检索结果-内蒙古大学图书馆

IEEE Information Technology, Networking, Electronic and Automation Control Conference

作者： Huibai Wang Ji Han Information Institute North China University Of Technology Beijing China

In the field of military equipment target recognition, in order to solve the problems of small target detection, such as easy loss and slow detection speed, this paper proposes a target recognition and detection algorithm YOLO-M, which not only keeps high accuracy, but also reduces the number of parameters. Firstly, C3CMix module is used to replace part of the structure of head in YOLO v5, which reduces the number of parameters and effectively reduces the model size. Then, the detection accuracy is further improved by replacing the activation function in the structure. The experimental results show that the average accuracy of YOLO-M is 95.2%, the parameters are reduced by 18.8%, and the calculation is reduced by 14.5%. Therefore, the improved algorithm in this paper is more suitable for deployment in the field of military equipment target recognition.

关键词： Automation Target recognition Object detection Real-time systems detection algorithms Military equipment

来源：评论

学校读者我要写书评

暂无评论

An Optimal Strategy of Sorting Neighborhood Method 6

An Optimal Strategy of Sorting Neighborhood Method

引用

6th International Symposium on Computer and Information Processing Technology, ISCIPT 2021

作者： Shijie, Zhou Yuanshen, Lou Hohai University Department of Computer and Information Nanjing China

ISBN: (纸本)9781665441377

Duplicate item detection is the process of identifying multiple representations of the same real entity in a data set. Due to the huge amount of data, and the same real entity may have different representations, this task is very difficult. For this task, in recent years, researchers have proposed the Sorted Neighborhood Method (SNM) and Repeat Counting Strategy (DCS), reduce the number of comparisons of record pairs during duplicate entries while maintaining high matching quality. Inspired by these two algorithms, we propose an improved SNM algorithm. The algorithm uses an objective method to select sort keywords, and optimizes the strategy of duplicate item comparison. At the same time, it uses a variable window method to avoid the problems caused by a fixed window in the detection process. © 2021 IEEE.

关键词： Machine learning algorithms Distributed databases Machine learning Task analysis detection algorithms Sorting

来源：评论

学校读者我要写书评

暂无评论

Real-Time Adaptive Threshold Adjustment for Lane detection Application under Different Lighting Conditions using Model-Free Control

引用

IFAC-PapersOnLine 2021年第20期54卷 147-152页

作者： Yujing Zhou Zejiang Wang Junmin Wang Walker Department of Mechanical Engineering The University of Texas at Austin Austin TX 78712 USA

Lane detection and tracking technologies are critical to automated vehicles and the lane departure warning (LDW) feature in advanced driver assistance systems (ADAS). The color difference between lane marks and their background is an important source for drivers to detect lanes. However, most existing lane detection techniques neglect the importance of color features contained within lane marks because the color is subject to lighting variations. The hue, saturation, and value (HSV) color space is commonly used for color identification, as it separates the color, saturation, and brightness information into the H, S, and V channels, respectively. Colors can be identified with specified thresholds in H, S, and V channels, but fixed thresholds can hardly treat the disturbances caused by variation of lighting conditions. This paper leverages the color information of lane marks to perform the lane detection task by adaptively changing the thresholds in the HSV space to handle lighting condition variations. In doing so, the vehicle is assumed to follow a straight lane with a reference steering angle of zero degree under varying lighting conditions. It is hard to model the relationship between the thresholds in the HSV space and the vehicle steering angle in automated driving, so we adopt the model-free control (MFC), which is a data-driven approach and does not require an explicit model for the system. Simulink-QUARC joint offline experiments demonstrate the effectiveness of the proposed real-time HSV threshold adjustment method for better lane detections.

关键词： Autonomous vehicles detection algorithms Threshold selection Adaptive systems Self-adjusting systems Model-free control

来源：评论

学校读者我要写书评

暂无评论

Tobacco Maturity Recognition Based on an Improved YOLOv8 Algorithm

Tobacco Maturity Recognition Based on an Improved YOLOv8 Alg...

引用

Smart and Healthy Cities (ISHC), International Symposium on

作者： Xin Zan Yang Chen Ziran Li Tao Wang Baolin Zhang Zhao Zhang Lingmei Li Tao Bai School of Automation Xi'an Jiaotong University Xi'an China Qujing Tobacco Company of Yunnan Province Qujing China

ISBN: (数字)9798350360745

ISBN: (纸本)9798350371536

Tobacco maturity is an important factor in determining tobacco baking, and the existing means of tobacco maturity detection are usually based on manual detection, lacking objective and uniform detection standards. In this paper, a new object detection algorithm GSSS-YOLO is proposed based on the YOLOv8 algorithm. The GSSS-YOLO introduces the CPCA attention mechanism to raise the precision of the detection, and the loss function SIoU is used to speed up the model's convergence speed. Finally, the tobacco maturity recognition model is trained using nearly 3700 tobacco leaves collected in a real environment, and the results of experiment indicate that the proposed detection algorithm has good detection accuracy and detection efficiency. The proposed GSSS-YOLO algorithm can be deployed on an embedded platform to realize on-line tobacco maturity recognition, which is very important for improving the efficiency of tobacco harvesting.

关键词： Convolution Urban areas Object detection Manuals Prediction algorithms Data mining detection algorithms

来源：评论

学校读者我要写书评

暂无评论

Joint Exploitation of Hemodynamic and Electrocardiographic Signals by Hidden Markov Models for Heartbeat detection 8th

Joint Exploitation of Hemodynamic and Electrocardiographic S...

引用

8th Latin American Conference on Biomedical Engineering (CLAIB) / 42nd National Conference on Biomedical Engineering (CNIB)

作者： Monroy, Nelson F. Altuve, Miguel Pontifical Bolivarian Univ Fac Syst Engn & Informat Bucaramanga Colombia Pontifical Bolivarian Univ Fac Elect & Elect Engn Bucaramanga Colombia

ISBN: (纸本)9783030306489;9783030306472

The detection of heartbeats is a fundamental task in the diagnosis of cardiovascular diseases. In a clinical setting, heartbeats are commonly detected from the electrocardiographic (ECG) signal, but it can also be estimated from hemodynamic signals, such as arterial blood pressure (ABP), pulmonary arterial pressure (PAP) and central venous pressure (CVP). In this paper, we conceived three HMM-based detectors that jointly exploit the information from ECG and ABP (D-ECG&ABP), ECG and PAP (D-ECG&PAP), and ECG and CVP (D-ECG&CVP). The HMM-based detectors are based on the comparison of the difference of log-likelihoods of observation of two competing models that learned the dynamics of the presence and absence of a heartbeat, respectively. The detection performances of the bivariate centralized detectors were similar, but the detector that simultaneously considers ECG and ABP provide the best result (sensitivity = 98.91% and positive predictivity = 99.11%). These results agree when a univariate HMM-based detector only considers the ECG signal in the decision process. The centralized bivariate detection approach proposed in this paper make uses of redundant information in the decision process and is particularly useful when one of the signals is too noisy or missing.

关键词： Hidden Markov model Heartbeat detection algorithms Electrocardiogram Blood pressure

来源：评论

学校读者我要写书评

暂无评论

Throwing Event detection using Acceleration Magnitude collected with Wrist-Worn Sensors

Throwing Event Detection using Acceleration Magnitude collec...

引用

Annual IEEE SoutheastCon Conference

作者： Schweiger, Daniel Critcher, Shelby Freeborn, Todd J. Hibberd, Elizabeth Univ Alabama Dept Elect Comp Engn Tuscaloosa AL 35401 USA Univ Alabama Dept Hlth Sci Tuscaloosa AL USA

ISBN: (数字)9781728168616

ISBN: (纸本)9781728168616

Recent advancements to reduce rates of injury and pain related to pitching include the release of "Pitch Smart" guidelines that specify the number of pitches per day and required rest recommendations for youth and high school baseball players. However, monitoring pitch counts for adherence to these guidelines requires significant human resources for tracking individuals throughout training and games. The availability of low-cost tools that can monitor and report pitch count and cumulative loads have the potential to improve adherence to pitch count guidelines, reduce the burden of monitoring on coaching staff, and reduce the chances of injury. In this work, acceleration datasets collected using a wrist-worn sensor during throwing events are explored as a method to identify throwing events. The detection algorithm, using a 18g threshold and 3 second window, identified 100% of the 161 pitching events from 8 participants, with only 1 false-positive. This supports that using acceleration datasets from wrist-worn sensors may be an effective method to identify and count throwing & pitching events.

关键词： Event detection Humping Pitching PITCHES Human Resource Management Medication Adherence Mucilages, Adhesive pitch acceleration detection algorithms Sensors

来源：评论

学校读者我要写书评

暂无评论

A Current-Based Approach for Short-Circuit Fault Diagnosis in Closed-Loop Current Source Inverter

引用

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS 2020年第9期67卷 7941-7950页

作者： Guo, Xiaoqiang Sui, Sen Wang, Baocheng Zhang, Wenlin Yanshan Univ Dept Elect Engn Qinhuangdao 066004 Hebei Peoples R China

In this article, a method based on output currents is proposed to diagnose short-circuit faults of single or double switches in current source converter. The method only needs the feedback currents and their corresponding reference signals that can be obtained from the main control system. This diagnosis algorithm needs to calculate the average values of the feedback currents and the average absolute values of the current reference signals in one power frequency cycle. Diagnosis variables can be obtained from the ratio of the two values. Then, three detection thresholds are set to judge and locate the short-circuit power switches. The method can be implemented in the main control program directly, without additional voltage sensors and other auxiliary detection devices. Meanwhile, the method is simple and easy to implement in practice. The effectiveness of proposed algorithm is validated with experiments.

关键词： Circuit faults Inverters Fault detection Fault diagnosis Sensors detection algorithms Current-based method current closed-loop control current source converter faults location short-circuit faults detection single or double switch faults diagnosis

来源：评论

学校读者我要写书评

暂无评论

Efficient Inference in Stochastic Block Models With Vertex Labels

引用

IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING 2020年第3期7卷 1215-1225页

作者： Stegehuis, Clara Massoulie, Laurent Eindhoven Univ Technol NL-5600 MB Eindhoven Netherlands Microsoft Res INRIA Joint Ctr F-54600 Villers Les Nancy France

We study the stochastic block model with two communities where vertices contain side information in the form of a vertex label. These vertex labels may have arbitrary label distributions, depending on the community memberships. We analyze a version of the popular belief propagation (BP) algorithm. We show that this algorithm achieves the highest accuracy possible whenever a certain function of the network parameters has a unique fixed point. When this function has multiple fixed points, the BP algorithm may not perform optimally, where we conjecture that a non-polynomial time algorithm may perform better than BP. We show that increasing the information in the vertex labels may reduce the number of fixed points and, hence, lead to optimality of BP.

关键词： Belief propagation Stochastic processes Analytical models detection algorithms Correlation Data mining Social networking (online) Network theory (graphs) clustering algorithms belief propagation

来源：评论

学校读者我要写书评

暂无评论

Making Neighbors Quiet: An Approach to Detect Virtual Resource Contention

引用

IEEE TRANSACTIONS ON SERVICES COMPUTING 2020年第5期13卷 843-856页

作者： Vallone, Joel Birke, Robert Chen, Lydia Y. IBM Res Zurich Cloud & Comp Infrastruct CH-8803 Ruschlikon Switzerland IBM Res Lab Zurich Cloud Server Technol Grp CH-8803 Ruschlikon Switzerland

It is imperative for public cloud providers to guarantee performance targets for tenants' virtual machines (VMs) while respecting strict business confidentiality, e.g., having no information on applications nor their performance. A large body of related work addresses the challenges of detecting performance interferences by leveraging client's quality of service (QoS) metrics, e.g., latency, and additional profiling servers. In this paper, we take the perspective of the cloud provider and propose a general black-box approach that detects different resource contentions by throttling neighboring VMs. Specifically, we design a three-phase detection algorithm that includes: (i) an alarm phase to identify statistical outliers using control charts;(ii) a passive clustering phase to match the current sample to historical behaviors;and (iii) an active throttling phase to discern contentions from application phase changes via throttling. The algorithm is specifically designed for scenarios where multiple co-located VMs request detection analysis simultaneously. We implement and evaluate the proposed three-phase algorithm on four latency sensitive applications, i.e., Wikimedia and three benchmarks from Cloudsuite. Our extensive experimental results show that we can reach an average detection accuracy above 90 percent while limiting the performance degradation experienced by offender workloads to short learning phases.

关键词： Measurement Monitoring Quality of service Cloud computing Servers Detectors detection algorithms Contention detection cloud QoS

来源：评论

学校读者我要写书评

暂无评论

Evaluating Overfit and Underfit in Models of Network Community Structure

引用

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 2020年第9期32卷 1722-1735页

作者： Ghasemian, Amir Hosseinmardi, Homa Clauset, Aaron Univ Colorado Dept Comp Sci Boulder CO 80309 USA Univ Southern Calif Inst Informat Sci Marina Del Rey CA 90292 USA Univ Colorado Boulder CO 80309 USA Univ Colorado Biofrontiers Inst Boulder CO 80305 USA Santa Fe Inst 1399 Hyde Pk Rd Santa Fe NM 87501 USA

A common graph mining task is community detection, which seeks an unsupervised decomposition of a network into groups based on statistical regularities in network connectivity. Although many such algorithms exist, community detection's No Free Lunch theorem implies that no algorithm can be optimal across all inputs. However, little is known in practice about how different algorithms over or underfit to real networks, or how to reliably assess such behavior across algorithms. Here, we present a broad investigation of over and underfitting across 16 state-of-the-art community detection algorithms applied to a novel benchmark corpus of 572 structurally diverse real-world networks. We find that (i) algorithms vary widely in the number and composition of communities they find, given the same input;(ii) algorithms can be clustered into distinct high-level groups based on similarities of their outputs on real-world networks;(iii) algorithmic differences induce wide variation in accuracy on link-based learning tasks;and, (iv) no algorithm is always the best at such tasks across all inputs. Finally, we quantify each algorithm's overall tendency to over or underfit to network data using a theoretically principled diagnostic, and discuss the implications for future advances in community detection.

关键词： Task analysis Clustering algorithms Probabilistic logic Bayes methods detection algorithms Prediction algorithms Partitioning algorithms Community detection model selection overfitting underfitting link prediction link description

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：