检索结果-内蒙古大学图书馆

Research on rectal tumor identification method by convolutional neural network based on multi-feature fusion

International Journal for engineering Modelling 2021年第2期34卷 31-41页

作者： Liang, Zhuang Wu, Jiansheng School of Computer Science and Software Engineering University of Science and Technology Liaoning Anshan114044 China

Aiming at the obscure features of tumors in rectal CT images and their complexity, this paper proposes a multi-feature fusion-based convolutional neural network rectal tumor recognition method and uses it to model rectal tumors for classification experiments. This method extracts the convolutional features from rectal CT images using Alexnet, VGG16, ResNet, and DenseNet, respectively. At the same time, local features such as histogram of oriented gradient, local binary pattern, and HU moment invariants are extracted from this image. The above local features are fused linearly with the convolutional features. Then we put the new fused features into the fully connected layer for image classification. The experimental results finally reached the accuracy rates of 92.6 %, 93.1 %, 91.7 %, and 91.1 %, respectively. Comparative experiments show that this method improves the accuracy of rectal tumor recognition. © 2021, University of Split. All rights reserved.

关键词： Tumors

来源：评论

学校读者我要写书评

暂无评论

Enhanced Diagnosis of Diabetic Retinopathy Utilizing Convolutional Neural Networks: A Deep Learning Approach

Enhanced Diagnosis of Diabetic Retinopathy Utilizing Convolu...

引用

2024 International Conference on Artificial Intelligence and Emerging Technology, Global AI Summit 2024

作者： Singh, Neelam Srivastava, Rohit Mahdi, Hussain Falih Sati, Subhangi Joshi, Purvika Sar, Ayan Singh, Devansh Choudhury, Tanupriya Graphic Era Deemed to be University Department of Computer Science and Engineering Uttarakhand Dehradun248002 India School of Computer Science University of Petroleum and Energy Studies Bidholi Uttarakhand Dehradun248007 India University of Diyala Department of Computer and Software Engineering Baquba Iraq Thapar University Punjab Patiala147001 India

ISBN: (纸本)9798350379716

Diabetic Retinopathy (DR) is one of the most severe eye complications associated with diabetes, hence the need to diagnose and treat at the initial stages. The current approaches of DR detection that use ophthalmoscopic inspection of the optic fundus images are subjective and take an ample amount of time. This paper focuses on the application of Convolutional Neural Networks (CNNs) in automating DR detection as well as its classification process using large archive annotated datasets for training and testing. The structure of our CNN excels in capturing stipples of the retinal image and reaches a level higher than 90% accuracy while remaining sensitive and specific. By presenting feature maps and class activation maps, there is an improved credibility as the process the model goes through is explained. The analysis shows that using the CNN-based system can bring notable benefits in the form of early detection, a decrease in the amount of work performed by ophthalmologists, and enhancement of patient outcomes. Future work seeks to implement this model in clinical practice and work to overcome problems such as data privacy and real-time processing, thus demonstrating the revolutionary impact that deep learning holds for ophthalmology. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

A Digital Twin Based Framework for Real-Time Machine Condition Monitoring 19

A Digital Twin Based Framework for Real-Time Machine Conditi...

引用

19th IEEE International Conference on Automation science and engineering, CASE 2023

作者： Chen, Zixiao Choudhury, Madhurjya Dev Blincoe, Kelly Dhupia, Jaspreet Singh University of Auckland Department of Mechanical and Mechatronics Engineering New Zealand School of Engineering and Computer Science Victoria University of Wellington New Zealand University of Auckland Department of Electrical Computer and Software Engineering New Zealand

ISBN: (纸本)9798350320695

Condition Monitoring (CM) is an important approach to extending the life of complex equipment by forecasting the outcome of an event before catastrophic failure occurs. Recent advancements in digital twins (DT) offer additional benefits to machine condition monitoring. In this study, a framework based on DT for real-time condition monitoring of industrial machines is proposed. The multi-layer DT framework consists of a physical entity (PE), virtual equipment (VE), edge device, fidelity service and digital twin services. The virtual equipment is a replica of the physical entity or the monitored machine. It also contains a cloud platform to store data online and an application to interface with the cloud enabling users to check the data remotely. The fidelity service ensures conformity between the PE and the VE. The digital service provides optimal operation and maintenance schedules based on the data from both physical and virtual spaces. The integration of the edge layer enables real-time handling of high-frequency machine data for effective health monitoring. The validity of the proposed framework is demonstrated with a case study based on monitoring a critical component of an industrial drivetrain test rig. The features of the framework allow end-users to visualize the component's real-time health status remotely. © 2023 IEEE.

关键词： Condition monitoring

来源：评论

学校读者我要写书评

暂无评论

Enhancing Federated Learning by One-Shot Transferring of Intermediate Features from Clients 10

Enhancing Federated Learning by One-Shot Transferring of Int...

引用

10th IEEE International Conference on Data science and Advanced Analytics, DSAA 2023

作者： Deng, Youxingzhu Zhou, Yipeng Liu, Gang Wang, Jessie Hui Shui, Yu Shenzhen University College of Computer Science and Software Engineering Shenzhen China Macquarie University School of Computing Faculty of Science and Engineering Sydney Australia Institute for Network Sciences and Cyberspace Tsinghua University Peking China University of Technology Sydney School of Computer Science Sydney Australia

ISBN: (纸本)9798350345032

Federated learning (FL) is an emerging paradigm using a parameter server (PS) to coordinate multiple decentralized clients for training a common model without exposing their raw data. Despite its amazing capability in preserving data privacy, FL confronts two significant challenges that have not been sufficiently addressed by existing works, which are: 1) heterogeneous data distributed on clients given that the PS cannot alter data locations. 2) limited computation resources of FL clients who may conduct model training with mobile devices. To tackle these challenges, we propose a novel federated one-shot transferring of intermediate features (FedOTF) algorithm. More specifically, FedOTF consists of two stages: feature extraction and model reconstruction. To overcome challenge 1, clients in stage 1 only collaboratively train a small model in a federated manner, with the objective to extract features. In stage 2, intermediate features (generated by the small model trained by stage 1) are transferred from clients to the PS so that a large model can be trained to overcome challenge 2. In addition, the design of FedOTF is robust, which can flexibly diminish the number of communication rounds of stage 1 when network capacity is limited, and reduce the amount of exposed features when privacy is concerned. To verify the superiority of FedOTF, we conduct comprehensive experiments with real datasets. The experiment results demonstrate that FedOTF can significantly improve the model utility of FL because a better model can be finally obtained by the PS without incurring heavy computational load on clients. Besides, we conduct robustness evaluation of FedOTF which can achieve stable performance when varying network capacity and privacy requirement. © 2023 IEEE.

关键词： Privacy-preserving techniques

来源：评论

学校读者我要写书评

暂无评论

Three-dimensional reconstruction of industrial parts from a single image

引用

Visual Computing for Industry,Biomedicine,and Art 2024年第1期7卷 340-351页

作者： Zhenxing Xu Aizeng Wang Fei Hou Gang Zhao School of Mechanical Engineering and Automation Beihang UniversityBeijing 100191China Key Laboratory of Aeronautics Smart Manufacturing Beihang UniversityBeijing 100191China State Key Laboratory of Computer Science Institute of SoftwareChinese Academy of SciencesBeijing 100190China

This study proposes an image-based three-dimensional(3D)vector reconstruction of industrial parts that can gener-ate non-uniform rational B-splines(NURBS)surfaces with high fidelity and *** contributions of this study include three parts:first,a dataset of two-dimensional images is constructed for typical industrial parts,including hex-agonal head bolts,cylindrical gears,shoulder rings,hexagonal nuts,and cylindrical roller bearings;second,a deep learning algorithm is developed for parameter extraction of 3D industrial parts,which can determine the final 3D parameters and pose information of the reconstructed model using two new nets,CAD-ClassNet and CAD-ReconNet;and finally,a 3D vector shape reconstruction of mechanical parts is presented to generate NURBS from the obtained shape *** final reconstructed models show that the proposed approach is highly accurate,efficient,and practical.

关键词： Three-dimensional reconstruction Non-uniform rational B-splines Industrial parts Deep learning

来源：评论

学校读者我要写书评

暂无评论

TFSG: A Self-Supervised Guided Graph Learning Method for Traffic Flow Prediction

引用

IEEE Internet of Things Journal 2025年第13期12卷 24345-24357页

作者： Zheng, Yingran Luo, Chao Shao, Rui Shandong Normal University School of Information Science and Engineering Jinan250014 China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Jinan250014 China China Mobile Shandong Co. Ltd center of network optimization Jinan250014 China

With the rapid development of the Internet of Things, a large amount of real-time traffic data has been collected. How to fully utilize these traffic data for rational planning of intelligent transportation systems is an important issue. Graph learning has proved to be potential tools for modeling complex spatiotemporal relationships in addressing traffic flow prediction problems. Existing methods usually relied on the graph structure constructed from predefined distance information between different regions. However, affected by factors such as road capacity, traffic rules etc., distance bias is non-trivial to affect traffic flow in reality. Besides, the distance based static transportation network structure is also difficult to describe the dynamic and uncertainty of road traffic in complex situations. To address the aforementioned issues, this article proposes an enhanced graph generation approach for traffic flow prediction, which is based on Temporal-Frequency Self-Supervised Guided Graph Learning (TFSG). The method progressively guides the feature latent representation to approach the target latent representation in both the time domain and frequency domain. Meanwhile, a multi-head self-attention mechanism is introduced to capture dependencies among different time series patches, constructing an enhanced graph structure for downstream prediction tasks. Experimental results first independently evaluated the qualities of the generated graphs. Subsequently, by using five public traffic datasets, the feasibility and effectiveness of the proposed method are further verified. © 2014 IEEE.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Intelligent Tutoring for Large-Scale Personalized Programming Learning Based on Knowledge Graph 53

Intelligent Tutoring for Large-Scale Personalized Programmin...

引用

53rd IEEE ASEE Frontiers in Education International Conference, FIE 2023

作者： Li, Ying Qiu, Jincheng Yang, Runze Zhu, Tongyu Sheng, Hao Gui, ShiJie Liang, Yu School of Computer Science and Engineering Beihang University Beijing China Beijing Engineering Research Center for IoT Software and Systems Beijing University of Technology Beijing China

ISBN: (纸本)9798350336429

The development of learning analytics technology has created the necessary conditions for the implementation of large-scale personalized education. By taking advantage of the current advancements available, we can adopt a deep integration of 'artificial intelligence + education', take educational big data as the research object, aim at achieving precise teaching, and use cognitive diagnosis as a means to construct an intelligent guidance model. The model will conform to a learner's cognitive patterns and curriculum characteristics to meet his personalized learning needs under different knowledge states. Firstly, use knowledge map and abstract syntax trees to extract the characteristics of teaching objectives from questions and standard programing codes. Secondly, Use 'target characteristics - learning characteristics - learning sequence - learning effect' cognitive diagnosis model to assess student achievement in learning. Finally, according to the correlation between knowledge points, the initial learning path consisting of all relevant knowledge points is identified from a knowledge map. And the most optimal learning path is selected based on the importance of knowledge points and student's learning situation. There are three innovation points in this paper. First, the traditional method which only focuses on discovering the one-dimensional knowledge features implicated in the questions, is extended to construct a knowledge-ability multidimensional feature matrix. Secondly, it optimized static cognitive diagnosis into dynamic cognitive arbitrariness, which has a modeling capability of dynamic temporal sequence. Thirdly, it regarded the correlation between knowledge points, and the importance of knowledge points as the heuristic conditions for constructing a learning path. © 2023 IEEE.

关键词： Cognitive Diagnosis Intelligent Guidance Model Knowledge Graph Personalized Learning Path

来源：评论

学校读者我要写书评

暂无评论

Deep-Learning-Assisted Complete Targets Coverage in Energy-Harvesting IoT Networks

引用

IEEE Internet of Things Journal 2025年第11期12卷 17780-17790页

作者： Wang, Kunsheng Yang, Changlin Chin, Kwan-Wu Xian, Jun Sun Yat-sen University School of Mathematics Guangdong Guangzhou510275 China Sun Yat-sen University School of Software Engineering Zhuhai519082 China University of Wollongong School of Electrical Computer and Telecommunications Engineering WollongongNSW2522 Australia Minnan Normal University School of Mathematics and Statistics Zhangzhou363000 China Sun Yat-sen University Guangdong Provincial Key Laboratory of Computational Science Guangzhou510275 China

Complete targets coverage is required by many Internet of Things (IoT) applications. In this respect, an important goal is to maximize the number of time slots with complete targets coverage. Achieving such coverage is challenging when devices experience spatio-temporal energy arrivals. To this end, this article outlines a deep learning assisted approach that has an offline stage whereby it determines and stores an exhaustive collection of optimal activation schedules based the energy levels and arrivals of devices. In addition, it presents a network partitioning and training strategy, and outlines an algorithm to mend coverage holes in its online stage. We have compared the proposed approach with the optimal solution, and also a state-of-the-art heuristic algorithm. The results show that our solution achieves 94% of the optimal coverage lifetime. Moreover, the proposed approach has a 35% smaller optimality gap as compared with the said heuristic algorithm. © 2014 IEEE.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Multi-turn dialogue model based on the improved hierarchical recurrent attention network

引用

International Journal for engineering Modelling 2021年第2期34卷 17-29页

作者： Miao, Jiawei Wu, Jiansheng School of Computer Science and Software Engineering University of Science and Technology Liaoning Anshan114044 China

When considering the multi-turn dialogue systems, the model needs to generate a natural and contextual response. At present, HRAN, one of the most advanced models for multi-turn dialogue problems, uses a hierarchical recurrent encoder-decoder combined with a hierarchical attention mechanism. However, for complex conversations, the traditional attention-based RNN does not fully understand the context, which results in attention to the wrong context that generates irrelevant responses. To solve this problem, we proposed an improved hierarchical recurrent attention network, a self-attention network (HSAN), instead of RNN, to learn word representations and utterances representations. Empirical studies on both Chinese and English datasets show that the proposed model has achieved significant improvement. © 2021, University of Split. All rights reserved.

关键词： Speech processing

来源：评论

学校读者我要写书评

暂无评论

Personalized Health Assistant with Reinforcement Learning 1

Personalized Health Assistant with Reinforcement Learning

引用

1st IEEE International Conference on Artificial Intelligence for Medicine, Health and Care, AIMHC 2024

作者： Jin, Jennifer Kim, Mira Kim, Soo Dong Computer Science and Engineering California State University San BernardinoCA92407 United States Department of Computer Science California State University FullertonCA92831 United States School of Software Soongsil University Seoul Korea Republic of

ISBN: (纸本)9798350371987

Integration of medical IoT devices with machine learning presents significant advantages for personalized healthcare. This research is to develop a software solution which leverages smart scale devices to measure health indexes, recommend personalized health-promoting activities. We designed the system with Q-learning elements and tailored the Q-learning policy to keep learning the personal effectiveness of health-promoting activities through users’ episodes. The system was implemented in Python, and the experiments demonstrated a substantial average enhancement of 25% across health indexes and 53% in optimizing the performance of the recommender policy. ©2024 IEEE.

关键词： Reinforcement learning

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：