检索结果-南通市图书馆

Lightweight Res-Connection Multi-Branch Network for Highly Accurate Crowd Counting and Localization

同方期刊数据库评论

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2024年第5期79卷 2105-2122页

作者： Mingze Li Diwen Zheng shuhua Lu College of Information and Cyber security People’s Public Security University of ChinaBeijing102600China

Crowd counting is a promising hotspot of computer vision involving crowd intelligence analysis,achieving tremendous success recently with the development of deep ***,there have been stillmany challenges including crowd multi-scale variations and high network complexity,*** tackle these issues,a lightweight Resconnection multi-branch network(LRMBNet)for highly accurate crowd counting and localization is ***,using improved shuffleNet V2 as the backbone,a lightweight shallow extractor has been designed by employing the channel compression mechanism to reduce enormously the number of network parameters.A light multi-branch structure with different expansion rate convolutions is demonstrated to extract multi-scale features and enlarged receptive fields,where the information transmission and fusion of diverse scale features is enhanced via residual *** addition,a compound loss function is introduced for training themethod to improve global context information *** proposed method is evaluated on the sHHA,sHHB,UCF-QNRF and UCF_CC_50 public *** accuracy is better than those of many advanced approaches,while the number of parameters is *** experimental results show that the proposed method achieves a good tradeoff between the complexity and accuracy of crowd counting,indicating a lightweight and high-precision method for crowd counting.

关键词： Crowd counting Res-connection multi-branch compound loss function

Heterophilic Graph Neural Network Based on spatial and Frequency Domain Adaptive Embedding Mechanism

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computer Modeling in Engineering & sciences 2024年第5期139卷 1701-1731页

作者： Lanze Zhang Yijun Gu Jingjie Peng College of Information and Cyber security People’s Public Security University of ChinaBeijing100038China

Graph Neural Networks(GNNs)play a significant role in tasks related to homophilic *** GNNs,based on the assumption of homophily,employ low-pass filters for neighboring nodes to achieve information aggregation and ***,in heterophilic graphs,nodes from different categories often establish connections,while nodes of the same category are located further apart in the graph *** characteristic poses challenges to traditional GNNs,leading to issues of“distant node modeling deficiency”and“failure of the homophily assumption”.In response,this paper introduces the spatial-Frequency domain Adaptive Heterophilic Graph Neural Networks(sFA-HGNN),which integrates adaptive embedding mechanisms for both spatial and frequency domains to address the aforementioned ***,for the first problem,we propose the“Distant spatial Embedding Module”,aiming to select and aggregate distant nodes through high-order randomwalk transition probabilities to enhance modeling *** the second issue,we design the“Proximal Frequency Domain Embedding Module”,constructing adaptive filters to separate high and low-frequency signals of nodes,and introduce frequency-domain guided attention mechanisms to fuse the relevant information,thereby reducing the noise introduced by the failure of the homophily *** deploy the sFA-HGNN on six publicly available heterophilic networks,achieving state-of-the-art results in four of ***,we elaborate on the hyperparameter selection mechanism and validate the performance of each module through experimentation,demonstrating a positive correlation between“node structural similarity”,“node attribute vector similarity”,and“node homophily”in heterophilic networks.

关键词： Heterophilic graph graph neural network graph representation learning failure of the homophily assumption

Attention-Enhanced Voice Portrait Model Using Generative Adversarial Network

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2024年第4期79卷 837-855页

作者： Jingyi Mao Yuchen Zhou YifanWang Junyu Li Ziqing Liu Fanliang Bu school of Information Network security People’s Public Security University of ChinaBeijing100038China

Voice portrait technology has explored and established the relationship between speakers’ voices and their facialfeatures, aiming to generate corresponding facial characteristics by providing the voice of an unknown *** to its powerful advantages in image generation, Generative Adversarial Networks (GANs) have now beenwidely applied across various fields. The existing Voice2Face methods for voice portraits are primarily based onGANs trained on voice-face paired datasets. However, voice portrait models solely constructed on GANs facelimitations in image generation quality and struggle to maintain facial similarity. Additionally, the training processis relatively unstable, thereby affecting the overall generative performance of the model. To overcome the abovechallenges,wepropose a novel deepGenerativeAdversarialNetworkmodel for audio-visual synthesis, namedAVPGAN(Attention-enhanced Voice Portrait Model using Generative Adversarial Network). This model is based ona convolutional attention mechanism and is capable of generating corresponding facial images from the voice ofan unknown speaker. Firstly, to address the issue of training instability, we integrate convolutional neural networkswith deep GANs. In the network architecture, we apply spectral normalization to constrain the variation of thediscriminator, preventing issues such as mode collapse. secondly, to enhance the model’s ability to extract relevantfeatures between the two modalities, we propose a voice portrait model based on convolutional attention. Thismodel learns the mapping relationship between voice and facial features in a common space from both channeland spatial dimensions independently. Thirdly, to enhance the quality of generated faces, we have incorporated adegradation removal module and utilized pretrained facial GANs as facial priors to repair and enhance the clarityof the generated facial images. Experimental results demonstrate that our AVP-GAN achieved a cosine similarity of0.511, ou

关键词： Cross-modal generation GANs voice portrait technology face synthesis

Exploring the correlation between temperature and crime:A case-crossover study of eight cities in America

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of safety science and Resilience 2024年第1期5卷 13-36页

作者： Jinming Hu Xiaofeng Hu Xin’ge Han Yan Lin Huanggang Wu Bing shen school of Information Technology and Cyber security People’s Public Security University of ChinaBeijing 100038China Center for Capital social safety People’s Public Security University of ChinaBeijing 100038China China Association for public safety Beijing 100084China school of Emergency Management and safety Engineering China University of Mining and TechnologyBeijing 100083China school of International Police studies People’s Public Security University of ChinaBeijing 100038China

Recent years have seen increasing academic interest in exploring the correlation between temperature and ***,it is uncertain whether similar long-term trends or seasonality(rather than causal effect)of temper-ature and crime is the major reason for the observed correlation between *** explore whether there is still a correlation between temperature and crime when long-term trends and seasonal cycles are filtered out,we use the Kalman filter to decompose the time series of temperature and crimes,and then the fast Fourier transform is used to calculate the exact circle of their seasonality *** on that,the box-plot method and linear regression are used to explore the correlation between temperature residuals and crime *** results show that more than half of the crime types have similar seasonal cycles(approximately 1 year)to that of ***,the daily residual analyses show that temperature residuals have a positive correlation with assault and robbery residuals in all cities,whose average slopes are more than *** other four types of crimes vary greatly from case to *** temperature residuals show a weak correlation with the residuals of some crime types.

关键词： Time series decomposition Crime Temperature Kalman filter Fast fourier transform

sAT-Based Automatic searching for Differential and Linear Trails: Applying to CRAX

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

chinese Journal of Electronics 2024年第1期33卷 72-79页

作者： Yiyi HAN Caibing WANG Zhongfeng NIU lei HU state Key Laboratory of Information security Institute of Information Engineering Chinese Academy of Sciences school of Cyber security University of Chinese Academy of Sciences

Boolean satisfiability problem(sAT) is now widely applied in differential cryptanalysis and linear cryptanalysis for various cipher *** generated many excellent results for some ciphers,for example,*** this research,we study the differential and linear propagations through the operations of addition,rotation and XOR(ARX),and construct the sAT *** apply the models to CRAX to search differential trails and linear trails *** this sense,our contnbution can be broadly divided into two *** give the bounds for differential and linear cryptanalysis of Alzette both up to 12 steps,by which we present a 3-round differential attack and a 3-round linear attack for *** construct a 4-round key-recovery attack for CRAX with time complexity289times of 4-round encryption and data complexity 225.

关键词： Ciphers Encryption Complexity theory Time complexity

Eu-MOF-Based Fluorescent Ratiometric sensor by Detecting 3,4,5-Trihydroxybenzoic for Fingerprint Visualization on Porous Objects

同方期刊数据库评论

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Photonic sensors 2024年第1期14卷 85-98页

作者： Muwen LIANG Yabin ZHAO Yaping LUO Bin DU Wei HU Bing LIU Xihui MU Zhaoyang TONG Department of Investigation People's Public Security University of ChinaBejing 100038China state Key Laboratory of NBC Protection for Civilian Beijing 102205China

Latent fingerprints(LFPs)at the crime scene are served as important clues to locate the trajectory of criminal behavior and portray the characteristics of the ***,visualizing LFPs is of considerable *** this work,the europium metal-organic framework(Eu-MOF)sensor was successfully constructed for sensitive detection of gallic acid(3,4,5-trihydroxybenzoic acid,GA)and visualization of the sweat *** boric-acid-modified Eu-MOF was prepared by using the simple one-pot solvothermal method using Eu as the metal i ion center and 3,5-dicarboxybenzeneboronic acid(BBDC)as the organic *** sensor showed desirable photoluminescent performance through the chelating of BBDC with Eu3+.The sensor exhibited the satisfactory linear relationship to GA in the range of 1 nM to 20 nM with a low detection limit of 0.34 nM under the optimized *** prepared sensor with ideal selectivity to GA was successfully applied for visualizing LFPs on porous substrates with the high contrast and superior *** the good performance of the sensor,all fingerprint images obtained from 1200 samples presented clear friction ridges and met the identification ***,the sensor had less impact on the subsequent deoxyribonucleic acid(DNA)detection,displaying a promising perspective for applications in extracting physical evidence of site investigation.

关键词： Europium metal-organic framework(Eu-MOF) photoluminescence 3,4,5-trihydroxybenzoic latent fingerprints(LFPs)development

Complementary monogamy and polygamy properties among multipartite systems

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

chinese Physics B 2024年第3期33卷 207-213页

作者：李陶周静怡孙琪靳志祥梁登峰罗婷 school of Mathematics and statistics Beijing Technology and Business UniversityBeijing 100048China school of Computer science and Techonology Dongguan University of TechnologyDongguan 523808China people's public security university of China Academy of Information Network SecurityBeijing 100038China

Monogamy and polygamy relations are essential properties of quantum entanglement,which characterize the distributions of entanglement in multipartite *** this paper,we establish the general monogamy relations forγ-th(0≤γ≤α,α≥1)power of quantum entanglement based on unified-(q,s)entanglement and polygamy relations forδ-th(δ≥β,0≤β≤1)power of entanglement of assistance based on unified-(q,s)entanglement of assistance,which provides a complement to the previous research in terms of different parameter regions ofγandδ.These results are then applied to specific quantum correlations,e.g.,entanglement of formation,Renyi-q entanglement of assistance and Tsallis-q entanglement of assistance to get the corresponding monogamy and polygamy ***,typical examples are presented for illustration.

关键词： monogamy relation polygramy relation

DLP:towards active defense against backdoor attacks with decoupled learning process

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Cybersecurity 2024年第1期7卷 122-134页

作者： Zonghao Ying Bin Wu state Key Laboratory of Information security Institute of Information EngineeringChinese Academy of SciencesBeijingChina school of Cyber security University of Chinese Academy of SciencesBeijingChina

Deep learning models are well known to be susceptible to backdoor attack,where the attacker only needs to provide a tampered dataset on which the triggers are *** trained on the dataset will passively implant the backdoor,and triggers on the input can mislead the models during *** study shows that the model shows different learning behaviors in clean and poisoned subsets during *** on this observation,we propose a general training pipeline to defend against backdoor attacks *** models can be trained from the unreli-able dataset by decoupling the learning process into three stages,i.e.,supervised learning,active unlearning,and active semi-supervised *** effectiveness of our approach has been shown in numerous experiments across various backdoor attacks and datasets.

关键词： Deep learning Backdoor attack Active defense

Lightweight Multi-Resolution Network for Human Pose Estimation

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computer Modeling in Engineering & sciences 2024年第3期138卷 2239-2255页

作者： Pengxin Li Rong Wang Wenjing Zhang Yinuo Liu Chenyue Xu school of Information and Cyber security People Public Security University of ChinaBeijing100038China Key Laboratory of security Prevention Technology and Risk Assessment of Ministry of public security Beijing100038China

Human pose estimation aims to localize the body joints from image or video *** the development of deeplearning,pose estimation has become a hot research topic in the field of computer *** recent years,humanpose estimation has achieved great success in multiple fields such as animation and ***,to obtainaccurate positioning results,existing methods may suffer from large model sizes,a high number of parameters,and increased complexity,leading to high computing *** this paper,we propose a new lightweight featureencoder to construct a high-resolution network that reduces the number of parameters and lowers the *** also introduced a semantic enhancement module that improves global feature extraction and networkperformance by combining channel and spatial ***,we propose a dense connected spatialpyramid pooling module to compensate for the decrease in image resolution and information loss in the ***,ourmethod effectively reduces the number of parameters and complexitywhile ensuring high *** experiments show that our method achieves a competitive performance while dramatically reducing thenumber of parameters,and operational ***,our method can obtain 89.9%AP score on MPIIVAL,while the number of parameters and the complexity of operations were reduced by 41%and 36%,respectively.

关键词： Lightweight human pose estimation keypoint detection high resolution network