检索结果-内蒙古大学图书馆

Deep Representation Learning With part Loss for person Re-Identification

IEEE TRANSACTIONS ON IMAGE pROCESSING 2019年第6期28卷 2860-2871页

作者： Yao, Hantao Zhang, Shiliang Hong, Richang Zhang, Yongdong Xu, Changsheng Tian, Qi Chinese Acad Sci Inst Automat Natl Lab Pattern Recognit Beijing 100190 Peoples R China Peking Univ Sch Elect Engn & Comp Sci Beijing 100871 Peoples R China Univ Technol Dept Comp Sci & Technol Hefei 230009 Anhui Peoples R China Chinese Acad Sci Inst Comp Technol Key Lab Intelligent Informat Proc Beijing 100190 Peoples R China Univ Chinese Acad Sci Dept Artificial Intelligence Beijing 100049 Peoples R China Univ Texas San Antonio Dept Comp Sci San Antonio TX 78249 USA

Learning discriminative representations for unseen person images is critical for person re-identification (ReID). Most of the current approaches learn deep representations in classification tasks, which essentially minimize the empirical classification risk on the training set. As shown in our experiments, such representations easily get over-fitted on a discriminative human body part on the training set. To gain the discriminative power on unseen person images, we propose a deep representation learning procedure named part loss network, to minimize both the empirical classification risk on training person images and the representation learning risk on unseen person images. The representation learning risk is evaluated by the proposed part loss, which automatically detects human body parts and computes the person classification loss on each part separately. Compared with traditional global classification loss, simultaneously considering part loss enforces the deep network to learn representations for different body parts and gain the discriminative power on unseen persons. Experimental results on three person ReID datasets, i.e., Market1501, CUHK03, and VIpeR, show that our representation outperforms existing deep representations.

关键词： person re-identification representation learning part lass networks convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Deep Multi-View Feature Learning for EEG-Based Epileptic Seizure Detection

引用

IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING 2019年第10期27卷 1962-1972页

作者： Tian, Xiaobin Deng, Zhaohong Ying, Wenhao Choi, Kup-Sze Wu, Dongrui Qin, Bin Wang, Jun Shen, Hongbin Wang, Shitong Jiangnan Univ Sch Digital Media Wuxi 214122 Jiangsu Peoples R China Jiangsu Key Lab Digital Design & Software Technol Wuxi 214122 Jiangsu Peoples R China Changshu Inst Technol Sch Comp Sci & Engn Changshu 215500 Jiangsu Peoples R China Hong Kong Polytech Univ Ctr Smart Hlth Hong Kong Peoples R China Huazhong Univ Sci & Technol Sch Automat Minist Educ Image Proc & Intelligent Control Key Lab Wuhan 430074 Hubei Peoples R China Shanghai Jiao Tong Univ Inst Image Proc & Pattern Recognit Shanghai 200240 Peoples R China Minist Educ China Key Lab Syst Control & Informat Proc Shanghai 200240 Peoples R China

Epilepsy is a neurological illness caused by abnormal discharge of brain neurons, where epileptic seizure can lead to life-threatening emergencies. By analyzing the encephalogram (EEG) signals of patients with epilepsy, their conditions can be monitored and seizure can be detected and intervened in time. As the identification of effective features in EEG signals is important for accurate seizure detection, this paper proposes a multi-view deep feature extraction method in attempt to achieve this goal. The method first uses fast Fourier transform (FFT) and wavelet packet decomposition (WpD) to construct the initial multi-view features. Convolutional neural network (CNN) is then used to automatically learn deep features from the initial multi-view features, which reduces the dimensionality and obtain the features with better seizure identification ability. Furthermore, the multi-view Takagi-Sugeno-Kang fuzzy system (MV-TSK-FS), an interpretable rule-based classifier, is used to construct a classification model with strong generalizability based on the deep multi-view features obtained. Experimental studies show that the classification accuracy of the proposed multi-view deep feature extraction method is at least 1% higher than that of common feature extraction methods such as principal component analysis (pCA), FFT and WpD. The classification accuracy is also at least 4% higher than the average accuracy achieved with single-view deep features.

关键词： EEG seizure detection multi-view feature extracting deep learning

来源：评论

学校读者我要写书评

暂无评论

A Fault Diagnosis Model for Rotating Machinery Using VWC and MSFLA-SVM Based on Vibration Signal Analysis

引用

SHOCK AND VIBRATION 2019年第1期2019卷

作者： You, Lei Fan, Wenjie Li, Zongwen Liang, Ying Fang, Miao Wang, Jin Chengdu Univ Sch Informat Sci & Engn Chengdu 610106 Sichuan Peoples R China Changsha Univ Sci & Technol Engn Lab Spatial Informat Technol Highway Geol Di Changsha 410114 Hunan Peoples R China Chengdu Aeronaut Polytech Coll Elect & Informat Engn Chengdu 610100 Sichuan Peoples R China Chengdu Univ Key Lab Pattern Recognit & Intelligent Informat P Chengdu 610106 Sichuan Peoples R China

Fault diagnosis of rotating machinery mainly includes fault feature extraction and fault classification. Vibration signal from the operation of machinery usually could help diagnosing the operational state of equipment. Different types of fault usually have different vibrational features, which are actually the basis of fault diagnosis. This paper proposes a novel fault diagnosis model, which extracts features by combining vibration severity, dyadic wavelet energy time-spectrum, and coefficient power spectrum of the maximum wavelet energy level (VWC) at the feature extraction stage. At the stage of fault classification, we design a support vector machine (SVM) based on the modified shuffled frog-leaping algorithm (MSFLA) for the accurate classifying machinery fault method. Specifically, we use the MSFLA method to optimize SVM parameters. MSFLA can avoid getting trapped into local optimum, speeding up convergence, and improving classification accuracy. Finally, we evaluate our model on real rotating machinery platform, which has four different states, i.e., normal state, eccentric axle fault (EAF), bearing pedestal fault (BpF), and sealing ring wear fault (SRWF). As demonstrated by the results, the VWC method is efficient in extracting vibration signal features of rotating machinery. Based on the extracted features, we further compare our classification method with other three fault classification methods, i.e., backpropagation neural network (BpNN), artificial chemical reaction optimization algorithm (ACROA-SVM), and SFLA-SVM. The experiment results show that MSFLA-SVM achieves a much higher fault classification rate than BpNN, ACROA-SVM, and SFLA-SVM.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

The study of traffic flow model based on cellular automata and Naive Bayes

引用

INTERNATIONAL JOURNAL OF MODERN pHYSICS C 2019年第5期30卷

作者： Zhang Hong Wei Juan Gao Xiaoling Hu Jun Chengdu Univ Coll Informat Sci & Engn Chengdu 610106 Sichuan Peoples R China Key Lab Pattern Recognit & Intelligent Informat P Chengdu 610106 Sichuan Peoples R China Chengdu Normal Univ Coll Comp Sci Chengdu 611130 Sichuan Peoples R China Chengdu Normal Univ Inst Intelligent Comp & Informat Technol Chengdu 611130 Sichuan Peoples R China

A new traffic flow model is proposed based on cellular automata and Naive Bayes theory to effectively describe the traffic flow velocity and flow state of the road. On the basis of NaSch model, the safety distance is fully considered in this model, and random deceleration and inflow rules of a vehicle are introduced. At the same time, vehicle acceleration, deceleration and lane change are optimized with Naive Bayes theory. Finally, experimental platform is used for numerical analysis, and the relationship between such parameters as average velocity of traffic flow, maximum velocity of traffic flow, number of inflow vehicles and random deceleration probability, etc. is studied in depth. The results show that the maximum velocity has a great effect on the traffic flow state, and when the vehicle inflow probability is lower, random deceleration probability has less effect on the average velocity and the number of vehicles waiting for inflow;on the contrary, the higher the random deceleration probability is, the more obvious the tendency of road congestion.

关键词： Traffic flow safety distance inflow cellular automata Naive Bayes

来源：评论

学校读者我要写书评

暂无评论

Topological Indices of the pent-Heptagonal Nanosheets VC5C7 and HC5C7

引用

ADVANCES IN MATERIALS SCIENCE AND ENGINEERING 2019年第1期2019卷

作者： Deng, Fei Zhang, Xiujun Alaeiyan, Mehdi Mehboob, Abid Farahani, Mohammad Reza Chengdu Univ Technol Coll Network Secur Chengdu 610059 Sichuan Peoples R China Chengdu Univ Inst Higher Educ Sichuan Prov Key Lab Pattern Recognit & Intelligent Informat P Chengdu 610106 Sichuan Peoples R China IUST Dept Math Tehran 16844 Iran Univ Educ Lahore Vehari Campus Vehari Pakistan

In this paper, we computed the topological indices of pent-heptagonal nanosheet. Formulas for atom-bond connectivity index, fourth atom-bond connectivity index, Randi connectivity index, sum-connectivity index, first Zagreb index, second Zagreb index, augmented Zagreb index, modified Zagreb index, hyper Zagreb index, geometric-arithmetic index, fifth geometric-arithmetic index, Sanskruti index, forgotten index, and harmonic index of pent-heptagonal nanosheet have been derived.

关键词： Nanosheets

来源：评论

学校读者我要写书评

暂无评论

Multi-Criteria Decision Making for pLM Maturity Analysis based on an Integrated Fuzzy AHp and VIKOR Methodology

引用

JOURNAL OF ADVANCED MANUFACTURING SYSTEMS 2018年第2期17卷 155-179页

作者： Yu, Xi Zhang, Haiqing Bouras, Abdelaziz Ouzrout, Yacine Sekhari, Aicha Chengdu Univ Sch Informat Sci & Engn Chengdu Sichuan Peoples R China Inst Higher Educ Sichuan Prov Key Lab Pattern Recognit & Intelligent Informat P Chengdu Sichuan Peoples R China Chengdu Univ Informat Technol Dept Software Engn Chengdu Sichuan Peoples R China Univ Lyon 2 Lumiere DISP Lab 160 Bd Univ F-69676 Bron France Qatar Univ Comp Sci Dept ictQATAR Box 2731 Doha Qatar

A wide range of product lifecycle management (pLM) maturity models are proposed to assess the relative position of companies on their road to complete pLM implementation. However, it is a tough job for the company to dynamically evaluate the gradual process of pLM maturity by using existing values and accurately make decisions of improving pLM maturity by selecting the optimum alternative. A fuzzy pLM components maturity model (pCMA) is presented to build the internal logical relationship between maturity levels and existing values that can automatically predict the unknown pLM maturity levels. A fuzzy AHp-VIKOR methodology is used to make a decision among option pLM strategies. The weights of the criteria are determined by fuzzy pairwise comparison matrices (pCM). The weights of alternatives with respect to criteria are calculated by fuzzy VIKOR. The fuzzy AHp-VIKOR is a compromise solution and has the ability of transfer subjective and implicit linguistics into objective and transparent data. A numerical example illustrates and clarifies the running steps of the proposed methodology.

关键词： Eco-design pLM maturity models fuzzy VIKOR fuzzy AHp group decision making pairwise comparison matrices

来源：评论

学校读者我要写书评

暂无评论

M-polynomial and Topological Indices of Benzene Ring Embedded in p-Type Surface Network

引用

JOURNAL OF CHEMISTRY 2019年第1期2019卷

作者： Yang, Hong Baig, A. Q. Khalid, W. Farahani, Mohammad Reza Zhang, Xiujun Chengdu Univ Inst Higher Educ Sichuan Prov Key Lab Pattern Recognit & Intelligent Informat P Chengdu 610106 Sichuan Peoples R China COMSATS Inst Informat Technol Dept Math Attock Campus Islamabad Pakistan Punjab Coll Commerce & Sci Attock Campus Lahore Pakistan Iran Univ Sci & Technol IUST Narmak Dept Appl Math Tehran 16844 Iran

The representation of chemical compounds and chemical networks with the M-polynomials is a new idea, and it gives nice and good results of the topological indices. These results are used to correlate the chemical compounds and chemical networks with their chemical properties and bioactivities. In this article, particular attention will be put on the derivation of M-polynomial for the benzene ring embedded in the p-type surface network in 2D. Furthermore, the topological indices based on the degrees are also derived by using the general form of M-polynomial of the benzene ring embedded in the p-type surface network BRm,n. In the end, the graphical representation and comparison of the M-polynomial and the topological indices of the benzene ring embedded in the p-type surface network in 2D are described.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Investigation of extreme ultraviolet spectra from highly charged holmium ions in 1μm laser-produced plasmas

引用

JOURNAL OF pHYSICS B-ATOMIC MOLECULAR AND OpTICAL pHYSICS 2020年第22期53卷

作者： Wu, Tao Kawasaki, Hiromu Shimada, Yuta Higashiguchi, Takeshi O'Sullivan, Gerry Wuhan Inst Technol Sch Math & Phys Hubei Key Lab Opt Informat & Pattern Recognit Wuhan 430205 Peoples R China Guangdong Intelligent Robot Inst Bldg 51Xuefu RdSSL Hi Tech Zone Dongguan 523808 Peoples R China Utsunomiya Univ Dept Elect & Elect Engn Fac Engn Yoto 7-1-2 Utsunomiya Tochigi 3218585 Japan Utsunomiya Univ Ctr Opt Res & Educ CORE Yoto 7-1-2 Utsunomiya Tochigi 3218585 Japan Univ Coll Dublin Sch Phys Dublin 4 Ireland

Absolute intensity calibrated extreme ultraviolet spectra radiated by highly charged holmium (Ho) ions from 1 mu m Nd:YAG laser generated plasmas in the 1-8 nm wavelength spectral range were measured and investigated. The spectral features show a broad structured continuum-like emission band, most prominent in the wavelength range 6-8 nm, which accounts for more than half of the emitted power. Assuming local thermodynamic equilibrium (LTE) in the laser produced plasmas and using the flexible atomic code and Cowan suite of codes, the wavelengths and LTE-gAvalues of unresolved transition arrays (UTAs) fromE(1)-contributing transition arrays 4p-4d, 4d-4f and super transition arrays such as 4p(6)4d(k-1)4f + 4p(5)4d(k+1)-4p(6)4d(k-2)4f(2)+ 4p(4)4d(k+2)+ 4p(5)4d(k)4f and 4d(10)(5s5p)(k)4f(m-k)-4d(9)(5s5p)(k)4f(m-k+1)were calculated and shown to be responsible for the strongest observed spectral structure. Transitions between these excited to excited states are considered, since even if only weakly present, they will not be influenced by opacity effects unlike resonance transitions involving ground configurations. Another two intense continuous emission bands at 3-6 nm and 2-3 nm that dominate the spectra in the shorter wavelength range mainly arise from 4d-5p, 4d5s-4d5s5p, 4f-5g, 4f5s-4f5s5g and 4d-5f, 4d5s-4d5s5f, 4p-5s transitions. Spectral line shape parameters of Gaussian fits to the LTE-gAweighted radiation spectra of each Ho ion were given, enabling direct explanation of recorded spectra. Based on the LTE-gAUTAs formalism parameters, reasonable agreement is obtained between the synthetic spectrum and experimental result enabling some of the observed spectral features to be identified.

关键词： holmium laser produced plasmas unresolved transition arrays flexible atomic code

来源：评论

学校读者我要写书评

暂无评论

A Novel Binary Chaotic Cloud Flower pollination Algorithm for Analog Fault Diagnosis 3

A Novel Binary Chaotic Cloud Flower Pollination Algorithm fo...

引用

3rd IEEE Advanced information Technology, Electronic and Automation Control Conference (IAEAC)

作者： Cui, Weijia He, Yuzhu Zhao, Dongsheng Zhu, Jing Beihang Univ Sch Instrumentat Sci & Optoelect Engn Beijing Peoples R China Beijing Jinghang Res Inst Comp & Commun Beijing Peoples R China Chengdu Aeronaut Polytech Chengdu Sichuan Peoples R China Chengdu Univ Key Lab Pattern Recognit & Intelligent Informat P Chengdu Sichuan Peoples R China

ISBN: (纸本)9781538645093

In analog fault diagnosis, support vector machine (SVM) classifier is a popular diagnosis technique. Before training the SVM classifier, feature extraction is an essential preprocessing step to enhance the identifiability of the fault samples. However, in the feature set, there exist the redundant and uncorrelated features, which have the inferior influence on the learning performance of the SVM classifier. Feature selection is to select the most representative feature subset to enhance the diagnosis accuracy. In addition, the penalty coefficient and kernel width parameter in SVM also impact the classification performance. Obviously, the selection of feature subset and SVM parameters can be considered as an optimization problem. Flower pollination algorithm (FpA) as a swarm intelligence optimization is derived from the pollination behavior of flowers. In this paper, we propose a novel binary version of FpA and then incorporate a cross search strategy based on chaos theory and cloud model into the novel BFpA to further enhance the solution quality, named BCCFpA. Moreover, different kinds of features are extracted by various wavelet basis functions based wavelet packet energy spectrum (wpes) and wavelet packet energy entropy (wpee) methods. The original feature set is built by combining these feature vectors. BCCFpA is used to simultaneously search the optimal feature subset and SVM parameters. The simulation experimental results prove that BCCFpA can effectively balance the diagnosis accuracy and the dimension of feature subset, and achieves the better diagnosis performance when compared with other methods.

关键词： binary FpA fault diagnosis of analog circuits feature selection SVM parameters optimization

来源：评论

学校读者我要写书评

暂无评论

MiRGOFS: a GO-based functional similarity measurement for miRNAs, with applications to the prediction of miRNA subcellular localization and miRNA-disease association

引用

BIOinformatICS 2018年第20期34卷 3547-3556页

作者： Yang, Yang Fu, Xiaofeng Qu, Wenhao Xiao, Yiqun Shen, Hong-Bin Shanghai Jiao Tong Univ Dept Comp Sci & Engn Shanghai 200240 Peoples R China Key Lab Shanghai Educ Commiss Intelligent Interac Shanghai 200240 Peoples R China Shanghai Jiao Tong Univ Inst Image Proc & Pattern Recognit Shanghai Peoples R China Minist Educ China Key Lab Syst Control & Informat Proc Shanghai 200240 Peoples R China

Motivation: Benefiting from high-throughput experimental technologies, whole-genome analysis of microRNAs (miRNAs) has been more and more common to uncover important regulatory roles of miRNAs and identify miRNA biomarkers for disease diagnosis. As a complementary information to the high-throughput experimental data, domain knowledge like the Gene Ontology and KEGG pathway is usually used to guide gene function analysis. However, functional annotation for miRNAs is scarce in the public databases. Till now, only a few methods have been proposed for measuring the functional similarity between miRNAs based on public annotation data, and these methods cover a very limited number of miRNAs, which are not applicable to large-scale miRNA analysis. Results: In this paper, we propose a new method to measure the functional similarity for miRNAs, called miRGOFS, which has two notable features: (i) it adopts a new GO semantic similarity metric which considers both common ancestors and descendants of GO terms;(i) it computes similarity between GO sets in an asymmetric manner, and weights each GO term by its statistical significance. The miRGOFS-based predictor achieves an F-1 of 61.2% on a benchmark dataset of miRNA localization, and AUC values of 87.7 and 81.1% on two benchmark sets of miRNA-disease association, respectively. Compared with the existing functional similarity measurements of miRNAs, miRGOFS has the advantages of higher accuracy and larger coverage of human miRNAs (over 1000 miRNAs).

关键词：

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：