检索结果-内蒙古大学图书馆

IEEE International Conference on Robotics and Automation (ICRA)

作者： Wenhui Huang Jason Gu Peiyong Duan Sujuan Hou Yuanjie Zheng School of Information Science and Engineering Shandong Normal University China Dalhousie University Halifax Canada Key Lab of Intelligent Computing & Information Security in Universities of Shandong Shandong Normal University China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Shandong Normal University China

Visual tracking is a fundamental capability for robots tasked with humans and environment interaction. However, state-of-the-art visual tracking methods are still prone to failures and are imprecise when applied to challenging stereos, and their results are generally confidence agonistic. These methods depend on an embedded deep learning model to provide deterministic features or regression maps. A deterministic output with low confidence can result in disastrous consequences and lacks evidence needed for subsequent operations. Moreover, training data ambiguities or noise in the observations (so-called data uncertainty) can also lead to inherent uncertainty. In this paper, we focus on exploiting probabilistic Siamese visual tracking with a conditional variational autoencoder (CVAE). First, we build a bridge between the Siamese architecture and the CVAE and propose a novel Bayesian visual tracking method. Second, the proposed method generates a complete probability distribution that enables the production of multiple plausible tracking outputs. Third, CVAE conditioned by ground truth data encodes a low-dimensional latent space and conducts noise-injection training to prevent overfitting. Our proposed tracking method outperformed the state-of-the-art trackers on the VOT2016, VOT2018 and TColor-128 datasets.

关键词： Bridges Training Visualization Uncertainty Training data Probabilistic logic Probability distribution

来源：评论

学校读者我要写书评

暂无评论

new Method of MSCA for Edge computing

New Method of MSCA for Edge Computing

引用

IEEE International Conference on Smart Internet of Things (SmartIoT)

作者： De-gan Zhang Hao-li Zhu Ting Zhang Jia-xu Wang Tianjin Key Lab of Intelligent Computing & Novel software Technology Tianjin University of Technology Tianjin China Key Laboratory of Computer Vision and System (Tianjin University of Technology) Ministry of Education Tianjin

ISBN: (数字)9781728165141

ISBN: (纸本)9781728165158

By innovative use of edge computing technology to construct a node data cache model and step-by-step calculation of node channel separation, a multi-strategy channel allocation algorithm (MSCA) for edge computing is proposed. The mechanism of this algorithm is to use edge computing technology to pre-store the data in the nodes in the multicast tree and calculate the channel separation between the nodes, and then select the transmission channel number of the node with the least interference to avoid mutual interference between node information transmissions. Through experimental tests and comparisons, the MSCA algorithm for edge computing presented in this paper can minimize channel interference and overall network energy consumption while satisfying throughput and end-to-end delay.

关键词： Edge computing Interference Computational modeling Delays Peer-to-peer computing Channel allocation Protocols

来源：评论

学校读者我要写书评

暂无评论

Accurate Latent Factor Analysis via Particle Swarm Optimizers

Accurate Latent Factor Analysis via Particle Swarm Optimizer...

引用

IEEE International Conference on Systems, Man and Cybernetics

作者： Jia Chen Xin Luo MengChu Zhou School of Cyber Science and Technology Beihang University Beijing China Chongqing Key Lab. of Big Data and Intelligent Computing and the Chongqing Engineering Research Center of Big Data Application for Smart Cities Chongqing Institute of Green and Intelligent Technology Chongqing School University of Chinese Academy of Sciences Chongqing China New Jersey Institute of Technology Newark NJ USA

ISBN: (纸本)9781665442084

A stochastic-gradient-descent-based Latent Factor Analysis (LFA) model is highly efficient in representative learning of a High-Dimensional and Sparse (HiDS) matrix. Its learning rate adaptation is vital in ensuring its efficiency. Such adaptation can be realized with an evolutionary computing algorithm. However, a resultant model tends to suffer from two issues: a) the pre-mature convergence of the swarm of learning rates as caused by an adopted evolution algorithm, and b) the pre-mature convergence of the LFA model as caused jointly by evolution-based learning rate adaptation and an optimization algorithm. This paper focuses on the methods to address such issues. A Hierarchical Particle-swarm-optimization-incorporated Latent factor analysis (HPL) model with a two-layered structure is proposed, where the first layer pre-trains desired latent factors with a position-transitional particle-swarm-optimization-based LFA model, and the second layer performs latent factor refining with a newly-proposed mini-batch particle swarm optimizer. With such design, an HPL model can well handle the pre-mature convergence, which is supported by the positive experimental results achieved on HiDS matrices from industrial applications.

关键词： Adaptation models Analytical models Computational modeling Refining Predictive models Sparse matrices Particle swarm optimization

来源：评论

学校读者我要写书评

暂无评论

A new method of mobile Ad Hoc network routing based on improved GF strategy 3

A new method of mobile Ad Hoc network routing based on impro...

引用

3rd IEEE International Conference on Smart Internet of Things, SmartIoT 2019

作者： Li, Ke Zhang, De-Gan Qiu, Jian-Ning Chen, Lu Tianjin Key Lab of Intelligent Computing and Novel Software Technology Tianjin University of Technology Tianjin300384 China Key Laboratory of Computer Vision and System Tianjin University of Technology Ministry of Education Tianjin China

ISBN: (纸本)9781728134888

We propose a new mobile ad hoc network routing method based on improved GF (greedy forwarding) strategy in this paper. In the greedy forwarding phase, the reliable communication area is calculated, and then the quality of the link is evaluated according to the relative displacement between the nodes and the maintenance time of the link. Then, according to the quality of the link, the distance from the candidate node to the destination node, and the number of the neighbor nodes, the metric value is obtained, and the node with the large metric value is selected as the next hop. When a routing hole occurs, the waiting forwarding mode is used for a period of time. After a period of time, if the current node is still a routing hole node, the right-hand rule and the left-hand rule are used, and the bypass mode is performed simultaneously. By considering three factors of the deflection angle and the residual energy of the node and the maintenance time of the link, the forwarding node as the next hop with the highest priority value is selected in each direction. According to experiments, the method of this paper has achieved good experimental results. © 2019 IEEE.

关键词： Energy utilization

来源：评论

学校读者我要写书评

暂无评论

Universal deep network for steganalysis of color image based on channel representation

arXiv

引用

arXiv 2021年

作者： Wei, Kangkang Luo, Weiqi Tan, Shunquan Huang, Jiwu The Guangdong Key Lab of Information Security Technology School of Computer Science and Engineering Sun Yat-Sen University Guangzhou510006 China The College of Computer Science and Software Engineering Shenzhen University Shenzhen518060 China The Guangdong Key Laboratory of Intelligent Information Processing Shenzhen Key Laboratory of Media Security National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen518060 China

Up to now, most existing steganalytic methods are designed for grayscale images, and they are not suitable for color images that are widely used in current social networks. In this paper, we design a universal color image steganalysis network (called UCNet) in spatial and JPEG domains. The proposed method includes preprocessing, convolutional, and classification modules. To preserve the steganographic artifacts in each color channel, in preprocessing module, we firstly separate the input image into three channels according to the corresponding embedding spaces (i.e. RGB for spatial steganography and YCbCr for JPEG steganography), and then extract the image residuals with 62 fixed high-pass filters, finally concatenate all truncated residuals for subsequent analysis rather than adding them together with normal convolution like existing CNN-based steganalyzers. To accelerate the network convergence and effectively reduce the number of parameters, in convolutional module, we carefully design three types of layers with different shortcut connections and group convolution structures to further learn high-level steganalytic features. In classification module, we employ a global average pooling and fully connected layer for classification. We conduct extensive experiments on ALASKA II to demonstrate that the proposed method can achieve state-of-the-art results compared with the modern CNN-based steganalyzers (e.g., SRNet and J-YeNet) in both spatial and JPEG domains, while keeping relatively few memory requirements and training time. Furthermore, we also provide necessary descriptions and many ablation experiments to verify the rationality of the network design. Copyright © 2021, The Authors. All rights reserved.

关键词： Steganography

来源：评论

学校读者我要写书评

暂无评论

Brief review of image denoising techniques

引用

Visual computing for Industry,Biomedicine,and Art 2019年第1期2卷 55-66页

作者： Linwei Fan Fan Zhang Hui Fan Caiming Zhang School of Software Shandong UniversityShunHua Road No.1500Jinan 250101China Shandong Co-Innovation Center of Future Intelligent Computing BinHai Road No.191Yantai 264005China Shandong Province Key Lab of Digital Media Technology Shandong University of Finance and EconomicsBicyclic East Road No.7366Jinan 250061China

With the explosion in the number of digital images taken every day,the demand for more accurate and visually pleasing images is ***,the images captured by modern cameras are inevitably degraded by noise,which leads to deteriorated visual image ***,work is required to reduce noise without losing image features(edges,corners,and other sharp structures).So far,researchers have already proposed various methods for decreasing *** method has its own advantages and *** this paper,we summarize some important research in the field of image ***,we give the formulation of the image denoising problem,and then we present several image denoising *** addition,we discuss the characteristics of these ***,we provide several promising directions for future research.

关键词： Image denoising Non-local means Sparse representation Low-rank Convolutional neural network

来源：评论

学校读者我要写书评

暂无评论

Epilepsy Classification for Mining Deeper Relationships between EEG Channels based on GCN

Epilepsy Classification for Mining Deeper Relationships betw...

引用

Computer Vision, Image and Deep Learning (CVIDL), International Conference on

作者： Xin Chen Yuanjie Zheng Yi Niu Chengiiang Li School of Information Science and Engineering Shandong Normal University Jinan China Key Lab of Intelligent Computing and Information Security Universities of Shandong Jinan China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Institute of Biomedical Sciences Jinan China Shandong University of Science and Technology Jinan China

ISBN: (数字)9781728194813

ISBN: (纸本)9781728194820

Epilepsy is a brain disorder caused by abnormal discharges of neurons in brain. It is one of the most commonly studied disorders in neurology. The research of epilepsy electroencephalogram (EEG) has become a hot research topic. We find that in epilepsy EEG detection task, many previous methods focused on directly collecting the data of each channel, but these methods seldom analyse relationships between signals. Therefore, we propose the Epilepsy EEG Graph Convolutional Network EGCN, which makes full use of correlations between channels to deeply mine data information. We specifically design 5-layer graph convolutional network structure for classification of healthy and epileptic patients. The method is applied to public data set (Boon and CHB-MIT) to establish a reasonable classification model. And we compare it with some advanced algorithms. The experimental results show that the E-GCN method is superior to many existing methods in classification accuracy. In brief, the E-GCN method can be effectively used in classification and detection for epilepsy. This provides new ideas for colleagues, who study epilepsy EEG. In addition, this also provides richer experience for diagnosis of epilepsy.

关键词： Epilepsy Electroencephalography Convolution Brain modeling Feature extraction Correlation Task analysis

来源：评论

学校读者我要写书评

暂无评论

A novel method of android malware detection based on ensemble learning algorithm 8

A novel method of android malware detection based on ensembl...

引用

2018 8th International Workshop on Computer Science and Engineering, WCSE 2018

作者： Zhao, Jingyi Mo, Xiuliang Zheng, Qiao Tianjin Intelligent Computing and Software New Technology Key Laboratory School of computer Science and Engineering Tianjin University of Technology China

ISBN: (纸本)9789811178610

Android is popular for mobile devices in recent years. As the risk of malware is sharply increasing in Android platform, Android malware detection has become an important research topic. However, the current work shows that the detection of malware still needs to be reformed. Many current results suggest a modification of permissions or a combination of permissions and intents, but Android's fragmentation issues and requiring rooting, hindering the widespread adoption of those methods. Many approaches have been proposed to detect this attack by modifying the Android OS. Existing anti-viruses depend on signature databases that need to be updated from time to time and are unable to detect zero-day malware. The Android Operating system allows inter-application communication through the use of component reuse by using intents. A malware detection model is developed based on ensemble learning algorithms and on the basis of the random forest in WEKA. The method we proposed can take a better efficiency and precision. © Proceedings of 2018 the 8th International Workshop on Computer Science and Engineering, WCSE 2018. All rights reserved.

关键词： Android (operating system)

来源：评论

学校读者我要写书评

暂无评论

Microblog text clustering based on BK-means algorithm 8

Microblog text clustering based on BK-means algorithm

引用

2018 8th International Workshop on Computer Science and Engineering, WCSE 2018

作者： Li, Qianru Mo, Xiuliang Wang, Chundong Tianjin Intelligent Computing and Software New Technology Key Laboratory School of Computer Science and Engineering Tianjin university of technology Tianjin300384 China

ISBN: (纸本)9789811178610

In recent years, the increasing popularity of social media such as WeChat and Weibo has facilitated the communication among people. However, due to the characteristics like large scale, fast propagation, low quality and diverse modalities of social short texts, the short text clustering faces the challenge of sparse features, high dimension and noise interference. The traditional clustering method based on vector space model is not good for short text data processing. With the improvement of K-means algorithm, this paper proposes a short-text clustering algorithm named BK-means which alleviates the effect of data sparseness. Firstly, we preprocess the wordset by means of word segmentation, stop-of-word and other operations, then extract the biterm using the BTM to model the document, and get the document-topic, the topic-word distribution matrix. Finally, we use the proposed BK-means algorithm to cluster short texts of documents represented by vectors. Experiments on the short text data of Sina Weibo have proved that the short text clustering algorithm based on BK-means is superior to the traditional one, and both the F-measure and the purity are improved. © Proceedings of 2018 the 8th International Workshop on Computer Science and Engineering, WCSE 2018. All rights reserved.

关键词： K-means clustering

来源：评论

学校读者我要写书评

暂无评论

A Kind of new parallax algorithm based on symmetric continuous optimization

引用

Journal of Physics: Conference Series 2021年第3期1748卷

作者： Kaiwen Jiang Degan Zhang Haixia Xu Key Laboratory of Computer Vision and System (Tianjin University of Technology) Ministry of Education 300384 China Tianjin Key Lab of Intelligent Computing & Novel software Technology Tianjin University of Technology Tianjin China School of Electronic and Information Engineering Tianjin Vocational Institute Tianjin 300410 China

With the fast development of 3D imaginations becomes more and more fascination, multi-view stereo based 3D reconstruction is a significant technique for those application. To facilitate the subsequent processing of 3D reconstruction and reduce the possibility of other algorithms falling into local optimal solutions, attempting to get better and faster performance, a new parallax calculation based on symmetric continuous optimization is proposed in this paper. The algorithm is proposed here will be tested respectively in the same data, in order to certificate the algorithm applied is better than traditional minimize El algorithm.

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：