classic JOC paradigm.ResultsBehavioral results showed an inverted-U shaped relationship between response time (RT) and JOCs, peaking at magnitude 3 before declining. There were significantly longer RT for low JOCs compared with high JOCs, along with lower recognition scores. The ERP results showed that low JOCs induced larger N400 in the right frontal lobe and right central area, while high JOCs induced larger slow-wave components (500 ~ 700ms) in the right frontal lobe.ConclusionsBased on these findings, the present study suggests that JOC involves two processing stages. N400 reflects the process of cue acquisition, while the slow-wave component reflects the process of cue application. Furthermore, a two-stage model was proposed and validated, enriching the study of metacognition monitoring mechanisms, offering insights into the processing mechanisms of retrospective metacognitive monitoring.

A Set of Models for Device Positioning in Sixth Generation Networks. Part 2. Review of Algorithms and Accuracy Assessment

Relevance. This work is the second part of a series devoted to the study of a set of positioning models in sixth-generation terahertz networks and solves the problems of systematizing algorithms and assessing the accuracy of determining the location of a user device depending on the configuration and size of the antenna array at the base station.Purpose. Within the framework of the scientific problem of searching for means of achieving decimeter accuracy of coordinate estimates, outlined in the first part of the cycle, the analysis of accuracy assessment models, a review of algorithms and ways of their optimization, as well as a numerical experiment, performed in this study serve the purpose of justifying the configuration and dimensions of the antenna array used at the base station.The research method is an analytical review of the state of the problem based on current scientific publications, conceptual modeling, categorical approach, expert combination, comparative analysis, formalization, mathematical and simulation modeling.Solution / results. The paper presents models for assessing the accuracy of positioning in 6G terahertz networks, formalizes the relationship between primary measurements and coordinate estimates for multi-position and single-position positioning in the near and far zones. It provides an overview of algorithms for geometric positioning and positioning with training for cases of one-stage and two-stage processing; analyzes the specifics of implementing algorithms for simultaneous tracking and map construction. It provides an analysis of the features of optimizing algorithms in offline and online modes. Simulation modeling is used to assess the accuracy for a scenario of territorial distribution with direct visibility and ideal synchronization.Novelty. Using simulation modeling tools, the achievement of decimeter accuracy of coordinate and orientation estimates of 1° in the terahertz range for a far-field model using a 1 GHz band and a composite antenna array of more than half a thousand elements has been scientifically substantiated.The theoretical significance lies in establishing the dependence of the accuracy of coordinate and orientation estimates of the device on the configuration and dimensions of the antenna array at the base station.The practical significance of the developed simulation model lies in the numerical justification of the limits of device positioning accuracy in sixth-generation networks depending on the antenna array used at the base station for a given scenario.

Eye-Tracking Evidence for the SNARC Effect: Unveiling Attentional Stages

The Spatial Numerical Association of Response Codes (SNARC) effect, which involves the conjunction of spatial and numerical response coding, was initially identified through reaction time experiments involving numerical stimuli. Research has revealed a common tendency among individuals to associate smaller numbers with the left side and larger numbers with the right side. Furthermore, scholars have identified additional dimensions, such as the “up-down” and “near-far” associations, that contribute to the complexity of the SNARC effect. However, the universality of the SNARC effect and the underlying psychological and physiological mechanisms continue to be a topic of debate within the academic community. Various studies have proposed a range of interpretations and understandings of this effect. The present study delves into the cognitive mechanisms of the SNARC effect, employing eye-tracking technology to objectively quantify participants’ responses. Utilizing the parity judgment task, a method frequently employed in SNARC effect research, in conjunction with eye-tracking, participants were instructed to judge the parity of numerical stimuli while directing their gaze to the left or right, and above or below the numbers as per the experimental requirements. Colored blocks were positioned around the numbers to serve as visual anchors, indicating the spatial directions of left/right and top/down. Data analysis revealed that when smaller numbers (1-4) were presented, participants exhibited a tendency to focus on the left color block adjacent to the number, characterized by a higher Total Fixation Duration (TFD) and a shorter Time to First Fixation (TFF). In contrast, when larger numbers (6-9) appeared, participants were more likely to focus on the right color block, indicating the presence of a SNARC effect in the horizontal dimension. However, no significant effect was observed in the vertical dimension. Additionally, this study innovatively employed a blank page in conjunction with a test page to differentiate between the early and late stages of participants’ attentional processing. The examination of TFF and TFD data uncovered a two-stage processing mechanism associated with the SNARC effect. This experimental approach and methodological framework offer novel perspectives that contribute to the advancement of research in this domain.

Improving mechanical and electrical properties of Cu-Ni-Si alloy via machine learning assisted optimization of two-stage aging processing

Recent studies have shown that synergistic precipitation of continuous precipitates (CPs) and discontinuous precipitates (DPs) is a promising method to simultaneously improve the strength and electrical conductivity of Cu-Ni-Si alloy. However, the complex relationship between precipitates and two-stage aging process presents a significant challenge for the optimization of process parameters. In this study, machine learning models were established based on orthogonal experiment to mine the relationship between two-stage aging parameters and properties of Cu-5.3Ni-1.3Si-0.12Nb alloy with preferred formation of DPs. Two-stage aging parameters of 400 °C/75 min + 400 °C/30 min were then obtained by multi-objective optimization combined with an experimental iteration strategy, resulting in a tensile strength of 875 MPa and a conductivity of 41.43 %IACS, respectively. Such an excellent comprehensive performance of the alloy is attributed to the combined precipitation of DPs and CPs (with a total volume fraction of 5.4% and a volume ratio of CPs to DPs of 6.7). This study could provide a new approach and insight for improving the comprehensive properties of the Cu-Ni-Si alloys.

Angular attitude estimation algorithm using signals of Global Navigation Systems

Navigation apparatus of Global Positioning Systems are wildly used in many applications one of which is attitude determination using global satellite navigation signals, one of which is angular attitude determination. One of effective approach to this task is based on representation an object motion as linear motion along line-of-site and rotation about object center of mass. From this consideration it followed deferent tracking systems on signal parameters that connected with linear motion along line-of-site and on signal parameters that connected with rotation about object center of mass. In the article this approach is developed for synthesis of angular attitude estimation using signals of Global Navigation Systems for two-stage processing in navigation apparatus in coherent mode. Objective. To get more effective algorithm of angular attitude estimation using signals of Global Navigation Systems.

The effect of Ar+ and N+ ion irradiation on the thermally induced evolution of the structural and magnetic properties of Co/Pt and Pt/Co bilayered stacks

The application of Co–Pt thin films as functional elements of novel nanoelectronics and spintronics devices requires the formation of a homogeneous ferromagnetic CoPt phase with tunable magnetic properties. A diffusion-controlled synthesis of this ferromagnetic phase can be implemented through the annealing of deposited Co/Pt bilayers. Apart from thermal treatment, both structural and magnetic properties of such layered stacks can be affected by ion preirradiation. In this work, we, therefore, studied the effect of a two-stage process consisting of preirradiation with 110 keV Ar+/N+ ions followed by post-annealing in vacuum at 550°С for 30 min on the evolution of the structural, chemical, and magnetic properties of Co/Pt/substrate and Pt/Co/substrate heterostructures. The results obtained for such two-stage processing were compared to those received after single-stage vacuum annealing. It was found that when ion preirradiation is followed by annealing, the diffusion-driven intermixing of Pt and Co leading to the formation of the ferromagnetic Co–Pt phase is slowed down compared to the non-irradiated samples, which is associated with the barrier effect of implanted projectiles. Furthermore, we demonstrate that preirradiation does not compromise the magnetic properties of the samples. For instance, preirradiation leads to a coercivity increase of up to 38 % compared to the non-irradiated annealed samples which is attributed to the presence of remaining paramagnetic Pt between the grains of the ferromagnetic A1-CoPt phase. We demonstrate that the applied two-stage processing (consisting of ion preirradiation followed by thermal annealing) of magnetic thin films is a promising approach for tailoring their magnetic properties such as the in-plane coercivity, saturation, and effective magnetization.

The effect of two-stage filler processing on the properties of a wood-polymer composite

The effect of two-stage filler processing on the properties of a wood-polymer composite

Analysis of peculiarities of polymer matrix microstructure in PCMs formed under the influence of electrophysical effects

The microstructure of monolayer samples formed by 3D printing from composite prepregs reinforced with continuous carbon fiber, followed by treatment in a cured state by exposure to ultrasonic vibrations and a microwave electromagnetic field, has been studied. It has been shown that high-power ultrasonic exposure causes macro-changes, which are characterized by a decrease in the number and transverse dimensions of delaminations (cracks) from 0.05–0.1 mm to 0.01–0.02 mm and a decrease in the height of microrelief elements by almost an order of magnitude. Changes caused by exposure to a microwave electromagnetic field are more apparent when examining the end surface with an increase from ×1000 to ×10000. The control samples revealed fibers consolidated by binder in combination with unfilled gaps in a ratio of about 50:50, with individual globular fragments of frozen binder on their surface. The samples after microwave exposure revealed the presence of large areas without defects, completely uniformly filled with binder, fibers from all sides are in contact with the binder, in which there are voids no larger than 5 × 1.5μm in size. Greater orderliness of the structure of monolayers is most typical for samples with two-stage processing, including preliminary ultrasonic and subsequent.

Spall size estimation for wind turbine pitch bearings: observation, signal processing method and experiments

It is essential to continuously monitor the spall size of wind turbine pitch bearings to prevent severe faults and catastrophic failure. In the field of spall size estimation for bearings, an essential step is to extract the entry and impact signals simultaneously. And this would become more difficult when it comes to the wind turbine pitch bearings due to the limited fault signals and heavy noise. In this paper, a coherent procedure is proposed to estimate the spall size for wind turbine pitch bearings. Firstly, the characteristics of entry and impact signals in actual wind turbine pitch bearings are observed to be low-frequency and high-frequency dominated, respectively. On the basis of characteristics analysis, a novel two-stage signal processing method called the wavelet augmented sparse dictionary, is proposed to extract the entry and impact signals, which combines the discrete wavelet transform and sparse representation technique. Finally, the spall size is calculated according to aforementioned extraction and geometric constraints. Results from real-world experiments demonstrate the effectiveness of the proposed method.

Toward Better Pedestrian Trajectory Predictions: The Role of Density and Time-to-Collision in Hybrid Deep-Learning Algorithms.

Predicting human trajectories poses a significant challenge due to the complex interplay of pedestrian behavior, which is influenced by environmental layout and interpersonal dynamics. This complexity is further compounded by variations in scene density. To address this, we introduce a novel dataset from the Festival of Lights in Lyon 2022, characterized by a wide range of densities (0.2-2.2 ped/m2). Our analysis demonstrates that density-based classification of data can significantly enhance the accuracy of predictive algorithms. We propose an innovative two-stage processing approach, surpassing current state-of-the-art methods in performance. Additionally, we utilize a collision-based error metric to better account for collisions in trajectory predictions. Our findings indicate that the effectiveness of this error metric is density-dependent, offering prediction insights. This study not only advances our understanding of human trajectory prediction in dense environments, but also presents a methodological framework for integrating density considerations into predictive modeling, thereby improving algorithmic performance and collision avoidance.

Microstructural characteristics and recrystallization mechanism of Ti-6.5Al–2Zr–1Mo–1V alloy during two-stage hot deformation

In this study, we investigated the two-stage deformation behavior of near α Ti-6.5Al–2Zr–1Mo–1V alloy at various temperatures (915 °C, 950 °C, 1020 °C) and strain rates (0.01 s−1, 0.1 s−1). The alloy underwent differing holding periods (500s, 1000s) between deformation stages. The results demonstrated an increase in flow stress with decreasing deformation temperature and increasing strain rate. Work hardening of flow stress was observed during the two-stage thermomechanical processing within the α+β two-phase region. Microstructural evolution was characterized by the formation of α lamellar kink bands, the separation and migration of grain boundaries, and grain spheroidization. An increase in deformation temperature and a decrease in strain rate enhanced grain refinement, accelerated structural evolution, and intensified the degree of spheroidization. Dynamic recrystallization (DRX) was more prevalent at higher temperatures and lower strain rates, as the critical nucleation rate for DRX was more readily achieved. The increased interval between thermomechanical processing stages led to a transition from low-angle to high-angle grain boundaries, and a reduction in dislocation density due to static recovery occurred during the holding period. At 915 °C, hot deformation induced the formation of {10-10} twinning, altering grain orientation and reducing stress. The Schmid factor for {10-10}<1120> slip was higher at lower temperatures, which reduced the critical resolved shear stress necessary for yielding in the α phase of the Ti-6.5Al–2Zr–1Mo–1V alloy, thereby facilitating further plastic deformation.

Purification of Copper Concentrate from Arsenic under Autoclave Conditions

This study presents the results of a two-stage autoclave processing of a copper–arsenic concentrate. Copper concentrate is an important raw material to produce copper and other metals. However, in some cases, the concentrate may contain increased amounts of arsenic, which makes further processing difficult. Therefore, the development of modern hydrometallurgical methods for processing copper concentrate with a high arsenic content is an urgent task, which could lead to the optimization of the raw material processing process and the improvement of the quality of the concentrate. It has been established that the optimal conditions for the sequential two-stage autoclave processing of copper–arsenic concentrate are: t = 220–225 °C, τoxidation = 20 min, τtot = 90 min, Po2 = 0.4 MPa, and L:S = 10:1, [H2SO4]initial = 40 g/dm3; in this case, 85% of zinc, 44% of iron, and 78% of arsenic, respectively, are extracted into the solution during both stages and the loss of copper was about 0.01%. This is explained by the fact that at the first stage (oxidation) of the autoclave processing of the copper–arsenic concentrate, copper, together with iron, leaches into the solution, and at the second stage (reduction), copper precipitates out of the solution in the form of chalcocite. Copper in the residue after autoclave leaching is in the form of Cu2S, iron is in the form of pyrite (FeS2), and lead is in the form of anglesite (PbSO4), respectively. The obtained micrographs and EDX mappings clearly show no iron arsenates. This confirms that at the oxidative stage of the developed process, arsenic, removed by 78%, remains in the solution. The remaining arsenic is associated with tennantite, indicating the effectiveness of the treatment process in removing arsenic from the copper–arsenic concentrate. A second important observation is the presence of pronounced areas of copper sulfides in the microphotos without iron and arsenic impurities. This confirms that copper is deposited as chalcocite during the reduction phase of the process, which is the desired result.

An Image Processing Approach for Real-Time Safety Assessment of Autonomous Drone Delivery

The aim of producing self-driving drones has driven many researchers to automate various drone driving functions, such as take-off, navigation, and landing. However, despite the emergence of delivery as one of the most important uses of autonomous drones, there is still no automatic way to verify the safety of the delivery stage. One of the primary steps in the delivery operation is to ensure that the dropping zone is a safe area on arrival and during the dropping process. This paper proposes an image-processing-based classification approach for the delivery drone dropping process at a predefined destination. It employs live streaming via a single onboard camera and Global Positioning System (GPS) information. A two-stage processing procedure is proposed based on image segmentation and classification. Relevant parameters such as camera parameters, light parameters, dropping zone dimensions, and drone height from the ground are taken into account in the classification. The experimental results indicate that the proposed approach provides a fast method with reliable accuracy based on low-order calculations.

Noncoherent Channel Combining for GNSS Signal Tracking with an Adaptive Antenna Array

Adaptive antenna arrays are widely used to protect Global Navigation Satellite System (GNSS) receivers from interference. However, conventional blind anti-jamming array processing methods can cause satellite signal attenuation, which reduces signal tracking precision and measurement accuracy. To alleviate this problem, this paper proposes the noncoherent channel combining method for GNSS signal tracking with an adaptive antenna array. A two-stage processing architecture is applied in the proposed method, where the first stage implements interference suppression, and the second stage achieves noncoherent channel combining. The proposed method has the advantage of blind characteristics that can be implemented without any other auxiliary information. It also reveals a unified processing architecture for anti-jamming array processing methods. Thus, a detailed comparison and analysis of the proposed method with several typical blind anti-jamming methods is performed. Simulation results demonstrate the effectiveness of the proposed method and its superiority over existing methods in terms of tracking performance.

System for analyzing data from camera traps based on a two-stage neural network approach for operational remote monitoring

Camera traps are a widely used tool for monitoring wildlife with minimal human intervention. Their number can amount to several hundred, and the accumulated volume can reach several terabytes. Often, photos and videos contain empty frames that are created by accidental triggering of camera trap detectors, such as by wind. The staff of nature reserves must process the images manually and sort them by animal species. In our study we propose to consider a technology for analysing data from camera traps using a two-stage neural network processing. The task of the first stage was to separate empty images from non-empty images. To do this, using a comparative analysis, we identified the most optimal detector model from the YOLO series. The task of the second stage was to classify the objects found by the detector. For this purpose, a comparative analysis of the architectures of classifiers from the ResNet series was carried out. Based on the selected algorithms, a two-stage system for processing data from camera traps was created in the form of a graphical interface with the ability to work on any operating system. The software will significantly reduce the processing time of data from camera traps and simplify environmental analysis.

INVESTIGATION OF THE YOLOv5 ALGORITHM EFFICIENCY FOR DRONE RECOGNIZATION

With the growth in the production and sale of drones, the number of offenses related to the use of drones in no-fly zones is increasing. Visual detection systems using neural networks for drone recognition effectively solve this problem. One of the most effective algorithms for objects of various classes is YOLO, which can be used to detect drones. However, this algorithm has a number of limitations that reduce the drone detection range. Therefore, the goal of the research is to evaluate the detecting drones' effectiveness at different distances using the fifth version of this algorithm, as well as to create algorithms for increasing the detection range. Based on the experimental data, datasets were created for training four main modifications: s, m, l, and x of the neural network of the YOLOv5 algorithm. These network modifications were trained for the visible and infrared (IR) ranges, as well as various image resolutions at the network input. After processing a dataset of drone flight videos, the effectiveness of various modifications of the neural network was evaluated, the dependences of the probability of detecting a drone on the distance and speed of the drone were plotted, and the maximum detection range was estimated. A two-stage algorithm has been created that makes it possible to increase the detection probability and increase the detection range as a result of a combination of the classical YOLOv5 algorithm at the first stage of processing and the convolutional neural network proposed by the authors at the second stage. An algorithm for synthesizing IR images has been created to supplement IR datasets with the necessary drone-background combinations when training neural networks. Practical recommendations are given for choosing the type of neural network and quantitative estimates of the YOLOv5 algorithm's efficiency in combination with a two-stage processing algorithm.

A novel hybrid model based on two-stage data processing and machine learning for forecasting chlorophyll-a concentration in reservoirs.

The accurate and efficient prediction of chlorophyll-a (Chl-a) concentration is crucial for the early detection of algal blooms in reservoirs. Nevertheless, predicting Chl-a concentration in multivariate time series poses a significant challenge due to the complex interrelationships within the aquatic environment and the discrete and non-stationary nature of online monitoring of water quality data. To address the aforementioned issue, this paper proposes a novel prediction model named SGMD-KPCA-BiLSTM (SKB) for predicting Chl-a concentration. The model combines two-stage data processing and machine learning (ML). To capture nonlinear relationships in multivariate time series data, the optimal data subset is determined by combining symplectic geometry mode decomposition (SGMD) and kernel principal component analysis (KPCA). This subset is then input into a bidirectional long short-term memory (BiLSTM) model, and the model's hyperparameters are optimized using the sparrow search algorithm (SSA) to improve the accuracy of predictions. The performance of the model was evaluated at Qiaodian Reservoir in Shandong, China. To assess its superiority, the evaluation criteria included the root mean square error (RMSE), mean absolute percentage error (MAPE), mean absolute error (MAE), coefficient of determination (R2), frequency histograms of the prediction error, and the Taylor diagram. The prediction performance of five single models, namely the back-propagation (BP) neural network, support vector regression (SVR), long short-term memory (LSTM), convolutional neural network with long short-term memory (CNN-LSTM), and BiLSTM, as well as three hybrid models, namely SGMD-LSTM, SGMD-KPCA-LSTM, and SGMD-BiLSTM, were compared against the SKB model. The results demonstrated that the SKB model performs best in predicting Chl-a concentration (R2 = 96.19%, RMSE = 1.05, MAE = 0.65, MAPE = 0.08). It significantly reduced the prediction error compared to other models for comparison. Furthermore, the multi-step predictive capabilities of the SKB model are also discussed. The analysis shows a decline in predictive performance with larger prediction time steps, and the SKB model exhibits slightly superior performance compared to the other model at corresponding prediction intervals. The model has significant advantages in terms of its ability to accurately predict the non-smooth and nonlinear Chl-a sequences observed by the online monitoring system. This study presents a potential solution for controlling and preventing reservoir eutrophication, as well as an innovative approach for predicting water quality.

Natural language processing system for rapid detection and intervention of mental health crisis chat messages

Patients experiencing mental health crises often seek help through messaging-based platforms, but may face long wait times due to limited message triage capacity. Here we build and deploy a machine-learning-enabled system to improve response times to crisis messages in a large, national telehealth provider network. We train a two-stage natural language processing (NLP) system with key word filtering followed by logistic regression on 721 electronic medical record chat messages, of which 32% are potential crises (suicidal/homicidal ideation, domestic violence, or non-suicidal self-injury). Model performance is evaluated on a retrospective test set (4/1/21–4/1/22, N = 481) and a prospective test set (10/1/22–10/31/22, N = 102,471). In the retrospective test set, the model has an AUC of 0.82 (95% CI: 0.78–0.86), sensitivity of 0.99 (95% CI: 0.96–1.00), and PPV of 0.35 (95% CI: 0.309–0.4). In the prospective test set, the model has an AUC of 0.98 (95% CI: 0.966–0.984), sensitivity of 0.98 (95% CI: 0.96–0.99), and PPV of 0.66 (95% CI: 0.626–0.692). The daily median time from message receipt to crisis specialist triage ranges from 8 to 13 min, compared to 9 h before the deployment of the system. We demonstrate that a NLP-based machine learning model can reliably identify potential crisis chat messages in a telehealth setting. Our system integrates into existing clinical workflows, suggesting that with appropriate training, humans can successfully leverage ML systems to facilitate triage of crisis messages.

A wind speed forecasting system for the construction of a smart grid with two-stage data processing based on improved ELM and deep learning strategies

The operation and scheduling management of smart grids are important aspects, and wind speed forecasting modules are indispensable in wind power system management. Researchers have contributed significantly to the development of accurate forecasting models. However, predicting the ideal performance remains a daunting task. Data preprocessing strategies are widely used to process the original wind speed sequences. To develop the utility of the data preprocessing module, a novel forecasting framework based on a two-stage data processing method is designed in this study. The designed system combines singular spectrum analysis and variational mode decomposition methods to decompose the trend term and multiple components of the residual term of the sequence to effectively capture its inherent characteristics. In addition, a multi-objective optimization strategy was applied to determine the weights of the prediction sequences obtained using deep learning techniques and improved extreme learning machine algorithms to obtain accurate forecasting results. The experimental results verified that the proposed wind forecasting framework is better than other benchmark comparison models, thereby establishing a feasible solution for wind speed forecasting and a powerful tool for power grid operation management.