检索结果-内蒙古大学图书馆

Surface defect detection method of aluminium alloy castings based on data enhancement and CRT-DETR

IET image processing 2024年第13期18卷 4275-4286页

作者： Li, Long Gao, Mingyang Tian, Xing Wang, Chengjun Yu, Jiangli Anhui Univ Sci & Technol Coll Artificial Intelligence 168 Taifeng Ave Huainan 232001 Anhui Peoples R China Anhui Univ Sci & Technol Affiliated Hosp 1 Huainan Peoples Hosp 1 Huainan Anhui Peoples R China

The surface quality of aluminium alloy castings is crucial to quality control. Aiming to address the challenges of limited samples and extensive computation in deep learning-based surface defect detection for aluminium alloy castings, this paper proposes a surface defect detection method based on data enhancement and the Casting real-time DEtection TRansformer. First, to tackle the issue of small sample sizes and uneven distribution in surface defect data sets of aluminium alloy castings, ECA-MetaAconC deep Convolution Generative Adversarial Networks is proposed for generating defects with fewer samples and employ the image augmentation (IMGAUG) library for sample enhancement. Second, building upon the real-time DEtection TRansformer (RT-DETR), a lightweight partial-rep convolution is designed to decrease the network's parameter count. Simultaneously, the Deformable attention module and the DRBC3 module are introduced to enhance the neck network, thereby improving the model's capability to capture information and enhancing its detection performance. Compared to RT-DETR, this method reduces the number of model parameters by 38.7%, increases mAP by 1.5%, and achieves a frame rate that is 1.58 times higher than the original model. The experimental results demonstrate that this method can effectively and accurately detect surface defects in aluminium alloy castings, satisfying industrial requirements. The surface quality of aluminium alloy castings is very important in quality control. Aiming at the problems of few samples, large amount of calculation, and poor real-time performance in the surface defect detection method of aluminium alloy castings based on deep learning, here a surface defect detection method based on data enhancement and Casting real-time DEtection TRansformer is designed. image

关键词： computer vision image processing

来源：评论

学校读者我要写书评

暂无评论

real-time Large-Motion Deblurring for Gimbal-Based Imaging Systems

引用

IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL processing 2024年第3期18卷 346-357页

作者： Varghese, Nisha Rajagopalan, A. N. Ansari, Zahir Ahmed Indian Inst Technol Madras Dept Elect Engn Chennai 600036 India IRDE DRDO Dehra Dun 248001 India

Robotic systems employed in tasks such as navigation, target tracking, security, and surveillance often use camera gimbal systems to enhance their monitoring and security capabilities. These camera gimbal systems undergo fast to-and-fro rotational motion to surveil the extended field of view (FOV). A high steering rate (rotation angle per second) of the gimbal is essential to revisit a given scene as fast as possible, which results in significant motion blur in the captured video frames. real-time motion deblurring is essential in surveillance robots since the subsequent image-processing tasks demand immediate availability of blur-free images. Existing deep learning (DL) based motion deblurring methods either lack real-time performance due to network complexity or suffer from poor deblurring quality for large motion blurs. In this work, we propose a Gyro-guided Network for real-time motion deblurring (GRNet) which makes effective use of existing prior information to improve deblurring without increasing the complexity of the network. The steering rate of the gimbal is taken as a prior for data generation. A contrastive learning scheme is introduced for the network to learn the amount of blur in an image by utilizing the knowledge of blur content in images during training. To the GRNet, a sharp reference image is additionally given as input to guide the deblurring process. The most relevant features from the reference image are selected using a cross-attention module. Our method works in real-time at 30 fps. As a first, we propose a Gimbal Yaw motion real-wOrld (GYRO) dataset of infrared (IR) as well as color images with significant motion blur along with the inertial measurements of camera rotation, captured by a gimbal-based imaging setup where the gimbal undergoes rotational yaw motion. Both qualitative and quantitative evaluations on our proposed GYRO dataset, demonstrate the practical utility of our method.

关键词： Robotic systems gimbal-based imaging real-time motion deblurring gyro-aided motion deblurring video surveillance reference-based motion deblurring contrastive learning Robotic systems gimbal-based imaging real-time motion deblurring gyro-aided motion deblurring video surveillance reference-based motion deblurring contrastive learning

来源：评论

学校读者我要写书评

暂无评论

deep learning-based medical image processing of lung CT

Deep learning-based medical image processing of lung CT

引用

Electronic Technology, Communication and Information (ICETCI), IEEE International Conference on

作者： Zeyu Pan College of Physics and Optoelectronic Engineering Shenzhen University Shenzhen China

ISBN: (数字)9798350361643

ISBN: (纸本)9798350361650

Aiming at the planning and navigation needs in lung disease surgery, and the time-consuming and laborious manual labeling of lung CT arterial segmentation, a 3DUNet architecture neural network incorporating CBAM attention mechanism is proposed for fast extraction of lung arterial veins. Based on the 2DUNet architecture, the dimension of the processed image is upgraded from two-dimensional to three-dimensional 3DUNet, and the CBAM attention module is added before pooling or up-convolution at the end of each convolution layer, and the parameter update is performed by using the Focal Loss loss function combined with stochastic gradient descent method. A three-fold cross-validation was performed using 95 lung CT data and ablation experiments were performed. After experimental validation, the model with the addition of CBAM attention module and Focal Loss loss function showed significant improvement in segmentation metrics, in which the accuracy, sensitivity, Dice coefficient, and precision for arteries and veins reached 99.8% and 99.8%, 84.9% and 87.2%, 82.5% and 84.8%, 81.1% and 83.5%, respectively.

关键词： deep learning Training image segmentation Accuracy Veins Computed tomography Lung

来源：评论

学校读者我要写书评

暂无评论

Enhancing fitness action recognition with ResNet-TransFit: Integrating IoT and deep learning techniques for real-time monitoring

引用

ALEXANDRIA ENGINEERING JOURNAL 2024年 109卷 89-101页

作者： Li, Jijie Gong, Ruyao Wang, Gang Kunsan Natl Univ Phys Educ Gunsan 54150 Jeollabuk South Korea Linyi Univ Coll Sports & Hlth Linyi 276000 Shandong Peoples R China NingboTech Univ Sch Comp & Data Engn Ningbo 315100 Peoples R China Imperial Coll London Dept Bioengn London SW7 2AZ England

With the growing popularity of fitness, the demand for real-time action recognition and feedback is increasing. Current research faces challenges in handling complex actions, real-time processing, and system integration. address these issues, we propose a novel fitness action recognition model that integrates ResNet, Transformer, and transfer learning techniques. Specifically, ResNet is used for image feature extraction, Transformer handles time-series data processing, and transfer learning accelerates the model's adaptation to new data. We evaluated our model on the NTU RGB+D action recognition dataset, achieving 48.5 ms latency, 29.1 fps throughput, and 93.7% accuracy, significantly outperforming other models. Our model achieved an accuracy improvement 5% over existing methods, demonstrating significant potential for real-time fitness monitoring. By incorporating IoT technology, our system enables real-time data processing and action recognition, making it ideal for smart fitness monitoring. Although the model has high complexity and memory usage, its efficiency and accuracy demonstrate its potential for widespread adoption. Future work will focus on optimizing the model structure and training methods to enhance applicability in resource-constrained environments, ensuring broader usability and efficiency in various real-world applications.

关键词： Action recognition image analysis ResNet Transformer Transfer learning

来源：评论

学校读者我要写书评

暂无评论

Practical real-time image compression for resource-challenged devices

Practical real-time image compression for resource-challenge...

引用

Conference on Multimodal image Exploitation and learning

作者： Pham, Kevin Depoian, Arthur C., II Bailey, Colleen P. Univ North Texas Dept Elect Engn Denton TX 76207 USA

ISBN: (纸本)9781510673854;9781510673847

In the era of rapidly expanding image data, the demand for improved image compression algorithms has grown significantly, particularly with the integration of deep learning approaches into traditional image processing tasks. However, many of the existing solutions in this domain are burdened by computational complexity, rendering them unsuitable for real-time deployment on standard devices as they often necessitate complex systems and substantial energy consumption. This work addresses the growing paradigm of edge computing for real-time applications by introducing a novel, on-edge device solution. This innovative approach aims to strike a balance between efficiency and accuracy, adhering to the practical constraints of real-world deployment. By presenting demonstrations of the proposed solution's performance on readily available devices, we provide tangible evidence of its applicability and viability in real-world scenarios. This advance contributes to the ongoing dialogue about the need for accessible and efficient image compression algorithms that can be deployed real-time applications on edge devices, bridging the gap between the demanding computational requirements of deep learning and the practical limitations of everyday hardware. As data continues to surge, solutions like this become ever more critical in ensuring effective image compression, aligning with on-edge computing within AI. This research paves the way for improved image processing in real-time applications while conserving computational resources and energy consumption.

关键词： image compression efficient AI real time processing

来源：评论

学校读者我要写书评

暂无评论

HDMA-CGAN: Advancing image Style Transfer with deep learning

引用

INTERNATIONAL JOURNAL OF PATTERN RECOGNITION AND ARTIFICIAL INTELLIGENCE 2024年第9期38卷

作者： Liu, Huaqun Hu, Benxi Cao, Yu Beijing Inst Graph Commun Sch New Media 1 Sect 2Xinghua St Beijing 102600 Peoples R China

The integration of artificial intelligence (AI) and deep learning heralds a transformative era in pattern recognition and computer vision, notably in image style transfer. We introduce the hierarchical dynamic multi-attention cycle generative adversarial network (HDMA-CGAN), an innovative deep learning architecture poised to redefine image style transfer capabilities. HDMA-CGAN employs a novel multi-attention mechanism and color optimization strategies, enabling precise style replication with improved fidelity and vibrancy. Our model surpasses existing benchmarks in image quality, validated by leading metrics such as peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and Fr & eacute;chet inception distance (FID). Although HDMA-CGAN advances the state of the art, it necessitates high computational resources and faces challenges with very high-resolution images. Future work could explore optimizing the model's efficiency for real-time applications and extending its application to video content. This work enhances the tools available for visual content creation and digital media enhancement, leveraging advanced pattern recognition and AI techniques to significantly impact computer vision and image processing.

关键词： deep learning image style transfer HDMA-CGAN generative adversarial networks multi-attention mechanisms color space optimization

来源：评论

学校读者我要写书评

暂无评论

An improved pulse coupled neural networks model for semantic IoT

引用

Digital Communications and Networks 2024年第3期10卷 557-567页

作者： Rong Ma Zhen Zhang Yide Ma Xiping Hu Edith C.H.Ngai Victor C.M.Leung School of Information Science and Engineering Lanzhou UniversityLanzhou 730000China School of Medical Technology Beijing Institute of TechnologyBeijing100081China Department of Electrical and Electronic Engineer The University of Hong KongHong Kong 999077China College of Computer Science and Software Engineering Shenzhen UniversityShenzhen 518060China Department of Electrical and Computer Engineering the University of British ColumbiaVancouverBC V6T 1Z4Canada

In recent years,the Internet of Things(IoT)has gradually developed applications such as collecting sensory data and building intelligent services,which has led to an explosion in mobile data ***,with the rapid development of artificial intelligence,semantic communication has attracted great attention as a new communication ***,for IoT devices,however,processing image information efficiently in real time is an essential task for the rapid transmission of semantic *** the increase of model parameters in deep learning methods,the model inference time in sensor devices continues to *** contrast,the Pulse Coupled Neural Network(PCNN)has fewer parameters,making it more suitable for processing real-time scene tasks such as image segmentation,which lays the foundation for real-time,effective,and accurate image ***,the parameters of PCNN are determined by trial and error,which limits its *** overcome this limitation,an Improved Pulse Coupled Neural Networks(IPCNN)model is proposed in this *** IPCNN constructs the connection between the static properties of the input image and the dynamic properties of the neurons,and all its parameters are set adaptively,which avoids the inconvenience of manual setting in traditional methods and improves the adaptability of parameters to different types of *** segmentation results demonstrate the validity and efficiency of the proposed self-adaptive parameter setting method of IPCNN on the gray images and natural images from the Matlab and Berkeley Segmentation *** IPCNN method achieves a better segmentation result without training,providing a new solution for the real-time transmission of image semantic information.

关键词： Internet of things(IoT) Semantic information real-time application Improved pulse coupled neural network image segmentation

来源：评论

学校读者我要写书评

暂无评论

Yolo-global: a real-time target detector for mineral particles

引用

JOURNAL OF real-time image processing 2024年第3期21卷 85-85页

作者： Wang, Zihao Zhou, Dong Guo, Chengjun Zhou, Ruihao Univ Elect Sci & Technol China 2006 Xiyuan Ave Chengdu Peoples R China Chongqing Changan Ind Grp Co Ltd 599 Airport Ave Chongqing Peoples R China

Recently, deep learning methodologies have achieved significant advancements in mineral automatic sorting and anomaly detection. However, the limited features of minerals transported in the form of small particles pose significant challenges to accurate detection. To address this challenge, we propose a enhanced mineral particle detection algorithm based on the YOLOv8s model. Initially, a C2f-SRU block is introduced to enable the feature extraction network to more effectively process spatial redundant information. Additionally, we designed the GFF module with the aim of enhancing information propagation between non-adjacent scale features, thereby enabling deep networks to more fully leverage spatial positional information from shallower networks. Finally, we adopted the Wise-IoU loss function to optimize the detection performance of the model. We also re-designed the position of the prediction heads to achieve precise detection of small-scale targets. The experimental results substantiate the effectiveness of the algorithm, with YOLO-Global achieving a mAP@.5 of 95.8%. In comparison to the original YOLOv8s, the improved model exhibits a 2.5% increase in mAP, achieving a model inference speed of 81 fps, meeting the requirements for real-time processing and accuracy.

关键词： Small target detection Lightweight deep learning Computer vision

来源：评论

学校读者我要写书评

暂无评论

Efficient physics-based learned reconstruction methods for real-time 3D near-field MIMO radar imaging

引用

DIGITAL SIGNAL processing 2024年 144卷

作者： Manisali, Irfan Oral, Okyanus Oktem, Figen S. Middle East Tech Univ METU Dept Elect & Elect Engn TR-06800 Ankara Turkiye

Near-field multiple-input multiple-output (MIMO) radar imaging systems have recently gained significant attention. These systems generally reconstruct the three-dimensional (3D) complex-valued reflectivity distribution of the scene using sparse measurements. Consequently, imaging quality highly relies on the image reconstruction approach. Existing analytical reconstruction approaches suffer from either high computational cost or low image quality. In this paper, we develop novel non-iterative deep learning-based reconstruction methods for real-time near-field MIMO imaging. The goal is to achieve high image quality with low computational cost at compressive settings. The developed approaches have two stages. In the first approach, physics-based initial stage performs adjoint operation to back-project the measurements to the image-space, and deep neural network (DNN)-based second stage converts the 3D backprojected measurements to a magnitude-only reflectivity image. Since scene reflectivities often have random phase, DNN processes directly the magnitude of the adjoint result. As DNN, 3D U-Net is used to jointly exploit range and cross-range correlations. To comparatively evaluate the significance of exploiting physics in a learning-based approach, two additional approaches that replace the physics-based first stage with fully connected layers are also developed as purely learning based methods. The performance is also analyzed by changing the DNN architecture for the second stage to include complex-valued processing (instead of magnitude-only processing), 2D convolution kernels (instead of 3D), and ResNet architecture (instead of U-Net). Moreover, we develop a synthesizer to generate large-scale dataset for training the neural networks with 3D extended targets. We illustrate the performance through experimental data and extensive simulations. The results show the effectiveness of the developed physics based learned reconstruction approach compared to commonly used ap

关键词： Radar imaging Near-field microwave imaging deep learning 3D inverse imaging problems Sparse MIMO array

来源：评论

学校读者我要写书评

暂无评论

real-time disease detection on bean leaves from a small image dataset using data augmentation and deep learning methods: real-time disease detection on bean leaves from a small image dataset using data augmentation..: E. Karantoumanis et al.

引用

Soft Computing 2024年第21期28卷 12929-12941页

作者： Karantoumanis, Emmanouil Balafas, Vasileios Louta, Malamati Ploskas, Nikolaos Department of Electrical and Computer Engineering University of Western Macedonia Campus ZEP Kozani50100 Greece

Disease detection in agricultural crops plays a pivotal role in ensuring food security and sustainable farming practices. deep learning models, known for their ability in image analysis, often demand extensive image datasets and annotations to achieve high accuracy. However, in the case of bean crops, the absence of a publicly available dataset has posed a significant challenge for applying deep learning algorithms to accurately predict diseases. Additionally, the manual annotation of images demands substantial time and resources. This paper introduces an innovative approach to tackle these issues. We introduce a solution for real-time disease detection on bean leaves, despite the lack of bean-specific image data. Initially, we generate a small dataset from real images and annotate them. Then, we utilize images from the existing dataset PlantDoc (Singh et al. in: Proceedings of the 7th ACM IKDD CoDS and 25th COMAD, Association for Computing Machinery, pp 249–253, 2020) from leaves of other plant species. Moreover, to compensate for the limitations of a small image dataset, we employ advanced data augmentation techniques, enriching the training data and enhancing the model’s ability to generalize. Our experimental study shows that data augmentation techniques can improve the accuracy of deep learning methods by up to 37%. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：