In today’s world of massive data and interconnected networks,it’s crucial to burgeon a secure and efficient digital watermarking method to protect the copyrights of digital content.Existing research primarily focuse...In today’s world of massive data and interconnected networks,it’s crucial to burgeon a secure and efficient digital watermarking method to protect the copyrights of digital content.Existing research primarily focuses on deep learning-based approaches to improve the quality of watermarked images,but they have some flaws.To overcome this,the deep learning digital image watermarking model with highly secure algorithms is proposed to secure the digital image.Recently,quantum logistic maps,which combine the concept of quantum computing with traditional techniques,have been considered a niche and promising area of research that has attracted researchers’attention to further research in digital watermarking.This research uses the chaotic behaviour of the quantum logistic map with Rivest–Shamir–Adleman(RSA)and Secure Hash(SHA-3)algorithms for a robust watermark embedding process,where a watermark is embedded into the host image.This way,the quantum chaos method not only helps limit the chance of tampering with the image content through reverse engineering but also assists in maintaining a high level of imperceptibility and strong robustness with efficient extraction or detection of watermark images.Lifting Wavelet Transformation(LWT)is a potential and computationally efficient version of traditional Discrete Wavelet Transform(DWT)where the host image is divided into four sub-bands to offer a multi-resolution view of an image with greater flexibility in watermarking methodologies.Furthermore,considering the robustness against attacks,a pre-trained Residual Neural Network(ResNet-50),a convolutional neural network with 50 layers deep,is used to better learn the complex features and efficiently extract the watermark from the image.By integrating RSA and SHA-3 algorithms,the proposed model demonstrates improved imperceptibility,robustness,and accuracy in watermark extraction compared to traditional methods.It achieves a Peak Signal-to-Noise Ratio(PSNR)of 49.83%,a Structural Similarity Index Measure(SSIM)of 0.98,and a Number of Pixels Change Rate(NPCR)of 99.79%,respectively.These results reflect the model’s effectiveness in delivering superior quality and security.Consequently,our proposed approach offers accurate results,exceptional invisibility,and enhanced robustness compared to the existing digital image watermarking techniques.展开更多
为解决现有工件分类模型在处理高分辨率图像时非感兴趣区域(region of interest,ROI)的冗余计算问题,提出了1种ROI自适应轮廓驱动裁剪的工件分类网络模型(a workpiece classification network model with ROI adaptive contour-driven c...为解决现有工件分类模型在处理高分辨率图像时非感兴趣区域(region of interest,ROI)的冗余计算问题,提出了1种ROI自适应轮廓驱动裁剪的工件分类网络模型(a workpiece classification network model with ROI adaptive contour-driven cropping,ACDC-ClassNet)。该模型利用轮廓检测定位图像最大轮廓及其中心,据此生成标准化方形ROI裁剪区域,消除背景干扰。该模型采用预训练的50层残差网络(residual network-50 layers,ResNet-50),调整其分类头以适应工件类别数量,实现高效的特征聚焦与分类。结果表明,该ROI裁剪策略平均减少72.15%的冗余区域面积,使模型更专注于工件细节。相较于原始ResNet-50,ACDC-ClassNet在准确率、精确率、召回率、F1分数等指标上分别提升3.83、4.04、3.64、4.13个百分点。同时,该策略也优于高效网络(efficient network,EfficientNet)、视觉变换器(vision transformer,ViT)模型,准确率分别提升8.40、4.27个百分点。ACDC-ClassNet为工业背景下的高效视觉检测提供了新的技术路径。展开更多
Accident detection plays a critical role in improving traffic safety by enabling timely emergency response and reducing the impact of road incidents.The main challenge lies in achieving real-time,reliable and highly a...Accident detection plays a critical role in improving traffic safety by enabling timely emergency response and reducing the impact of road incidents.The main challenge lies in achieving real-time,reliable and highly accurate detection across diverse Internet-of-vehicles(IoV)environments.To overcome this challenge,this paper leverages deep learning to automatically learn patterns from visual data to detect accidents with high accuracy.A visual classification model based on the ResNet-50 architecture is presented for distinguishing between accident and non-accident images.The model is trained and tested on a labeled dataset and achieves an overall accuracy of 91.84%,with a precision of 94%,recall of 90.38%,and an F1-score of 92.14%.Training behavior is observed over 100 epochs,where the model has shown rapid accuracy gains and loss reduction within the first 30 epochs,followed by gradual stabilization.Accuracy plateaues between 90−93%,and loss values remain consistent between 0.1 and 0.2 in later stages.To understand the effect of training strategy,the model is optimized using three different algorithms,namely,SGD,Adam,and Adadelta with all showing effective performance,though with varied convergence patterns.Further,to test its effectiveness,the proposed model is compared with existing models.In the end,the problems encountered in implementing the model in practical automotive settings and offered solutions are discussed.The results support the reliability of the approach and its suitability for real-time traffic safety applications.展开更多
为提高电力设备状态监测的智能化水平,设计一种基于无线移动通信技术的电力故障诊断系统。该系统采用“采集-传输-诊断”三层架构,集成STM32H743处理器、5G/窄带物联网(Narrow Band Internet of Things,NB-IoT)双模通信和ResNet-50改进...为提高电力设备状态监测的智能化水平,设计一种基于无线移动通信技术的电力故障诊断系统。该系统采用“采集-传输-诊断”三层架构,集成STM32H743处理器、5G/窄带物联网(Narrow Band Internet of Things,NB-IoT)双模通信和ResNet-50改进模型,通过自适应采样机制、MQTT 5.0和深度学习算法实现电力设备状态的实时监测与智能诊断。实验结果表明,该系统在有效采样数据占比、平均通信时延、数据丢包率、平均诊断准确率以及整体能耗5项关键性能指标上均显著优于传统数据采集与监控(Supervisory Control And Data Acquisition,SCADA)系统,为智能电网建设提供可靠的技术支撑。展开更多
文摘In today’s world of massive data and interconnected networks,it’s crucial to burgeon a secure and efficient digital watermarking method to protect the copyrights of digital content.Existing research primarily focuses on deep learning-based approaches to improve the quality of watermarked images,but they have some flaws.To overcome this,the deep learning digital image watermarking model with highly secure algorithms is proposed to secure the digital image.Recently,quantum logistic maps,which combine the concept of quantum computing with traditional techniques,have been considered a niche and promising area of research that has attracted researchers’attention to further research in digital watermarking.This research uses the chaotic behaviour of the quantum logistic map with Rivest–Shamir–Adleman(RSA)and Secure Hash(SHA-3)algorithms for a robust watermark embedding process,where a watermark is embedded into the host image.This way,the quantum chaos method not only helps limit the chance of tampering with the image content through reverse engineering but also assists in maintaining a high level of imperceptibility and strong robustness with efficient extraction or detection of watermark images.Lifting Wavelet Transformation(LWT)is a potential and computationally efficient version of traditional Discrete Wavelet Transform(DWT)where the host image is divided into four sub-bands to offer a multi-resolution view of an image with greater flexibility in watermarking methodologies.Furthermore,considering the robustness against attacks,a pre-trained Residual Neural Network(ResNet-50),a convolutional neural network with 50 layers deep,is used to better learn the complex features and efficiently extract the watermark from the image.By integrating RSA and SHA-3 algorithms,the proposed model demonstrates improved imperceptibility,robustness,and accuracy in watermark extraction compared to traditional methods.It achieves a Peak Signal-to-Noise Ratio(PSNR)of 49.83%,a Structural Similarity Index Measure(SSIM)of 0.98,and a Number of Pixels Change Rate(NPCR)of 99.79%,respectively.These results reflect the model’s effectiveness in delivering superior quality and security.Consequently,our proposed approach offers accurate results,exceptional invisibility,and enhanced robustness compared to the existing digital image watermarking techniques.
文摘为解决现有工件分类模型在处理高分辨率图像时非感兴趣区域(region of interest,ROI)的冗余计算问题,提出了1种ROI自适应轮廓驱动裁剪的工件分类网络模型(a workpiece classification network model with ROI adaptive contour-driven cropping,ACDC-ClassNet)。该模型利用轮廓检测定位图像最大轮廓及其中心,据此生成标准化方形ROI裁剪区域,消除背景干扰。该模型采用预训练的50层残差网络(residual network-50 layers,ResNet-50),调整其分类头以适应工件类别数量,实现高效的特征聚焦与分类。结果表明,该ROI裁剪策略平均减少72.15%的冗余区域面积,使模型更专注于工件细节。相较于原始ResNet-50,ACDC-ClassNet在准确率、精确率、召回率、F1分数等指标上分别提升3.83、4.04、3.64、4.13个百分点。同时,该策略也优于高效网络(efficient network,EfficientNet)、视觉变换器(vision transformer,ViT)模型,准确率分别提升8.40、4.27个百分点。ACDC-ClassNet为工业背景下的高效视觉检测提供了新的技术路径。
基金the Deanship of Graduate Studies and Scientific Research at Najran University for funding this work under the Growth Funding Program grant code(NU/GP/SERC/13/358-6)。
文摘Accident detection plays a critical role in improving traffic safety by enabling timely emergency response and reducing the impact of road incidents.The main challenge lies in achieving real-time,reliable and highly accurate detection across diverse Internet-of-vehicles(IoV)environments.To overcome this challenge,this paper leverages deep learning to automatically learn patterns from visual data to detect accidents with high accuracy.A visual classification model based on the ResNet-50 architecture is presented for distinguishing between accident and non-accident images.The model is trained and tested on a labeled dataset and achieves an overall accuracy of 91.84%,with a precision of 94%,recall of 90.38%,and an F1-score of 92.14%.Training behavior is observed over 100 epochs,where the model has shown rapid accuracy gains and loss reduction within the first 30 epochs,followed by gradual stabilization.Accuracy plateaues between 90−93%,and loss values remain consistent between 0.1 and 0.2 in later stages.To understand the effect of training strategy,the model is optimized using three different algorithms,namely,SGD,Adam,and Adadelta with all showing effective performance,though with varied convergence patterns.Further,to test its effectiveness,the proposed model is compared with existing models.In the end,the problems encountered in implementing the model in practical automotive settings and offered solutions are discussed.The results support the reliability of the approach and its suitability for real-time traffic safety applications.
文摘为提高电力设备状态监测的智能化水平,设计一种基于无线移动通信技术的电力故障诊断系统。该系统采用“采集-传输-诊断”三层架构,集成STM32H743处理器、5G/窄带物联网(Narrow Band Internet of Things,NB-IoT)双模通信和ResNet-50改进模型,通过自适应采样机制、MQTT 5.0和深度学习算法实现电力设备状态的实时监测与智能诊断。实验结果表明,该系统在有效采样数据占比、平均通信时延、数据丢包率、平均诊断准确率以及整体能耗5项关键性能指标上均显著优于传统数据采集与监控(Supervisory Control And Data Acquisition,SCADA)系统,为智能电网建设提供可靠的技术支撑。
