检索结果-南通市图书馆

Towards imbalanced motion:part-decoupling network for video portrait segmentation

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2024年第7期67卷 197-210页

作者： Tianshu YU Changqun XIA Jia LI State Key laboratory of Virtual Reality Technology and systems School of Computer Science and EngineeringBeihang University Peng Cheng laboratory

Video portrait segmentation(VPS), aiming at segmenting prominent foreground portraits from video frames, has received much attention in recent years. However, the simplicity of existing VPS datasets leads to a limitation on extensive research of the task. In this work, we propose a new intricate large-scale multi-scene video portrait segmentation dataset MVPS consisting of 101 video clips in 7 scenario categories,in which 10843 sampled frames are finely annotated at the pixel level. The dataset has diverse scenes and complicated background environments, which is the most complex dataset in VPS to our best *** the observation of a large number of videos with portraits during dataset construction, we find that due to the joint structure of the human body, the motion of portraits is part-associated, which leads to the different parts being relatively independent in motion. That is, the motion of different parts of the portraits is imbalanced. Towards this imbalance, an intuitive and reasonable idea is that different motion states in portraits can be better exploited by decoupling the portraits into parts. To achieve this, we propose a part-decoupling network(PDNet) for VPS. Specifically, an inter-frame part-discriminated attention(IPDA)module is proposed which unsupervisedly segments portrait into parts and utilizes different attentiveness on discriminative features specified to each different part. In this way, appropriate attention can be imposed on portrait parts with imbalanced motion to extract part-discriminated correlations, so that the portraits can be segmented more accurately. Experimental results demonstrate that our method achieves leading performance with the comparison to state-of-the-art methods.

关键词： video portrait segmentation imbalanced motion unsupervised part decoupling motion correlation inter-frame attention

Deep learning based water leakage detection for shield tunnel lining

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Frontiers of Structural and Civil Engineering 2024年第6期18卷 887-898页

作者： Shichang LIU Xu XU Gwanggil JEON Junxin CHEN Ben-Guo HE College of computer Science Sichuan UniversityChengdu 610065China School of computer Science and Engineering Northeastern UniversityShenyang 110004China Department of embedded systems Engineering Incheon National UniversityIncheon 22012Korea School of Software Dalian University of TechnologyDalian 116621China Key laboratory of Ministry of Education on Safe Mining of Deep Metal Mines Northeastern UniversityShenyang 110819China

Shield tunnel lining is prone to water leakage,which may further bring about corrosion and structural damage to the walls,potentially leading to dangerous *** avoid tedious and inefficient manual inspection,many projects use artificial intelligence(Al)to detect cracks and water leakage.A novel method for water leakage inspection in shield tunnel lining that utilizes deep learning is introduced in this *** proposal includes a ConvNeXt-S backbone,deconvolutional-feature pyramid network(D-FPN),spatial attention module(SPAM).and a detection *** can extract representative features of leaking areas to aid inspection *** further improve the model's robustness,we innovatively use an inversed low-light enhancement method to convert normally illuminated images to low light ones and introduce them into the training *** experiments are performed,achieving the average precision(AP)score of 56.8%,which outperforms previous work by a margin of 5.7%.Visualization illustrations also support our method's practical effectiveness.

关键词： water leakage detection deep learning deconvolutional-feature pyramid spatial attention

A Survey on Graph Neural Network Acceleration: A Hardware Perspective

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3期33卷 601-622页

作者： Shi CHEN Jingyu LIU Li SHEN School of computer National University of Defense Technology Key laboratory of Advanced Microprocessor Chips and systems

Graph neural networks(GNNs) have emerged as powerful approaches to learn knowledge about graphs and *** rapid employment of GNNs poses requirements for processing *** to incompatibility of general platforms,dedicated hardware devices and platforms are developed to efficiently accelerate training and inference of *** conduct a survey on hardware acceleration for *** first include and introduce recent advances of the domain,and then provide a methodology of categorization to classify existing works into three ***,we discuss optimization techniques adopted at different *** finally we propose suggestions on future directions to facilitate further works.

关键词： Graph neural networks Deep learning acceleration Domain-specific architecture Hardware accelerator

Friendship Inference Based on Interest Trajectory Similarity and Co-occurrence

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2024年第3期33卷 708-720页

作者： Junfeng TIAN Zhengqi HOU School of Cyber Security and computer Hebei University Hebei Key laboratory of High Confidence Information systems Hebei University

Most of the current research on user friendship speculation in location-based social networks is based on the co-occurrence characteristics of users,however,statistics find that co-occurrence is not common among all users;meanwhile,most of the existing work focuses on mining more features to improve the accuracy but ignoring the time complexity in practical *** this basis,a friendship inference model named ITSIC is proposed based on the similarity of user interest tracks and joint user location *** utilizing MeanShift clustering algorithm,ITSIC clustered and filtered user check-ins and divided the dataset into interesting,abnormal,and noise *** interest trajectories were constructed from user interest check-in data,which allows ITSIC to work efficiently even for users without *** the same time,by application of clustering,the single-moment multi-interest trajectory was further proposed,which increased the richness of the meaning of the trajectory *** extensive experiments on two real online social network datasets show that ITSIC outperforms existing methods in terms of AUC score and time efficiency compared to existing methods.

关键词： Data privacy Social networking (online) Noise Clustering algorithms Prediction algorithms Inference algorithms Data models

Multi-Scale Binocular Stereo Matching Based on Semantic Association

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2024年第04期33卷 1010-1022页

作者： Jin ZHENG Botao JIANG Wei PENG Qiaohui ZHANG State Key laboratory of Virtual Reality Techonology and systems Beihang University School of computer Science and Engineering Beihang University

Aiming at the low accuracy of existing binocular stereo matching and depth estimation methods, this paper proposes a multi-scale binocular stereo matching network based on semantic association. A semantic association module is designed to construct the contextual semantic association relationship among the pixels through semantic category and attention mechanism. The disparity of those regions where the disparity is easily estimated can be used to assist the disparity estimation of relatively difficult regions, so as to improve the accuracy of disparity estimation of the whole image. Simultaneously, a multi-scale cost volume computation module is proposed. Unlike the existing methods, which use a single cost volume, the proposed multi-scale cost volume computation module designs multiple cost volumes for features of different scales. The semantic association feature and multi-scale cost volume are aggregated, which fuses the high-level semantic information and the low-level local detailed information to enhance the feature representation for accurate stereo matching. We demonstrate the effectiveness of the proposed solutions on the KITTI2015 binocular stereo matching dataset, and our model achieves comparable or higher matching performance, compared to other seven classic binocular stereo matching algorithms.

关键词： Costs Accuracy Attention mechanisms Fuses Computational modeling Semantics Estimation

Parsing Objects at a Finer Granularity: A Survey

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Machine Intelligence Research 2024年第3期21卷 431-451页

作者： Yifan Zhao Jia Li Yonghong Tian School of computer Science Peking UniversityBeijing100871China State Key laboratory of Virtual Reality Technology and systems School of Computer Science and EngineeringBeihang UniversityBeijing100191China

Fine-grained visual parsing, including fine-grained part segmentation and fine-grained object recognition, has attracted considerable critical attention due to its importance in many real-world applications, e.g., agriculture, remote sensing, and space technologies. Predominant research efforts tackle these fine-grained sub-tasks following different paradigms, while the inherent relations between these tasks are neglected. Moreover, given most of the research remains fragmented, we conduct an in-depth study of the advanced work from a new perspective of learning the part relationship. In this perspective, we first consolidate recent research and benchmark syntheses with new taxonomies. Based on this consolidation, we revisit the universal challenges in fine-grained part segmentation and recognition tasks and propose new solutions by part relationship learning for these important challenges. Furthermore, we conclude several promising lines of research in fine-grained visual parsing for future research.

关键词： Finer granularity visual parsing part segmentation fine-grained object recognition part relationship

For Mega-Constellations: Edge Computing and Safety Management Based on Blockchain Technology

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

China Communications 2024年第2期21卷 59-73页

作者： Zhen Zhang Bing Guo Chengjie Li College of computer Science Sichuan UniversityChengdu 610064China The Key laboratory for computer systems of State Ethnic Affairs Commission School of Computer Science and TechnologySouthwest Minzu UniversityChengdu 610225China National Key laboratory of Science and Technology on Communications University of Electronic Science and Technology of ChinaChengdu 611731China

In mega-constellation Communication systems, efficient routing algorithms and data transmission technologies are employed to ensure fast and reliable data transfer. However, the limited computational resources of satellites necessitate the use of edge computing to enhance secure *** edge computing reduces the burden on cloud computing, it introduces security and reliability challenges in open satellite communication channels. To address these challenges, we propose a blockchain architecture specifically designed for edge computing in mega-constellation communication systems. This architecture narrows down the consensus scope of the blockchain to meet the requirements of edge computing while ensuring comprehensive log storage across the network. Additionally, we introduce a reputation management mechanism for nodes within the blockchain, evaluating their trustworthiness, workload, and efficiency. Nodes with higher reputation scores are selected to participate in tasks and are appropriately incentivized. Simulation results demonstrate that our approach achieves a task result reliability of 95% while improving computational speed.

关键词： blockchain consensus mechanism edge computing mega-constellation reputation management

维普期刊数据库

For LEO Satellite Networks: Intelligent Interference Sensing and Signal Reconstruction Based on Blind Separation Technology

在线全文

学校读者我要写书评

暂无评论

China Communications 2024年第2期21卷 85-95页

作者： Chengjie Li Lidong Zhu Zhen Zhang The Key laboratory for computer systems of State Ethnic Affairs Commission School of Computer Science and Technology(Southwest Minzu University)Chengdu 610000China National Key laboratory of Science and Technology on Communications(University of Electronic Science and Technology of China) Chengdu 610000China College of computer Science(Sichuan University) Chengdu 610000China

In LEO satellite communication networks,the number of satellites has increased sharply, the relative velocity of satellites is very fast, then electronic signal aliasing occurs from time to time. Those aliasing signals make the receiving ability of the signal receiver worse, the signal processing ability weaker,and the anti-interference ability of the communication system lower. Aiming at the above problems, to save communication resources and improve communication efficiency, and considering the irregularity of interference signals, the underdetermined blind separation technology can effectively deal with the problem of interference sensing and signal reconstruction in this scenario. In order to improve the stability of source signal separation and the security of information transmission, a greedy optimization algorithm can be executed. At the same time, to improve network information transmission efficiency and prevent algorithms from getting trapped in local optima, delete low-energy points during each iteration process. Ultimately, simulation experiments validate that the algorithm presented in this paper enhances both the transmission efficiency of the network transmission system and the security of the communication system, achieving the process of interference sensing and signal reconstruction in the LEO satellite communication system.

关键词： blind source separation greedy optimization algorithm interference sensing LEO satellite communication networks signal reconstruction

维普期刊数据库

DCFNet:Discriminant Correlation Filters Network for Visual Tracking

在线全文

学校读者我要写书评

暂无评论

Journal of computer Science & Technology 2024年第3期39卷 691-714页

作者：胡卫明王强高晋李兵 Stephen Maybank State Key laboratory of Multimodal Artificial Intelligence systems Institute of AutomationChinese Academy of Sciences Beijing 100190China Department of computer Science and Information systems Birkbeck CollegeLondon WC1E 7HXU.K.

CNN(convolutional neural network)based real time trackers usually do not carry out online network update in order to maintain rapid tracking *** inevitably influences the adaptability to changes in object *** filter based trackers can update the model parameters online in real *** this paper,we present an end-to-end lightweight network architecture,namely Discriminant Correlation Filter Network(DCFNet).A differentiable DCF(discriminant correlation filter)layer is incorporated into a Siamese network architecture in order to learn the convolutional features and the correlation filter *** correlation filter can be efficiently updated *** previous work,we introduced a joint scale-position space to the DCFNet,forming a scale DCFNet which carries out the predictions of object scale and position *** combine the scale DCFNet with the convolutional-deconvolutional network,learning both the high-level embedding space representations and the low-level fine-grained representations for *** adaptability of the fine-grained correlation analysis and the generalization capability of the semantic embedding are complementary for visual *** back-propagation is derived in the Fourier frequency domain throughout the entire work,preserving the efficiency of the *** evaluations on the OTB(Object Tracking Benchmark)and VOT(Visual Object Tracking Challenge)datasets demonstrate that the proposed trackers have fast speeds,while maintaining tracking accuracy.

关键词： correlation filter convolutional neural network(CNN) visual tracking