检索结果-内蒙古大学图书馆

20th International Symposium on Medical Information Processing and Analysis

作者： Navarrete, Sebastian Mangin, Jean-Francois Hernandez, Cecilia Guevara, Pamela Univ Concepcion Fac Engn Concepcion Chile Univ Paris Saclay Baobab Neurospin CEACNRS Gif Sur Yvette France

ISBN: (纸本)9798331528669;9798331528652

Whole brain tractography data contain a large number of streamlines that require algorithms such as clustering to group the data into smaller sets for visualization and analysis. We present a deep-learning clustering algorithm based on the latent space of a variational autoencoder trained on the direct and flipped versions of the streamlines from 10 tractograms (17,294,232 streamlines in total). The model takes advantage of the low-dimensional representation of the data in latent space to apply an HDBSCAN clustering algorithm to perform automatic and fast clustering of the tractography datasets. The proposed method was evaluated in terms of segmentation quality using the Davies-Bouldin index (DB) and execution time against two other state-of-the-art methods, QuickBundles (QB) and FFClust. The results show that the proposed method has the best performance in terms of the DB index, closely followed by QB, and is the second fastest method, only slightly surpassed by FFClust. In addition, the proposed method allows for obtaining meaningful large clusters because it uses metrics based on the density of the groups instead of a distance threshold as the other methods.

关键词： clustering algorithm White Matter Tractography Deep-learning

来源：评论

学校读者我要写书评

暂无评论

Outlier detection method based on improved DPC algorithm and centrifugal factor

引用

INFORMATION SCIENCES 2024年 682卷

作者： Xia, Hao Zhou, Yu Li, Jiguang Yue, Xuezhen Li, Jichun North China Univ Water Resources & Elect Power Sch Elect Engn Zhengzhou 450045 Peoples R China Univ Salford Sch Sci Engn & Environm Salford M5 4NT England Newcastle Univ Sch Comp Newcastle Upon Tyne NE4 5TG England

Outlier detection aims to identify data anomalies exhibiting significant deviations from normal patterns. However, existing outlier detection methods based on k-nearest neighbors often struggle with challenges such as increasing outlier counts and cluster formation issues. Additionally, selecting appropriate nearest-neighbor parameters presents a significant challenge, as researchers commonly evaluate detection accuracy across various k values. To enhance the accuracy and robustness of outlier detection, in this paper we propose an outlier detection method based on the improved DPC algorithm and centrifugal factor. Initially, we leverage k-nearest neighbors, kreciprocal nearest neighbors, and Gaussian kernel function to determine the local density of samples, particularly addressing scenarios where the DPC algorithm struggles to identify cluster centers in sparse clusters. Subsequently, to reduce the DPC algorithm's computational complexity, we screen the samples based on mutual nearest neighbor counts and select cluster centers accordingly. Non-central points are then distributed using k-nearest neighbors, k-reciprocal nearest neighbors, and reverse k-nearest neighbors. The centrifugal factor, whose magnitude reflects the outlier degree of samples, is then computed by calculating the ratio of the local kernel density at the cluster center to that of samples. Finally, we propose a method for choosing the nearest neighbor parameter, k. To comprehensively evaluate the outlier detection performance of the proposed algorithm, we conduct experiments on 12 complex synthetic datasets and 25 public real-world datasets, comparing the results with 12 state-of-the-art outlier detection methods.

关键词： Outlier detection clustering algorithm Centrifugal factor k -nearest neighbor Local density Local kernel density

来源：评论

学校读者我要写书评

暂无评论

POSTERIOR BELIEF clustering algorithm FOR ENERGY-EFFICIENT TRACKING IN WIRELESS SENSOR NETWORKS

引用

INTERNATIONAL JOURNAL ON SMART SENSING AND INTELLIGENT SYSTEMS 2014年第3期7卷 925-941页

作者： Wu, Bo Feng, Yanpeng Zheng, Hongyan Shenzhen Polytech Educ Technol & Informat Ctr Shenzhen 518055 Guangdong Peoples R China

In this paper, we propose a novel posterior belief clustering (PBC) algorithm to solve the tradeoff between target tracking performance and sensors energy consumption in wireless sensor networks. We model the target tracking under dynamic uncertain environment using partially observable Markov decision processes (POMDPs), and transform the optimization of the tradeoff between tracking performance and energy consumption into yielding the optimal value function of POMDPs. We analyze the error of a class of continuous posterior beliefs by Kullback-Leibler (KL) divergence, and cluster these posterior beliefs into one based on the error of KL divergence. So, we calculate the posterior reward value only once for each cluster to eliminate repeated computation. The numerical results show that the proposed algorithm has its effectiveness in optimizing the tradeoff between tracking performance and energy consumption.

关键词： Partially observable Markov decision processes wireless sensor networks target tracking energy consumption posterior belief clustering algorithm

来源：评论

学校读者我要写书评

暂无评论

Safety risk assessment for automotive battery pack based on deviation and outlier analysis of voltage inconsistency

引用

JOURNAL OF CLEANER PRODUCTION 2024年 466卷

作者： Huang, Peifeng Liu, Shoutong Ren, Yinghui He, Yanyun Chao, Peipei Feng, Renlang Liu, Chuan Li, Zhen Bai, Zhonghao Hunan Univ State Key Lab Adv Design & Mfg Technol Vehicle Changsha Peoples R China China Automot Engn Res Inst Co Ltd Chongqing Peoples R China

Safety risk assessment is essential for evaluating the health status and averting sudden battery failures in electric vehicles. This study introduces a novel safety risk assessment approach for battery systems, addressing both cell and pack levels with three key indexes. The core of the assessment lies in representing the relative deviation of cell voltages through scatter diagrams across various stages of service life. Specifically, the study quantifies voltage deviations and deviation angles within different state of charge intervals to gauge safety risks at the cell level. Leveraging clustering algorithms aids in identifying outlier values. Furthermore, the dispersion of scatter points is utilized to assess safety risks at the pack level. Validation of this proposed model is conducted through cycling tests on battery modules with a deformed cell, demonstrating its efficacy in capturing inconsistent features of mechanical deformation abuse across the three indexes, thereby triggering alarms during operation. Moreover, the method is applied to assess the safety risks of five hazardous and accident-prone vehicles, enabling comprehensive evaluation of potential faulty cells and safety risks at the pack level. This proposed approach offers a fresh perspective on the comprehensive safety risk assessment of battery systems.

关键词： Voltage inconsistency Battery pack Safety risk assessment clustering algorithm

来源：评论

学校读者我要写书评

暂无评论

A fast and lightweight train image fault detection model based on convolutional neural networks

引用

IMAGE AND VISION COMPUTING 2025年 154卷

作者： Zhang, Longxin Zeng, Wenliang Zhou, Peng Deng, Xiaojun Wu, Jiayu Wen, Hong Hunan Univ Technol Sch Comp Sci Zhuzhou 412007 Peoples R China

Trains playa vital role in the life of residents. Fault detection of trains is essential to ensuring their safe operation. Aiming at the problems of many parameters, slow detection speed, and low detection accuracy of the current train image fault detection model, a fast and lightweight train image fault detection model using convolutional neural network (FL-TINet) is proposed in this study. First, the joint depthwise separable convolution and divided-channel convolution strategy are applied to the feature extraction network in FL-TINet to reduce the number of parameters and computation amount in the backbone network, thereby increasing the detection speed. Second, a mixed attention mechanism is designed to make FL-TINet focus on key features. Finally, an improved discrete K-means clustering algorithm is designed to set the anchor boxes so that the anchor box can cover the object better, thereby improving the detection accuracy. Experimental results on PASCAL 2012 and train datasets show that FL-TINet can detect faults at 119 frames per second. Compared with the state-of-the-art CenterNet, RetinaNet, SSD, Faster R-CNN, MobileNet, YOLOv3, YOLOv4, YOLOv7Tiny, YOLOv8_n and YOLOX-Tiny models, FL-TINet's detection speed is increased by 96.37% on average, and it has higher detection accuracy and fewer parameters. The robustness test shows that FL-TINet can resist noise and illumination changes well.

关键词： Attention mechanism clustering algorithm Convolutional neural network Depthwise separable convolution Fault detection

来源：评论

学校读者我要写书评

暂无评论

A data-driven exploration of the complexity of the phosphorus-centered pnictogen bonds

引用

MOLECULAR PHYSICS 2025年第3期123卷

作者： Liu, Fang Wang, Haiyan Du, Likai Bohai Univ Coll Informat Sci & Technol Jinzhou Liaoning Peoples R China Shandong Univ Aeronaut Sch Informat Engn Jinan Shandong Peoples R China Jinzhou Med Univ Sch Biol Informat Engn Jinzhou Liaoning Peoples R China

Phosphorus-centered pnictogen bonds, similar to well-studied halogen bonds, play a vital role in molecular recognition and assembly. This study aims to explain how pnictogen bonds interact with other noncovalent interactions. A model system, phosphines (PH3), which is also present in phosphorus cycles and reduces planetary atmospheres, is examined. The investigation of the phosphorus trifluoride (PF3) molecule explores its substituent effects. To generate a broad spectrum of molecular configurations, a specially tailored protocol for sampling and optimisation was implemented. The configurations were refined to uncover primary interaction patterns, and a clustering algorithm revealed unique interaction patterns. This report presents the energy stability and distribution of all the clusters. Our findings verify the prevalent presence of pnictogen bonds, identified by their geometric characteristics and co-occurrence with hydrogen bonds, which show an almost linear correlation. Another significant discovery is the correlation between various energy decomposition elements, especially regarding the electrostatic energy and overall binding energy. These results are anticipated to significantly contribute towards our comprehension of non-covalent interactions among phosphorus-containing molecules and the formulation of empirical models having physical interpretations.

关键词： Data-driven phosphorus-centered pnictogen bonds clustering algorithm linear correlation energy decomposition

来源：评论

学校读者我要写书评

暂无评论

Load Balanced clustering Based on Imperialist Competitive algorithm in Wireless Sensor Networks

引用

WIRELESS PERSONAL COMMUNICATIONS 2020年第1期112卷 371-385页

作者： Dehestani, Fahimeh Jabraeil Jamali, Mohammad Ali Islamic Azad Univ Shabestar Branch Dept Comp Engn Shabestar Iran

Energy consumption is one of the most serious issues in designing Wireless Sensor Networks (WSNs) for maximizing its lifetime and stability. clustering is considered as one of the topology control methods for maintaining the stability of WSNs which can significantly reduce energy consumption in WSNs. However, using different methods for the selection of cluster head is an important challenge in this domain of research. Load balanced clustering is known as an NP-hard problem for a WSN along with unequal load for sensor nodes. The Imperialist Competitive algorithm (ICA) is regarded as an evolutionary method which can be used for finding a quick and efficient solution to such problems. In this paper, a clustering method with an evolutionary approach is introduced which investigates the issues of load balance and energy consumption of WSNs in the equal and unequal load modes so as to select optimal cluster heads. Simulation of the proposed method, carried out via NS2, indicated that it improves the criteria of energy consumption, the number of active sensor nodes and execution time.

关键词： Wireless sensor network clustering algorithm Imperialist competitive algorithm Load balancing

来源：评论

学校读者我要写书评

暂无评论

Multi-satellite flyby trajectory planning for uncertain target search

引用

ACTA ASTRONAUTICA 2025年 228卷 58-68页

作者： Wang, Ke-Mao Zhang, Jin Sun, Xue Gao, Lin-Tao Natl Univ Def Technol Coll Aerosp Sci & Technol Changsha 410073 Peoples R China Hunan Key Lab Intelligent Planning & Simulat Aeros Changsha 410073 Peoples R China Shanghai Inst Aerosp Syst Engn Shanghai 201109 Peoples R China

With the increase of space targets, it is significant to conduct rapid flyby monitoring of satellite clusters and identify Special Targets (STs), which have the distribution uncertainty among clusters. Moreover, there are observation uncertainties during the flybys of space targets. In order to complete the flyby and observation mission more efficiently, the trajectory design of Observation Satellites (OSs) is constructed as a Mixed-Integer Nonlinear Programming (MINLP) model. A two-stage optimization method is proposed to solve the MultiSatellite Flyby (MSF) planning problem. For the first stage, an improved K-means algorithm is used to determine the target assignment groups and the orbital planes of OSs. For the second stage, the mixed integer Differential Evolution (DE) algorithm is utilized to optimize the flyby sequences and maneuver impulses, and to determine the initial orbit parameters of OSs, thereby obtaining the optimal probability of successful search. The simulation results show that the proposed method is effective and can obtain the optimal MSF trajectory under the velocity increment and time constraints, and that the increase of allowed maximum velocity increment could expand the mission benefits. Furthermore, compared with non-revisitable mode which is commonly used in multi-satellite visit problems, the revisitable mode shows the considerable advantage.

关键词： Multi-satellite flyby Uncertain optimization Two-stage planning clustering algorithm Probability of success

来源：评论

学校读者我要写书评

暂无评论

Research on clustering of non-uniformly distributed point clouds in road scenes 24

Research on clustering of non-uniformly distributed point cl...

引用

16th International Conference on Machine Learning and Computing (ICMLC)

作者： Li Lin Zhao Haili Hui Zhen Mi Rongxin Li Weikuang Chen Zixuan Coordinat Ctr China Natl Comp Network Emergency Response Tech Team Beijing Peoples R China Changan Commun Technol Co LTD Beijing Peoples R China

ISBN: (纸本)9798400709234

This paper presents a clustering algorithm for non-uniformly distributed point clouds in road scenes, which is used to alleviate the performance effect of classical Density-Based Spatial clustering of Applications with Noise (DBSCAN) algorithm in non-uniformly distributed scenes. Because of the limitations of the DBSCAN algorithm, it's difficult to show good results in the space where the parameters aren't convergent. So we proposes a solution, which calculates the node density, average density, density variation coefficient and other parameters of each point which is divided the space into several small spaces with uniform density. Through this method we can achieve better clustering effect in small spaces. Finally, we analyze the proposed solution through Python code on some KITTI data sets. The analysis results show that our proposed scheme can effectively improve the performance of classical DBSCAN algorithm in non-density uniform space.

关键词： DBSCAN clustering algorithm point cloud road scenes

来源：评论

学校读者我要写书评

暂无评论

Multi-dimensional DBSCAN clustering Approach for Millimeter Wave Radar Point Cloud Data 16

Multi-dimensional DBSCAN Clustering Approach for Millimeter ...

引用

16th International Conference on Microwave and Millimeter Wave Technology (ICMMT)

作者： Li, Hanchen Fu, Chaowei Xi, Guangrong Wang, Qingmian Peng, Hongli Shanghai Jiao Tong Univ State Key Lab Radio Frequency Heterogeneous Integ Shanghai Peoples R China Shanghai Radio Equipment Res Inst Shanghai Peoples R China Shanghai Shentian Ind Co Ltd Shanghai Peoples R China

ISBN: (纸本)9798350389968

Addressing the poor resolution challenge of traditional Density-Based Spatial clustering of Applications with Noise (DBSCAN) algorithms in differentiating objects in close proximity, this paper introduces a multi-dimensional DBSCAN (MD-DBSCAN) clustering approach, considering the Doppler characteristics of millimeter wave (mmW) radar. Our approach is based on the idea that the whole clustering procedure is divided into two stages, i.e., velocity clustering and spatial one. Specifically, for each group, the clusters of varying radial velocities are firstly segmented, then followed by spatial dimension clustering. Experimental results demonstrate a 14.2% improvement in number of point clouds with this approach over traditional methods, underscoring its effectiveness in more accurately distinguishing closely situated objects in autonomous driving scenarios.

关键词： mmW Radar DBSCAN clustering algorithm

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：