Adaptive Threshold Method Research Articles

Enhancement of Small Ship Detection Using Polarimetric Combination from Sentinel−1 Imagery

Speckle noise and the spatial resolution of the Sentinel−1 Synthetic Aperture Radar (SAR) image can cause significant difficulties in the detection of small objects, such as small ships. Therefore, in this study, the Polarimetric Combination-based Ship Detection (PCSD) approach is proposed for enhancing small ship detection performance, which combines three different characteristics of polarization: newVH, enhanced VH, and enhanced VV. Employing the Radar Cross Section (RCS) value in three stages, the newVH was utilized to detect Automatic Identification System (AIS) -ships and small ships. In the first step, the adaptive threshold (AT) method was applied to newVH with a high RCS condition (>−10.36 (dB)) for detecting AIS-ships. Secondly, the first small ship target was detected with the maximum suppression of false alarms by using the AT with a middle RCS condition (>−16.98 (dB)). In the third step, a candidate group was identified by applying a condition to the RCS values (>−23.01 (dB)), where both small ships and speckle noise were present simultaneously. Subsequently, the enhanced VH and VV polarizations were employed, and an optimized threshold value was selected for each polarization to detect the second small ship while eliminating noise pixels. Finally, the results were evaluated using the AIS and small fishing vessel tracking system (V-Pass) based on the detected ship positions and ship lengths. The average matching results from 26 scenes in 2022 indicated a matching rate of over 86.67% for AIS-ships. Regarding small ships, the detection performance of PCSD was 42.27%, which was over twice as accurate as the previous Constant False Alarm Rate (CFAR) ship detection model. As a result, PCSD enhanced the detection rate of small ships while maintaining the capacity for detecting AIS-equipped ships.

Series–Parallel System Analysis and Reliability Research of Grid Structures Considering Adaptive Dynamic Bounding Threshold

There are many bars in grid structures, and their topological relationships are complex, so the workload of calculations involved in researching their reliability is large. In this paper, based on the geometric topological relationship within a grid structure, a series–parallel system analysis of the structure is carried out by means of node analysis, constraint grade division, and hierarchical combination analysis, and the series–parallel analysis results for the structure are used to judge the failure of the structure. In order to avoid the problem of missing important components due to unreasonable selection of the bounding threshold when identifying failed components, an adaptive dynamic bounding threshold method with high accuracy is introduced. The failure probability value of the failure path is calculated using the direct numerical integration method, and the reliability of the whole structure is obtained. Finally, the validity of this method is verified using an example of a square pyramid grid structure. The results show that the method is feasible for series–parallel system analysis, structural failure judgment, and reliability analysis of a grid structure.

Method of comparing and transforming images obtained using UAV

The subject matter of this article involves reviewing and developing methods for the comparison and transformation of images obtained using UAV via Computer Vision tools. The goal is to improve methods for image comparison and transformation. Various image-processing methods were employed to achieve the goal of this study,thereby contributing to the development of practical algorithms and approaches for image analysis and comparison. The tasks can be described as follows: 1) development of image comparison methods: design tools for the comparison of images from UAV that efficiently detect differences using algorithms such as cv2.absdiff and the PIL module; 2) Image transformation: implement transformation methods for images from UAV, including perspective transformation and thresholding, to enhance the quality and accuracy of image analysis. The methods used were algorithm development, image transformation methods, statistical analysis, experimental testing, and performance evaluation. The metrics used in this article are response time and accuracy. Algorithms for image comparison have also been refined, particularly those transformed through Global Threshold Value, Adaptive Mean Thresholding, and Adaptive Gaussian Thresholding. A novel change filtering method was introduced to enhance the precision of image comparison by filtering out insignificant alterations following image transformation. Comprehensive investigation of image comparison involving edge detection methods has been systematically presented. The results contain the development of practical algorithms and approaches for image analysis and comparison applicable in diverse areas such as military, security, and agriculture. Possibilities of applying our methods and algorithms in the context of drones were also considered, which is particularly relevant in tasks related to computer vision in unmanned aerial vehicles, where limited resources and the need for real-time processing of a large volume of data create unique challenges. Conclusions. The results contain OpenCV and PIL image comparison methods. OpenCV pixel-by-pixel comparison algorithm showed a better response time with the same accuracy. OpenCV method has 92,46% response time improvement compared with PIL and is 276ms. As for image thresholding with comparison, a method based on Global Threshold Value showed the shortest response time (266ms) and the lowest accuracy. The highest accuracy and response time (366ms) were obtained using the Adaptive Gaussian Thresholding method.

Keeping Pathologists in the Loop and an Adaptive F1-Score Threshold Method for Mitosis Detection in Canine Perivascular Wall Tumours

Performing a mitosis count (MC) is the diagnostic task of histologically grading canine Soft Tissue Sarcoma (cSTS). However, mitosis count is subject to inter- and intra-observer variability. Deep learning models can offer a standardisation in the process of MC used to histologically grade canine Soft Tissue Sarcomas. Subsequently, the focus of this study was mitosis detection in canine Perivascular Wall Tumours (cPWTs). Generating mitosis annotations is a long and arduous process open to inter-observer variability. Therefore, by keeping pathologists in the loop, a two-step annotation process was performed where a pre-trained Faster R-CNN model was trained on initial annotations provided by veterinary pathologists. The pathologists reviewed the output false positive mitosis candidates and determined whether these were overlooked candidates, thus updating the dataset. Faster R-CNN was then trained on this updated dataset. An optimal decision threshold was applied to maximise the F1-score predetermined using the validation set and produced our best F1-score of 0.75, which is competitive with the state of the art in the canine mitosis domain.

Full-flow surface defect identification method based on spot scanning scattering for unpatterned wafer

As the critical dimensions of semiconductor manufacturing processes gradually decrease, the requirements for production yield management become increasingly stringent. During the manufacturing process, there are many different types of defects, such as micron-sized particles, millimeter-sized scratches, etc. Multiple categories and different scales bring great challenges to the detection and identification of defects. This paper provides a full-flow surface defect identification method based on spot scanning scattering for unpatterned wafers. First, an adaptive threshold method with dynamic kernel windows is used to perform line-by-line scanning inspection of the wafer Mercator image. The 3σ decision strategy is used to avoid the impact of defects on background estimation and to improve detection sensitivity. After morphological processing, connected domain analysis is performed to obtain the defect mask, and feature information such as the shape, size, and distribution of the defect is extracted. Finally, the defect identification is performed by rules based binning, and the identified defects are converted into wafer polar coordinate image for display and analysis. In the experiments, the proposed method is used to identify micron-scale particles as well as large scratches on the millimeter scale for SiC wafers. Relative to the actual production rate requirement of 20 wafers per hour, the analysis time for a 6-inch wafer is 24.4 s, which can meet the requirement. Meanwhile, the test results illustrate the effectiveness of the method. The proposed method is recommended for early-stage defect detection and identification of unpatterned wafers.

Image preprocessing techniques applied on NIR images for fruit bruise detection

This study investigates the transformative potential of image preprocessing techniques when applied to near-infrared (NIR) images for early bruise detection. It emphasizes the nuanced selection of filters to retain essential image features while accentuating bruise characteristics. Filters as noise-reduction tools, rendering bruises more visible without erasing critical details. Subsequently, the limitations of conventional edge detection filters were examined such as Sobel, Prewitt, and Canny, which excel in outlining fruit edges but fall short in delineating bruises. Adaptive thresholding methods were introduced, exemplified by Otsu’s, showcasing their capacity to distinguish objects from backgrounds while acknowledging their challenge in preserving crucial edge pixels. Image enhancement techniques, such as Histogram Equalization, Contrast Stretching, and Sigmoid Correction, enhance fruit edge visibility and elevate bruise detection. In the frequency domain, filters such as Ideal Lowpass, Bandpass, and Highpass were harnessed to accentuate diverse bruise types. The Butterworth filter was introduced, capable of concurrently highlighting all relevant features, a pivotal innovation in comprehensive bruise detection. Through extensive experimentation and analysis of NIR images of various fruit varieties, including plums, peaches, and apples, our findings underscore the significance of tailored preprocessing techniques for optimal fruit bruise detection. These insights offer promise for agricultural industries and quality control processes seeking to enhance fruit quality assessment.

An Adaptive Thresholding Method With Local Binary Patterns Features for Image Reconstruction

An Adaptive Thresholding Method With Local Binary Patterns Features for Image Reconstruction

Research on the optimization model of personalized training for athletes based on machine learning in college physical education

Abstract We use deep learning to extract and identify the physiological indicators, sports, and data analysis parameters, providing scientific references for athletes’ personalized training. This paper aims to construct an optimization system for personalized training for athletes based on machine learning. Initially, we designed the system’s overall structure, hardware, and software scheme, and then we modeled the algorithms for measuring four physiological parameters. The NTC thermistor thermometry method was utilized to measure body temperature, the photoelectric volumetric method for determining blood oxygen and heart rate, and the ECG signal method for calculating blood pressure. The adaptive threshold method uses the human body’s steady-state recognition algorithm, based on a single triaxial acceleration sensor, to recognize steady-state motion states such as resting, running, and walking. After the construction of the system was completed, it was applied to the sports team of a university in Guangzhou, and a semester of personalized training optimization experiments for athletes was conducted. The system found a maximum error of 0.02 °C between the measured body temperature and the results from the medical temperature gun, ensuring stable performance during short-term body temperature changes like training. Except for B, C, D, and G errors of 1%, the rest of the athletes monitored by the system in this paper’s blood oxygen results are consistent with the oximeter. The accuracy of recognizing each sport is above 90%. In the eight swimming events that need to be assessed, the average score before the beginning of the semester was 65.43, and after the experiment, it was 87.14, an improvement of 21.71 points. The system in this paper has a great auxiliary effect on the optimization of personalized training for athletes in terms of physical function monitoring, sports status statistics, and training intensity setting, which provides new ideas and methods for the integration of personalized training for athletes with cutting-edge information technology.

Comparison Between Threshold Method and Artificial Intelligence Approaches for Early Warning of Respiratory Infectious Diseases - Weifang City, Shandong Province, China, 2020-2023.

Respiratory infectious diseases, such as influenza and coronavirus disease 2019 (COVID-19), present significant global public health challenges. The emergence of artificial intelligence (AI) and big data offers opportunities to improve traditional disease surveillance and early warning systems. The study analyzed data from January 2020 to May 2023, comprising influenza-like illness (ILI) statistics, Baidu index, and clinical data from Weifang. Three methodologies were evaluated: the adaptive dynamic threshold method (ADTM) for dynamic threshold adjustments, the machine learning supervised method (MLSM), and the machine learning unsupervised method (MLUM) utilizing anomaly detection. The comparison focused on sensitivity, specificity, timeliness, and warning consistency. ADTM issued 37 warnings with a sensitivity of 71% and a specificity of 85%. MLSM generated 35 warnings, with a sensitivity of 82% and a specificity of 87%. MLUM produced 63 warnings with a sensitivity of 100% and specificity of 80%. The initial warnings from ADTM and MLUM preceded those from MLSM by five days. The Kappa coefficient indicated moderate agreement between the methods, with values ranging from 0.52 to 0.62 (P<0.05). The study explores the comparison between traditional methods and two machine learning approaches for early warning systems. It emphasizes the validation of machine learning's reliability and underscores the unique advantages of each method. Furthermore, it stresses the significance of integrating machine learning models with various data sources to enhance public health preparedness and response, alongside acknowledging limitations and the need for broader validation.

Adaptive thresholding increases sensitivity to detect changes in the rate of skin conductance responses to psychologically arousing stimuli in both laboratory and ambulatory settings

Psychophysiologists recording electrodermal activity (EDA) often derive measures of slow, tonic activity—skin conductance level (SCL)—and faster, more punctate changes—skin conductance responses (SCRs). A SCR is conventionally considered to have occurred when the local amplitude of the EDA signal exceeds a researcher-determined threshold (e.g., 0.05 μS), typically fixed across study participants and conditions. However, fixed SCR thresholds can preferentially exclude data from individuals with low SCL because their SCRs are smaller on average, thereby reducing statistical power for group-level analyses. Thus, we developed a fixed plus adaptive (FA) thresholding method that adjusts identification of SCRs based on an individual's SC at the onset of the SCR to increase statistical power and include data from more participants. We assess the utility of applying FA thresholding across two independent samples and explore age and race-related associations with EDA outcomes. Study 1 uses wired EDA measurements from 254 healthy adults responding to evocative images and sounds in a laboratory setting. Study 2 uses wireless EDA measurements from 20 children with autism in a clinical environment while they completed behavioral tasks. Compared to a 0.01, 0.03, and 0.05 μS fixed threshold, FA thresholding at 1.9% modestly increases statistical power to detect a difference in SCR rate between tasks with higher vs. lower subjective arousal and reduces exclusion of participants by up to 5% across both samples. This novel method expands the EDA analytical toolbox and may be useful in populations with highly variable basal SCL or when comparing groups with different basal SCL. Future research should test for reproducibility and generalizability in other tasks, samples, and contexts. Impact statementsThis article is important because it introduces a novel method to enhance sensitivity and statistical power in analyses of skin conductance responses from electrodermal data.

Automatic detection and characterization of uterine contraction using Electrohysterography

Preterm birth is the leading cause of perinatal morbidity and mortality. In clinical practice, the information of uterine contraction is an important reference for preterm delivery. The commonly used tocography method for detecting uterine contractions has low sensitivity and is not suitable for long-term measurement. The electrohysterography (EHG) can record the electrical activity of uterine muscle cells associated with contractions through electrodes on the abdomen. Therefore, this paper proposes a new automatic detection method of uterine contractions based on EHG signals. Specifically, utilizing the nonlinear property of entropy analysis, contraction and spike noise are distinguished in the complexity domain to highlight contraction activity. Then, the location of the contraction is detected by an adaptive thresholding method. Finally, the detected contractions are analyzed for pregnancy and labor. The results show that compared with the existing root mean square methods, the proposed method is less affected by spike noise, and the contraction detection rate on the Icelandic EHG database reaches 87.9%. Meanwhile, features extracted from detected contractions have significant differences between non-labor and labor categories. This demonstrates the feasibility of contraction detection by EHG signal, which can contribute to better care for high-risk pregnant women.

Mapping large-scale aquaculture ponds in Jiangsu Province, China: An automatic extraction framework based on Sentinel-1 time-series imagery

Aquaculture holds significant importance in supporting food security, economic development, resource conservation, and ecological balance. However, the rapid expansion of aquaculture ponds has caused many environmental problems. Studies on the trends and spatial distribution of aquaculture ponds in Jiangsu province, China, are limited, and there is a lack of aquaculture datasets with a 10-m resolution. This study aimed to develop a novel assessment framework for mapping aquaculture ponds using the Google Earth Engine (GEE) platform and Sentinel-1 time-series images. We integrated multi-threshold segmentation and object-oriented classification to extract spatial data on aquaculture ponds from 2017 to 2021 in Jiangsu Province. Results showed that (1) local adaptive threshold methods could extract data about water bodies more effectively. Based on the time series of each image element, calculating its water frequency could effectively exclude seasonal water. (2) This method could effectively map aquaculture ponds in Jiangsu Province. The overall identification accuracy was higher than 93%, and the Kappa coefficient was higher than 0.87. These results are in accordance with those in the statistical yearbook. (3) The total aquaculture pond area was largest in 2017 (3404.00 km2) and smallest in 2021 (2939.51 km2). The aquaculture pond area in Jiangsu Province showed a slight downward trend over the five-year period. Aquaculture ponds were mainly distributed in central and southern Jiangsu and the coastal areas, with the highest density at the junction of Yangzhou, Taizhou, and Yancheng. (4) From 2017 to 2021, Aquaculture ponds spanning 104.65 km2 (accounting for 55.28% of the total reduced area) were transformed into wetlands. In the coastal zone. Our results substantiate that aquaculture pond removal can accelerate wetland restoration. This study maximised the identification of aquaculture ponds as independent targets. Taken together, these results can provide a reference for aquaculture green development and tail water management in Jiangsu Province.

Innovative Application of Visual Communication of Yi Lacquer Art under Internet Technology

Abstract Based on archaeological and historical data, this paper firstly researches and explains the source and flow of Yi lacquer art in Liangshan and proposes the visual communication innovation strategy of Yi lacquer art. Secondly, it uses Internet technology to collect images of Yi lacquer art, grayscale them on the basis of obtaining images of Yi lacquer art, and then facilitates the adaptive threshold method to realize the conversion of the bitmap into a vector map. Then, considering that the visual database pattern resource of Yi lacquer art contains not only pattern information but also color information, we constructed the visual communication database of Yi lacquer art images and carried out an example analysis of the visual communication of Yi lacquer art images based on adaptive thresholding method. The results show that the accuracy, precision, and DSC of R-AMRFCM are improved by 0.0978, 0.4537, and 0.5436, respectively, compared with the GSRG algorithm in algorithmic analysis. In the analysis of Yi lacquer art communication behaviors, behavioral cognition (t=−4.108, p&lt;0.01), behavioral habit (t=−3.090, p&lt;0.01), behavioral intention (t =−2.848, p&lt;0.01), emotional experience (t=−2.277p&lt;0.05), subjective criteria (t=−4.241, p&lt;0.01), and attitude towards Yi lacquer art communication (t=−5.268, p&lt;0.01) are all significantly different. This study optimizes the details of Yi lacquer art dissemination in Internet technology and effectively promotes the visual dissemination and high-quality development of Yi lacquer art.

Optimized variational mode decomposition algorithm based on adaptive thresholding method and improved whale optimization algorithm for denoising magnetocardiography signal

In the diagnosis and treatment of cardiac diseases, magnetocardiography (MCG) technology is characterized by non-invasive, non-contact, and high precision. Therefore, it has currently become a research hotspot in the field of new medical technologies. However, due to the weak signal of the magnetocardiography, the noise needs to be filtered after acquisition. In this paper, an optimized variational mode decomposition algorithm based on the improved threshold method and the improved whale optimization algorithm (WOA) is proposed to process the MCG signal. In order to improve the denoising accuracy, the improved whale optimization algorithm, VMD algorithm, and the improved threshold method are combined. Firstly, the correlation coefficients are obtained by the improved whale optimization algorithm to decompose the IMFs, then the baseline drift is removed by using the moving average method for the low-frequency IMFs, and then the improved thresholding algorithm is applied to each IMF. Finally, the denoised signal is obtained by integration. Experimental tests show that the algorithm has good denoising performance compared with similar algorithms and can filter environmental noise as much as possible without changing the original signal information. The proposed method has the highest Signal-to-Noise Ratio improvement (SNRimp) and Correlation Coefficient (CC) and the lowest Percentage Root Mean Square Difference (PDR). Also, the method is validated in real MCG signal processing. The proposed algorithm can be applied in the field of signal denoising, and it has a wide range of application backgrounds.

Heating pipeline identification and leakage detection method based on improved R3Det

In response to the frequent occurrence of leakage accidents in heating pipelines, timely detection of leakage points in such pipelines is of great significance to ensure the safe operation of heating systems. This article proposes a method for detecting leakage points in heating pipelines using drones equipped with infrared thermal imagers, employing a combination of the improved R3Det algorithm and the adaptive threshold method. Firstly, the algorithm identifies the area of the heating pipeline and then employs the adaptive threshold method to detect the presence of leakage points in the identified pipeline area. Additionally, taking into account the morphological characteristics of heating pipelines, the R3Det network is enhanced by introducing variable convolution, enabling more precise extraction of pipeline features. To reduce model overfitting and enhance network expression capabilities, the H-swish activation function is employed to replace the original activation function. Furthermore, candidate anchor boxes are clustered using the K-means++ clustering algorithm to obtain better position regression results and improve training efficiency. The improved algorithm demonstrates significantly better positioning precision compared to the original network. Moreover, an adaptive threshold algorithm is proposed for leak detection and labelling, utilising the original temperature information contained in infrared images. The experimental results demonstrate that this method achieves higher accuracy in detecting leaks in heating pipelines.

A self-evolving deep learning algorithm for automatic oil spill detection in Sentinel-1 SAR images

Oil spill accidents are one of the major problems causing marine pollution, and thus such accidents require rapid detection for early response. In recent years, deep learning algorithms for oil spill detection have been developed for analyzing SAR images. Nevertheless, to generation of deep learning training data using visual inspection is not only a time-consuming and labor-intensive, but also cause bias and lack of diversity in oil slick training data. This has easily led to false positives in challenging SAR images, such as large scale look-alikes, biologic surface films or oil slicks surrounded by look-alikes. In order to accommodate a broader range of SAR oil spill scenes, including those under challenging conditions, it is essential to continuously enhance the performance of oil spill detection algorithms. In this study, a novel self-evolving algorithm for automatic oil spill detection is proposed, which consists of three inter-connected modules: 1) oil spill detection, 2) generation of new training data, and 3) enhancement of deep learning models. The algorithm detected oil slicks automatically while the new SAR image was added, and then the additional high-quality training data was generated by an adaptive thresholding method to increase the performance of deep learning models. In order to detect oil slicks from whole SAR image with oil look-alikes, the variation of the backscattering coefficients near the oil boundary was considered as a key parameter to distinguish oil spills from look-alikes. As a result, after 21 self-evolving training cycles, 27 new training data were generated with an improved F1-score rising from 0.8423 to 0.8896. Despite the existence of look-alikes in various magnitudes, the algorithm successfully identified oil slicks, and the parameters of deep learning models were automatically updated. It was demonstrated that the performance of the algorithm was gradually improved by self-evolving without any human intervention. Additionally, the fundamental limitations of oil spill detection algorithms were mitigated by the proposed approaches, providing more possibilities for subsequent studies.

An adaptive threshold method for multi-faults diagnosis of lithium-ion batteries based on electro-thermal model

Lithium-ion batteries have emerged as the primary power source for electrochemical energy storage, making battery safety a crucial concern. This study focuses on the detection of thermal faults in lithium-ion batteries, specifically heat generation faults, thermal parameter faults, and temperature sensor faults. To address the issue of low accuracy in temperature indirect estimation models, a framework that combines electrical and thermal models is proposed. This framework employs particle swarm optimization to identify optimal thermal model parameters using real-time open circuit voltage and terminal voltage data. Additionally, an unscented Kalman filter estimator is utilized to predict core temperature based on surface temperature measurements in real time. The results demonstrate that the proposed framework achieves rapid convergence of core and surface temperature biases within a range of 1 ℃ and 0.1 ℃, respectively. Furthermore, an adaptive threshold method is developed to diagnose faults by analyzing the change rates of the estimated temperatures. Experimental results reveal that the proposed methods exhibit quick response times for four different fault conditions, indicating their effectiveness and accuracy in multi-fault diagnosis.

Influence of natural factors and land use change on changes in the main lake area in China over the past 30 years

Being one of the main elements of water resources, lakes perform an indispensable part in social production as well as in life. During recent decades, the area of the lake has significantly altered due to global climate variability and growing intensity of the anthropogenic activities. The shrinkage and expansion of lakes have seriously affected both its ecological function and the sustainable development of the regional economy. With this study, we obtain the areas of 38 lakes in China by using the Landsat serial of remote-sensing satellite images in the period 1989–2019 in combination with NDWI water body index and Otsu adaptive thresholding method according to Google Earth Engine (GEE) to explore the trends of lake area changes and related drivers. The outcomes revealed that (1) the overall lake area in the Yunnan-Guizhou Plateau Lake Region (YGP_LR) showed a decreasing trend during the period from 1989 to 2019; the total area of the two lakes(Selin Co and Yangzhuoyong Co) in the Qinghai-Tibet Plateau (QTP_LR) shows an expanding trend; the area of Wulungu Lake in the Mong-Xin Plateau Lake Region (MXP_LR) and Xingkai Lake in the Northeast Plain and Mountain Lake Region (NPM_LR) on a tendency to expand; the area of lakes in the Eastern Plain Lake Region (EP_LR) fluctuated greatly, and the overall area shows a decreasing trend. (2) The increase of temperature promoted the expansion of lake area fed by glacier melt. The increase of rainfall is conducive to the expansion of lake area. The lake area of YGP_LR and EP_LR is closely related to the conversion of other land use types. (3) The lake area of the QTP_LR (Selin Co, Yangzhuoyong Co) varies are mainly affected by climatic factors. Lake area changes in the YGP_LR, Northeast Plain and Mountainous Lake Region (Xingkai Lake), MXP_LR (Wulungu Lake), and EP_LR are influenced by a combination of both climatic changes and the activities of humans. Impact factors of area change in the NPM_LR (Xingkai Lake) need to be further explored.

An efficient load balancing technique using CAViaR-HHO enabled VM migration and replica management in cloud computing

Cloud computing is immense technology that offers distributed resources to a number of users who are present throughout the world. Cloud model is comprised of numerous virtual machines (VMs) and physical machines (PMs) to carry out user tasks effectively in a parallel manner but in some cases, the demand of the users may be high that resulting in the overloading of PMs and this condition deteriorates the performance of cloud network. For achieving effective virtualization in the cloud paradigm, energy and resource utilization are major properties that should be handled effectively and such properties are accomplished through effective management of workload by distributing load equivalently among VMs. By doing so, resource utilization of the network is enhanced and it only requires minimum energy to process the tasks. Numerous load-balancing algorithms have been introduced earlier to maintain load in a cloud environment, nevertheless, they are devoid of mitigating the number of task migrations. Hence, this research proposes an effective load balancing algorithm and replica management method using the proposed Conditional Autoregressive Value at risk by Regression Quantiles-Horse Herd Optimization (CAViaR-HHO) model. Here, the load is computed by considering some factors like Central Processing Unit (CPU), Million Instructions per Second (MIPS), bandwidth, memory, and frequency. VM migration and replica migration is effectively carried out using the proposed CAViaR-HHO model. Meanwhile, the developed method is devised by integration of Conditional Autoregressive Value at risk by Regression Quantiles (CAViaR) with Horse Herd Optimization Algorithm (HOA). However, the proposed CAViaR-HHO has achieved a load with a minimum value of 0.109, capacity with a maximum value of 0.591, resource utilization with a maximum value of 0.467, and minimum cost of 0.344. Using setup-1, when the number of tasks is 500, the capacity of the proposed method is 5.58%, 3.89%, 2.87%, 1.52%, and 0.67% higher when compared to the existing approaches namely, C-FDLA, K-means clustering + LB, Adaptive starvation threshold, EIMORM, and Dynamic replica creation method.

A holistic human activity recognition optimisation using AI techniques

AbstractBuilding on previous radar‐based human activity recognition (HAR), we expand the micro‐Doppler signature to 6 domains and exploit each domain with a set of handcrafted features derived from the literature and our patents. An adaptive thresholding method to isolate the region of interest is employed, which is then applied in other domains. To reduce the computational burden and accelerate the convergence to an optimal solution for classification accuracy, a holistic approach to HAR optimisation is proposed using a surrogate model‐assisted differential evolutionary algorithm (SADEA‐I) to jointly optimise signal processing, adaptive thresholding and classification parameters for HAR. Two distinct classification models are evaluated with holistic optimisation: SADEA‐I with support vector machine classifiers (SVM) and SADEA‐I with AlexNet. They achieve an accuracy of 89.41% and 93.54%, respectively. This is an improvement of ∼11.3% for SVM and ∼2.7% for AlexNet when compared to the performance without SADEA‐I. The effectiveness of our holistic approach is validated using the University of Glasgow human radar signatures dataset. This proof of concept is significant for dimensionality reduction and computational efficiency when facing a multiplication of radar representation domains/feature spaces and transmitting/receiving channels that could be individually tuned in modern radar systems.