检索结果-内蒙古大学图书馆

4th International Conference on computer Vision, Image and Deep Learning, CVIDL 2023

作者： Dang, Yuanyuan Li, Chenxi Zhang, Lijie Gao, Yu College of Computer Science and Engineering Changchun University of Technology Changchun China

ISBN: (纸本)9798350326444

Monocular depth estimation is an important research field in computer vision, with many applications in autonomous driving, 3D reconstruction, and other fields. However, previous research on monocular depth estimation often resulted in significant computational overhead. Therefore, we design a lightweight model for monocular depth estimation, which is Reducing Channel Convolution Network (RCCNet). RCCNet introduces Reducing Channel Convolution (RCConv) to extract local features with a lower computation cost. Extensive experiments show that this method can achieve comparable results while a tiny computation cost. © 2023 IEEE.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Personal Identification Based on Joint Motions with Normalization against Size Changes of Targets

Personal Identification Based on Joint Motions with Normaliz...

引用

2024 International Symposium on Intelligent Signal Processing and Communication Systems, ISPACS 2024

作者： Yuda, Yuki Adachi, Miho Kamiya, Sho Aoyagi, Kosuke Miyamoto, Ryusuke Department of Computer Science Graduate School of Science and Technology Meiji University Kawasaki Japan Department of Computer Science School of Science and Technology Meiji University Kawasaki Japan

ISBN: (纸本)9798350389210

Gait analysis is useful for personal identification in public spaces. Recent, advancements in deep learning technology have enabled highly accurate estimation of human joint positions in images, making practical applications of gait analysis possible using only visible light cameras. We propose a method to improve the performance of personal identification methods based on joint movements captured by cameras, aiming to make them more practical for use in real-world scenarios. Specifically, we propose a novel normalization method for feature vectors to address variations in the size of target humans during walking sequences. Such variations have a significant impact on identification accuracy in gait analysis. The proposed method normalizes feature vectors based on the distance between the hip and neck joints, reducing the impact of differences in target human size in the images. To evaluate the effectiveness of this method, we created two datasets by intentionally varying the magnification of existing datasets. Experimental results showed that the proposed method improved identification accuracy by approximately 12%. Furthermore, even when multiple datasets were combined, the accuracy increased from 75.5% to 84.7%, demonstrating that the proposed method is highly robust under various conditions. © 2024 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Logistics distribution and inspection of unmanned vehicles based on improved YOLOv5 9

Logistics distribution and inspection of unmanned vehicles b...

引用

9th International Conference on Electromechanical Control technology and Transportation, ICECTT 2024

作者： Hu, Jiayu Sun, Zhijin Li, Dalin School of Computer Science Zhuhai College of Science and Technology Zhuhai China School of Electronic Information Engineering Zhuhai College of Science and Technology Zhuhai China College of Computer Science and Technology Jilin University Changchun China

ISBN: (纸本)9781510682399

With the continuous development and application of unmanned vehicle technology, more and more unmanned vehicles must work in a variety of different bad weather and environmental conditions, which also brings higher requirements to the research of unmanned vehicle perception systems. Especially in foggy days or other severe weather conditions, the perception of complex road states becomes more difficult. Therefore, to improve the perception effect of unmanned vehicles in bad weather, this paper proposes a multi-functional unmanned vehicle visual perception system based on YOLOv5. This paper proposes three aspects to improve the problem of the perception effect of unmanned vehicles in bad weather. First, the model balance between computational efficiency and accuracy is improved by including the Ghost Bottleneck module. Secondly, the CBATM module is used to enhance the target perception ability of the model, especially the detection accuracy in foggy scenarios. Finally, the MSR algorithm is combined to enhance the robustness of the model in foggy scenarios and improve the ability of the model to perceive targets in a complex environment. This YOLOv5-based multi-functional unmanned vehicle visual perception system has a wide application prospect in the application of multifunctional unmanned vehicles integrating distribution and inspection and provides strong support for the realization of intelligent perception and decision-making. © 2024 SPIE.

关键词： Medical imaging

来源：评论

学校读者我要写书评

暂无评论

Partially training, isolating and unlearning, mitigating backdoor attack

Partially training, isolating and unlearning, mitigating bac...

引用

2024 IEEE International Conference on Big Data, BigData 2024

作者： Li, Yong Gao, Han Chen, Hao Wang, Yiming Yu, Chun School of Computer Science and Engineering Changchun University of Technology Changchun130012 China Ai Research Institute Changchun University of Technology Changchun130012 China School of Computer Science and Technology Jilin University Changchun130012 China Non-Commissioned Officer School Army Academy of Armored Forces Changchun130051 China

ISBN: (纸本)9798350362480

Neural networks are widely known to be vulnerable to backdoor attacks, witch are full of threats, a method that poisons a portion of the training data to make the target model perform well on normal data sets, while outputting attacker-specified or random categories on the poisoned samples. Nowadays, poisoned samples are becoming more and more similar to corresponding normal samples, even the human eye cannot easily distinguish them and the accuracy of models carrying backdoors on normal samples is no different from that of clean models witch also leads to the inability to determine whether a model carries a backdoor based on its accuracy. In this article, by observing the characteristics of backdoor attacks, We provide a new model training method, Partial Training(PT), that freezes part of the model to train a model that can isolate suspicious samples. Then, on this basis, a clean model is fine-tune to resist backdoor attacks. We used three datasets and four backdoor attack methods to conduct our experiments. The experiments showed that PT performed very well in defending against backdoor attacks mentioned above. © 2024 IEEE.

关键词： Adversary Attack Backdoor Attack

来源：评论

学校读者我要写书评

暂无评论

LMSPNet: Improved Lightweight Network for Multi-Person Sitting Posture Recognition 3

LMSPNet: Improved Lightweight Network for Multi-Person Sitti...

引用

3rd IEEE International Conference on computer Communication and Artificial Intelligence, CCAI 2023

作者： Jiao, Shuyang Xiao, Yubin Wu, Xuan Liang, Yanchun Liang, Yi Zhou, You Jilin University Computer Science and Technology Changchun China Zhuhai College of Science and Technology School of Computer Science Zhuhai China

ISBN: (纸本)9798350335262

Incorrect sitting posture may lead to health problems. Therefore, effective sitting posture recognition can remind individuals to maintain correct sitting posture and reduce discomfort. Traditional methods for sitting posture recognition have limitations in terms of high cost and slow inference speed. To address these issues, we propose a novel model called LMSPNet for multi-person sitting posture recognition. This model first employs the Light Convolution Core (LCC) to reduce the complexity of the model and then introduces Convolutional Block Attention Module (CBAM) to adaptively adjust the receptive field in the neural network to capture global contextual information, thereby enabling the model to better learn relationships between different channels. We construct the first human sitting posture dataset to evaluate the performance of LMSPNet. Experimental results demonstrate that, compared to the baseline models, our LMSPNet achieves state-of-the-art results with an accuracy of 99.57%. Therefore, our model is expected to become a powerful tool for multi-person sitting posture recognition. © 2023 IEEE.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

LRS: Enhancing Adversarial Transferability through Lipschitz Regularized Surrogate 38

LRS: Enhancing Adversarial Transferability through Lipschitz...

引用

38th AAAI Conference on Artificial Intelligence, AAAI 2024

作者： Wu, Tao Luo, Tie Wunsch, Donald C. Department of Computer Science Missouri University of Science and Technology United States Department of Electrical and Computer Engineering Missouri University of Science and Technology United States

ISBN: (纸本)1577358872

The transferability of adversarial examples is of central importance to transfer-based black-box adversarial attacks. Previous works for generating transferable adversarial examples focus on attacking given pretrained surrogate models while the connections between surrogate models and adversarial trasferability have been overlooked. In this paper, we propose Lipschitz Regularized Surrogate (LRS) for transfer-based black-box attacks, a novel approach that transforms surrogate models towards favorable adversarial transferability. Using such transformed surrogate models, any existing transfer-based black-box attack can run without any change, yet achieving much better performance. Specifically, we impose Lipschitz regularization on the loss landscape of surrogate models to enable a smoother and more controlled optimization process for generating more transferable adversarial examples. In addition, this paper also sheds light on the connection between the inner properties of surrogate models and adversarial transferability, where three factors are identified: smaller local Lipschitz constant, smoother loss landscape, and stronger adversarial robustness. We evaluate our proposed LRS approach by attacking state-of-the-art standard deep neural networks and defense models. The results demonstrate significant improvement on the attack success rates and transferability. Our code is available at https://***/TrustAIoT/LRS. Copyright © 2024, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Fuzzy validation of Taguchi-based convolutional fuzzy neural classifier for lung cancer imaging

引用

Multimedia Tools and Applications 2024年第41期83卷 89415-89437页

作者： Chang, Tsang-Chuan Lin, Cheng-Jian Yang, Tang-Yun Department of Intelligent Production Engineering National Taichung University of Science and Technology Taichung404336 Taiwan Department of Computer Science and Information Engineering National Chin-Yi University of Technology Taichung411030 Taiwan

While deep learning technology is widely used in the field of image classification and recognition, parameter setting for convolutional neural networks is complex, and a high number of parameters make the technology difficult to apply in practice. Therefore, this study proposes a Taguchi-based convolutional fuzzy neural classifier (T-CFNC) to classify computed tomography (CT) images. Two layers of convolution and pooling are used to extract features of the input images, and a fuzzy neural network is used to replace fully-connected neural networks to reduce the number of model parameters. To reduce cost and time, the Taguchi experimental design method determines the optimal combination of model parameters with the minimal number of experiments. The SPIE-AAPM lung CT challenge dataset was used to validate the proposed T-CFNC model. Experimental results indicate accuracy of 99.95%, a true positive rate of 99.97%, and a true negative rate of 99.94%. While the confusion matrix is commonly applied to evaluate model performance, its accuracy varies with the quality of model training, and it is easily affected by extreme values. Single and average values also incur the possibility of misjudgment. We therefore further propose fuzzy validation for performance evaluation. Results confirm the superiority of the proposed T-CFNC model in terms of lung cancer image classification over the traditional CFNC model. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

Optimizing Broadcast Utilization for Efficient Disaster Management Using Wireless Ad Hoc Networks and Novel Energy-Saving Algorithms

International Journal of Interactive Mobile Technologies

引用

International Journal of Interactive Mobile Technologies 2024年第20期18卷 142-156页

作者： Alzyoud, Faisal Y. Tarawneh, Monther Almaghthawi, Ahmed Altalidi, Askar Asiri, Layla Alrehaili, Meaad Department of Computer Science Faculty of Information Technology Isra University Amman Jordan Department of Information Technology College of Information Technology and Communications Tafila TechnicaUniversity Tafila Jordan Department of Computer Science College of Science & Art at Mahayil King Khalid University Abha Saudi Arabia Department of Information Science College of Science & Art at Mahayil King Khalid University Abha Saudi Arabia Departement of Computer Information System Applied college at Mahayil King Khalid University Abha Saudi Arabia Department of Computer Sciences and Artificial Intelligence Collège of Computer Science and Engineering University of Jeddah Jeddah Saudi Arabia

Globalization, industry, and population growth is considered the main reasons for climate change, so when some natural calamity or disaster occurs in any area, it leads to isolating this area from the rest of the region. Wireless ad hoc networks are important in disaster management to reduce losses and costs as they reduce the response, aiding time to save lives. However, they suffer from limited energy, which controls the lifetime of the sensor. Different methods are applied to save energy and increase battery life, such as sleep-awake schedules which assume similar initial energy for all nodes. Domestic partition in a unit disk graph is the production of the maximum possible number of such disjoint dominants. However, it assumes different initial energy levels for all nodes, which is another problem. In this paper, a new broadcast approach based on minimum set data broadcasting is proposed to mini-mize the number of broadcasting messages to increase the battery life during disasters and enhance rescue operations. The proposed minimum rebroadcast algorithm (MSRA) method balances coverage and utilization efficiency by dynamically adjusting node selection based on network conditions. The results show that MSRA significantly reduces power consumption and transmission latency, opening the way to find a new solution for disaster management using wireless ad hoc networks. © 2024 by the authors of this article.

关键词： Mobile ad hoc networks

来源：评论

学校读者我要写书评

暂无评论

High-resolution temperature and salinity model analysis using support vector regression

引用

Journal of Ambient Intelligence and Humanized Computing 2024年第2期15卷 1517-1525页

作者： Jiang, Yu Zhang, Tong Gou, Yu He, Lili Bai, Hongtao Hu, Chengquan College of Computer Science and Technology Jilin University Changchun130012 China Symbol Computation and Knowledge Engineer of Ministry of Education Jilin University Changchun130012 China

Temperature and salinity in marine data has been studied widely to enhance the marine environmental trend analysis. This work reports a new methodology for deriving high resolution monthly averages of temperature and salinity fields for the Berkeley Canyon based on the use of a support vector regression model. The data used in this paper is from WOA13, BOA Argo and ONC observations. The experimental results show that RBF Kernel function is satisfied with huge, complex marine data. The model is of great robust and adaptability. High resolution climatological means are critical for discerning oceanic features that are of great importance not just to climate systems but also to nutrient cycling and biological habitat. The method will be satisfied with other marine area. © Springer-Verlag GmbH Germany, part of Springer Nature 2018.

关键词： Regression analysis

来源：评论

学校读者我要写书评

暂无评论

3D Structure Estimation of Room Environment Using Semantic Segmentation and Model Fitting

3D Structure Estimation of Room Environment Using Semantic S...

引用

2024 IEEE International Workshop on Metrology for Living Environment, MetroLivEnv 2024

作者： Morioka, Junya Miyamoto, Ryusuke Graduate School of Science and Technology Meiji University Department of Computer Science Kawasaki Japan School of Science and Technology Meiji University Department of Computer Science Kawasaki Japan

ISBN: (纸本)9798350385014

Recent advances in VR devices and technology have increased the demand for technologies that reflect real space in virtual space. In particular, technologies that reflect the indoor environment in a virtual space are used in various fields such as architecture, interior design, and robotics. To reflect a real space in a virtual space, estimating the 3D structure of the real space is significant. Yet, this requires expensive equipment and poses challenges in computational complexity and ease of measurement. In this study, we propose a method of estimating 3D structure of indoor environments using a single image. Existing methods generally use depth estimation, but in this study, we introduce a new approach using semantic segmentation and rectangular fitting. Experimental results with 3D CG show that the proposed method outperforms existing methods in accuracy and competitive results. © 2024 IEEE.

关键词： Semantics

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：