检索结果-内蒙古大学图书馆

International Symposium on Parallel and Distributed processing with Applications, ISPA

作者： Min Huang Shuaihui Wang Shuaili Luo Qiuhong Sun Xiaokang Zhou School of Information Science and Engineering Hebei University of Science and Technology Shijiazhuang China Faculty of Business Data Science Kansai University Osaka Japan

ISBN: (数字)9798331509712

ISBN: (纸本)9798331509729

Aiming to address challenges in SAR image ship detection tasks, such as the dense concentration of inshore ships, small target sizes, and large overlapping areas, this paper proposes an improved algorithm LAOBB-YOLOv8n based on YOLOv8n for detecting rotating ships in SAR images. In this model, we propose the Detect_LSCD detection head, which replaces the BatchNorm in the convolutional layers of the Detect module in YOLOv8n with GroupNorm to improve the positioning and classification capabilities of the detection head; Replace the original Spatial Pyramid Pooling-Fast (SPPF) module with the Attention-based Intrascale Feature Interaction (AIFI) module, which focuses on processing advanced image features through self-attention mechanism, thus improving the flexibility and accuracy of the model in processing complex scenes while reducing unnecessary computational consumption; Based on the original network structure, we add a small target detection layer to improve the detection ability for small targets; Use rotating bounding box instead of traditional horizontal bounding box, which can better accommodate the rotating posture of the target and reduce the overlapping areas between detection boxes, thereby improving the accuracy of inshore target detection. The comparative and ablation experiments are conducted on the RSDD-SAR dataset to verify the model’s effectiveness. The experimental results show that the mAP50 of LAOBB-YOLOv8n reaches 98%, which is 1.4% higher than that of YOLOv8n. Although the number of parameters of LAOBB-YOLOv8n is larger than that of YOLOv8n, it is still very small compared with other baseline models, and the accuracy of LAOBB-YOLOv8n is much higher than other models.

关键词： Distributed processing Head Accuracy Computational modeling Object detection Feature extraction Radar polarimetry Robustness Marine vehicles signal to noise ratio

来源：评论

学校读者我要写书评

暂无评论

Yolov9-Spe: A Novel Deep Learning Method For Classroom Student Behavior Detection

SSRN

引用

SSRN 2024年

作者： Liu, Junxiu Wu, Fengxiang Lu, Baoshan Qin, Sheng Fu, Qiang Luo, Yuling Guangxi Key Lab of Brain-inspired Computing and Intelligent Chips School of Electronic and Information Engineering Guangxi Normal University Guilin541004 China Guangxi Wireless Broadband Communication and Signal Processing Key Laboratory Guilin University of Electronic Technology Guilin541004 China

Deep learning can significantly enhance student behavior detection in classrooms. However, challenges such as small target recognition, blurry data, occlusion, and multi-scale detection persist. To address these, we propose YOLOv9-SPE, an improved model based on YOLOv9, incorporating Space-to-Depth Convolution (SPD-Conv), Partial Convolution (PConv), and an Efficient Multi-Scale Attention (EMA) Module. SPD-Conv enhances small target recognition and improves feature extraction for blurry data, while PConv, adapted from FasterNet, addresses occlusion issues. EMA is integrated to handle multi-scale targets in classroom student behavior detection. Experimental results on a student behavior dataset show that YOLOv9-SPE achieves a 2.1% improvement in mean average precision over the original YOLOv9. These enhancements provide a more accurate and efficient method for monitoring and analyzing student behavior in educational settings. © 2024, The Authors. All rights reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

A Tobacco Pest Detection System Based on the Improved YOLO v5 Model

A Tobacco Pest Detection System Based on the Improved YOLO v...

引用

6th International Conference on Intelligent Computing and signal processing (ICSP)

作者： Shaohui Zhou Jinchao Zhao Yong Gan Yuefeng Wang College of Computer Science and Technology Zhengzhou University of Light Industry Zhengzhou China

ISBN: (数字)9798350376548

ISBN: (纸本)9798350376555

During the storage process, tobacco leaves are vulnerable to infestation by Lasioderma serricorne. To achieve precise detection of this pest within stored tobacco leaves and provide real-time updates on pest activity, we propose a detection system based on an improved YOLOv5 model. Initially, an open-source hardware system is established to capture images of the pests, which are then transmitted to a computer terminal via a Wi-Fi network. Random data augmentation techniques, including translation, rotation, and the addition of Gaussian noise, are applied to expand the training set of Lasioderma serricorne images. Subsequently, the BiFormer attention mechanism is integrated to enhance the model's accuracy in detecting small targets such as Lasioderma serricorne. By combining the improved YOLOv5 model with hardware facilities and the detection system, a comprehensive device is formed. The results demonstrate that the improved YOLOv5 network model achieves an accuracy rate of 93.26% and an average detection precision of 92.40% for Lasioderma serricorne. Compared to models such as YOLO v3, Retinanet, Faster R-CNN, the original YOLOv5, and SSD, the improved YOLOv5 model exhibits superior detection performance, with efficiency improvements of 2.76% and 2.46% over the existing YOLOv5 algorithm, respectively. The integration of the improved YOLOv5 model with hardware and the detection system into a comprehensive device fully harnesses the latest advancements in modern computer vision and artificial intelligence technologies. This device not only offers efficient object detection capabilities but also adapts well to various environments and scenarios, opening up a wide range of application possibilities. In this study, the device accurately identifies and counts Lasioderma serricorne, providing technical support for early warning and pest control in tobacco storage systems.

关键词： YOLO Training Adaptation models Accuracy Computational modeling signal processing algorithms signal processing Pest control Real-time systems Wireless fidelity

来源：评论

学校读者我要写书评

暂无评论

Extracting Sinusoidal Frequency Modulation signal Signature Using Fourier-Bessel Transform and Short-Time Fraction Fourier Transform

Extracting Sinusoidal Frequency Modulation Signal Signature ...

引用

IEEE International Conference on Information Communication and signal processing (ICICSP)

作者： Xiaomeng Ma Meiguo Gao School of Information and Communication Beijing Institute of Technology Beijing China

A method using Fourier-Bessel transform (FBT) and short-time fraction Fourier transform (ST-FRFT), for accurate and robust parameter estimate and separation of sinusoidally frequency modulated (SFM) nonstationary signal, is presented. The rotation micro-motion coupling with the rigid body's motion provides an approximately sinusoidal phase modulation, which is known as the micro-Doppler (m-D) signal. According to priori research, the SFM signal shows the nearly linear frequency modulation characteristic in the local part and sparsity features in fraction domain. Firstly, in order to extract the slope of nearly FM and the rotation target's angular frequency, here using FBT method can decompose the rotation target's echo into stationary and nonstationary component. Then, using ST-FRFT method separate the target's signal into the fraction domain where through peak search to catch the SFM signal's slope and rotation angular frequency. This function shows enormous potential in solving the problem that nonstationary component always can't be extracted because of weak echo amplitude. The rotation angle frequency is an important characteristic of the rotating target, which shows great significance to the classification of low-altitude and low-speed small targets. This proposed method has been demonstrated by the air target's simulation data which shows effective results of estimating these parameters.

关键词：

来源：评论

学校读者我要写书评

暂无评论

雷达信号与遥感地图融合的深度学习低慢小目标检测算法

引用

信号处理 2024年第1期40卷 82-93页

作者：高梅国林升泰北京理工大学信息与电子学院北京100081

雷达复杂环境低慢小目标检测是一项具有挑战性的任务,而利用深度学习以及数据特征融合等方法是解决这一难题的有效手段。本文在雷达地图融合检测网络(Radar Map fusion Detection Network,RMDN)的基础上进行了优化,主要优化方向为将雷... 详细信息

雷达复杂环境低慢小目标检测是一项具有挑战性的任务,而利用深度学习以及数据特征融合等方法是解决这一难题的有效手段。本文在雷达地图融合检测网络(Radar Map fusion Detection Network,RMDN)的基础上进行了优化,主要优化方向为将雷达与地图信息在检测过程中进行重要性程度区分,具体优化内容为减少地图特征提取模块的网络深度,加入通道注意力机制,让神经网络自主学习雷达信息与地图信息特征的权重,使神经网能够更好地利用地图信息对雷达目标进行辅助检测。在此优化基础上,本文重新设计出了雷达地图融合检测网络RMDN-V2。算法的主要思想为利用卫星遥感地图来提供背景环境信息,作为雷达信号检测的辅助,通过将目标背景中的特征信息融入检测决策中,提高目标检测的准确性和鲁棒性,减少对强杂波和移动物体的干扰敏感性,改善目标检测算法在复杂环境下的表现。最后的无人机雷达实测数据实验结果表明,本文所做的针对性优化是有效的,RMDN-V2的检测性能优于原始的RMDN,同时本文算法检测性能远超传统的雷达检测算法,同时也优于目前主流的一些深度学习雷达目标检测算法。本文为解决当下低慢小目标检测的难题提出了新的算法。

关键词：雷达目标检测深度学习雷达信号和遥感地图融合低慢小目标检测

来源：评论

学校读者我要写书评

暂无评论

A Low Slow small Target Stable Tracking Method for Vehicle-Borne Radar During Movement

A Low Slow Small Target Stable Tracking Method for Vehicle-B...

引用

China International SAR Symposium (CISS)

作者： Zhengkang Feng Jiayi Zhang Panpan Huang Jiahui Xu Junji Chen Zhou Ye Shanghai Aerospace Electronic Technology Institute Shanghai China Aerospace System Engineering Shanghai Shanghai China

When the vehicle is working in motion, firstly, combined with the characteristics of low slow small targets, the problems affecting radar stable tracking are analyzed, including the impact on target detection and target tracking. Then, starting from the radar system software module, the Doppler frequency is compensated in the signal processing software. In the data processing software. the attitude and position of the point are finely compensated. In the resource management software, the predicted attitude is used to compensate the beam, and the shaking state of the vehicle is judged to arrange the beam management. Finally, through the multiple flight tests of a phased array radar on the UAV target, it is shown that the method can effectively complete the function of stable target tracking.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Transformer-in-Transformer Network with A Complex-Valued Mixup Module for SAR Automatic Target Recognition

Transformer-in-Transformer Network with A Complex-Valued Mix...

引用

signal, Information and data processing (ICSIDP), IEEE International Conference on

作者： Zhenghua Ze Jianghong Han Qiyuan Lab Beijing China

ISBN: (数字)9798331515669

ISBN: (纸本)9798331515676

Convolutional neural network (CNN) has been demonstrated an effective approach for synthetic aperture radar (SAR) automatic target recognition (ATR) in recent years. However, CNNs usually extract characteristics through local receptive fields, resulting in the loss of global feature information of targets. Moreover, satisfactory target recognition performance using CNN generally depends on sufficient labeled sample images, which are difficult to obtain in SAR ATR conditions. To address these issues, we propose a Complex-Valued Mixup (CVM) module- aided Transformer-in-Transformer (TNT) network (CVM-TNT) for SAR ATR. The CVM module generates new samples through linear interpolation and utilizes both amplitude and phase information to extract richer electromagnetic scattering feature. The TNT network takes the advantages of global and local features to compensate the deficiency of traditional Transformer network that only extracts global features. Firstly, we divide the input images including amplitude and phase into several local patches as "visual sentences". Then, we use the outer conventional Transformer block to calculate global features and their global attentions of visual sentences. Subsequently, dividing the local patches into sub-patches as "visual words" and embedding an inner sub-Transformer block into the network excavates the local phase and amplitude features of visual words. Finally, both word-level and sentence-level features will be combined to boost the representation capability. Experiments implemented on the moving and stationary target acquisition and recognition (MSTAR) dataset have shown that the proposed CVM-TNT model outperforms the state-of-the-art (SOTA) methods in the case of small datasets.

关键词： Visualization Interpolation Target recognition Transformers Feature extraction Excavation Radar polarimetry Convolutional neural networks data mining Synthetic aperture radar

来源：评论

学校读者我要写书评

暂无评论

单帧红外弱小目标检测技术研究现状与展望

引用

信号处理 2024年第5期40卷 887-906页

作者：杨德贵韩同欢胡亮白正阳中南大学自动化学院湖南长沙410083 光电智能测控湖南省重点实验室湖南长沙410083

红外探测技术具有探测距离远、抗干扰能力强、隐蔽性强和全天候等优势在天基预警、末敏弹制导等领域得到了广泛应用。通过红外成像技术能够得到目标图像从而对目标进行预警、识别和跟踪。在实际场景中,目标图像往往所占像素比例小,信号... 详细信息

红外探测技术具有探测距离远、抗干扰能力强、隐蔽性强和全天候等优势在天基预警、末敏弹制导等领域得到了广泛应用。通过红外成像技术能够得到目标图像从而对目标进行预警、识别和跟踪。在实际场景中,目标图像往往所占像素比例小,信号强度低,容易湮没在背景图像中;背景图像变化剧烈,存在较强的结构信息、边缘和噪声,红外图像信噪比低,目标检测难度较大,一直是目标检测领域的研究难点和热门话题。为提高红外图像弱小目标检测能力,大量的弱小目标检测算法被提出。现有的主流的检测手段根据图像数据检测方式分为单帧检测和多帧检测两大类,多帧算法依赖大量的图像数据支撑,响应周期长,应用潜力低,而单帧检测算法凭借复杂度低、时效性强、便于硬件实现等特点,被广泛应用在高速运动目标检测、预警等领域。因此本文以单帧红外弱小目标检测算法为主体,从技术方向入手,着重阐述了基于滤波、基于对比度与显著性分析、基于数据优化和深度学习四类单帧弱小目标检测算法的原理与近年来的典型应用,通过仿真试验和算法复杂度对比了不同算法的性能、优势和不足,总结了弱小目标检测算法的研究现状并对本领域的发展趋势进行了展望。本文的工作能帮助读者快速了解本领域的研究现状,为研究人员提供参考。

关键词：红外检测单帧滤波算法视觉显著性数据优化

来源：评论

学校读者我要写书评

暂无评论

Search for supersymmetric sleptons and charginos with the ATLAS detector

引用

JOURNAL OF INSTRUMENTATION 2023年第8期18卷 C08009页

作者： Ballabene, Eric Univ Bologna Dept Phys & Astron Augusto Righi Viale Berti Pichat 6-2 Bologna Italy Ist Nazl Fis Nucl Sez Bologna Viale Berti Pichat 6-2 Bologna Italy

A new search for the electroweak production of supersymmetric particles decaying into two leptons with missing transverse momentum is presented. Two simplified models are considered: direct pair production of sleptons decaying into the lightest neutralinos through leptons of the Standard Model (SM) and direct pair production of the lightest charginos decaying into the lightest neutralinos through bosons of the SM. The analysis targets phase space regions where the difference in mass between the slepton or the lightest chargino and the lightest neutralino is close to or below the mass of the boson. Such regions with compressed mass spectra have not been covered by any searches conducted so far due to the low cross-section of the supersymmetric signal. Therefore, improved analysis strategies are crucial to separate the supersymmetric signal from the SM backgrounds. A search for an excess of same-flavour lepton pairs in opposite-sign lepton events is made in the direct slepton pair production analysis while a machine learning approach using boosted decision trees is exploited in the chargino pair production analysis and considering both the same-flavour and different-flavour channels. No significant excesses over the expected background are observed using proton-proton collisions data collected by the ATLAS experiment at root = 13 TeV and exclusion limits at 95% confidence level are set for each considered model. Exclusion limits are also set for selectrons and smuons separately and portions of the region excluded by the search of smuons pair production are expected to be compatible with the - 2 anomaly for small tan values.

关键词： Analysis and statistical methods data processing methods

来源：评论

学校读者我要写书评

暂无评论

MKD-YOLO: Multi-Scale and Knowledge-Distilling YOLO for Efficient PPE Compliance Detection

MKD-YOLO: Multi-Scale and Knowledge-Distilling YOLO for Effi...

引用

International Conference on Acoustics, Speech, and signal processing (ICASSP)

作者： Juntao Zan Yang Fang Qilie Liu Uswah Khairuddin Yan Li Kaiwei Sun School of Communications and Information Engineering Chongqing University of Posts and Telecommunications Chongqing China Chongqing Key Laboratory of Image Cognition Chongqing University of Posts and Telecommunications Chongqing China Malaysia-Japan International Institute of Technology University of Technology Malaysia Kuala Lumpur Malaysia Department of Electrical and Computer Engineering Inha University Incheon South Korea Key Laboratory of Data Engineering and Visual Computing Chongqing University of Posts and Telecommunications Chongqing China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

YOLO-based models are widely used for personal protective equipment (PPE) compliance detection due to their excellent detection performance and efficiency. However, most YOLO models are not competent for detection tasks in complex industrial scenarios such as remote surveillance and extremely small targets. In addition, there is a lack of effective model lightweighting and knowledge transfer approaches for industrial deployment. To this end, this paper proposes a Multi-scale and Knowledge-Distilling YOLO (MKD-YOLO) based on YOLOv8n for efficient PPE compliance detection. Specifically, in backbone stage, we design an Efficient Multi-Scale Enhanced Convolution (C2f-EMSEC) module and Large Spatial Pyramid Pooling-Fast (LSPPF) module for multi-scale and global-contextual feature learning as well as reducing model complexity. Then, in neck stage, a refined Bidirectional feature Pyramid Network (BPNet) is designated to capture fine-grained details for extremely small object detection. Moreover, we apply channel-wise knowledge distillation to facilitate model lightweighting and domain-specific knowledge transfer learning. Experiments on our proposed dataset and public datasets show that the proposed MKD-YOLO achieves a new state-of-the-art (SOTA) detection performance and efficiency for practical PPE compliance detection tasks. Codes and the dataset are available at https://***/z1Zjt/MKD-YOLO.

关键词： Personal protective equipment YOLO Representation learning Convolution Surveillance Speech enhancement Feature extraction Neck Kernel Knowledge transfer

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：