检索结果-内蒙古大学图书馆

5th International Conference on computer information Science and Artificial Intelligence, CISAI 2022

作者： Li, JiuHui Yao, JiaMin Xie, XiaoPeng Luo, YiMing Huang, ZhiPeng Chen, ZuDong School of Computer Science and Cyber Engineering Guangzhou University Guangzhou510006 China School of Mathematics and Information Science Guangzhou University Guangzhou510006 China School of Mechanical and Electrical Engineering Guangzhou University Guangzhou510006 China

ISBN: (纸本)9781510662520

In response to the problems of complex implementation, low accuracy, poor applicability and high latency of some contemporary human fall detection algorithms that do not achieve real-time results, this paper proposes a fall detection based on Tiny-YOLO target detection algorithm, Kalman target trajectory tracking, Alphapose human pose recognition and spatio-temporal convolutional network, a combination of multiple algorithms. It solves the problem that human target features are difficult to extract and subsequent tracking is lost. And the training speed is accelerated to achieve fast detection of human falls through video and improve the accuracy. The algorithm is tested on the publicly available datasets UR Fall Detection Dataset and Lei2 fall Detection Dataset, and the experimental results prove that the algorithm in this paper has high detection accuracy and effectively reduces the false detection rate. © 2023 SPIE.

关键词： Fall detection

来源：评论

学校读者我要写书评

暂无评论

A Systematic Review on Deepfake Technology

A Systematic Review on Deepfake Technology

引用

International Conference on Data Analytics and Management, ICDAM 2022

作者： Khan, Ihtiram Raza Aisha, Saman Kumar, Deepak Mufti, Tabish Department of Computer Science and Engineering School of Engineering Sciences and Technology Jamia Hamdard New Delhi India Amity Institute of Information Technology Amity University Noida India

ISBN: (纸本)9789811976148

The twenty-first century has seen technology become an integral part of human survival. Living standards have been affected as the technology continues to advance. Deep fake is a deep learning-powered application that recently appeared on the market. It allows to make fake images and videos that people cannot discern from the real ones and is a recent technique that allows swapping two identities within a single video. A wide range of factors, including communities, organizations, security, religions, democratic processes as well as personal lives, are being impacted by deep fakes. In the case of images and videos being presented to an organization as evidence, if the images or videos are altered, the entire truth will be transformed into a lie. It is inevitable that with every new and beneficial technology, some of its adverse effects will follow, causing world problems. There have been several instances in which deep fakes have alarmed the entire network. During the last few years, the number of altered images and videos has grown exponentially, posing a threat to society. Additionally, this paper explores where deep fakes are used, the impacts made by them, as well as the challenges and difficulties associated with deep fakes in this rapidly developing society. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

ANN Based Fake Profile Detection and Categorization Using Premetric Paradigms On Instagram

ANN Based Fake Profile Detection and Categorization Using Pr...

引用

2023 Innovations in Power and Advanced Computing Technologies, i-PACT 2023

作者： Fathima, Afifa Salsabil Reema, Syeda Ahmed, Syed Thouheed School of Computer Science and Engineering REVA University Bangalore India School of Computing and Information Technology REVA University Bangalore India Indian Institute of Technology Hyderabad India

ISBN: (纸本)9798350325188

This study, focused on the behavioral analysis for the identification of social media accounts, endeavors to discern counterfeit profiles within social media networks by employing machine learning methodologies. The investigation is underpinned by a comprehensive dataset encompassing various user behavior-related attributes, including but not limited to the count of followers, followings, posts, private account status, and other relevant factors. The data is subjected to rigorous preprocessing procedures, which involve feature scaling and the division of the dataset into distinct training and testing subsets. Subsequently, a neural network model is trained utilizing the training dataset and rigorously evaluated using performance metrics such as accuracy and loss on the testing dataset. Additionally, the study delves into an examination of the confusion matrix to provide insight into the classification performance and offers a visual representation of the distribution of genuine and counterfeit accounts. The experimental results substantiate the model's efficacy in successfully identifying fake accounts through an in-depth analysis of their behavioral patterns. © 2023 IEEE.

关键词： ANN Confusion Matrix data analysis data modelling Data preprocessing fake account Feature extraction Heatmap Instagram neural network

来源：评论

学校读者我要写书评

暂无评论

Advanced Deep Learning Framework for Automated Lung Cancer Detection Using DeepResNet 6

Advanced Deep Learning Framework for Automated Lung Cancer D...

引用

6th International Conference on Cybernetics, Cognition and Machine Learning Applications, ICCCMLA 2024

作者： Ahammed, Syed Zaheer Baskar, Radhika Nalinipriya, G. Saveetha Institute of Medical and Technical Sciences Saveetha University Saveetha School of Engineering Department of Computer Science & Engineering Chennai India Saveetha Institute of Medical and Technical Sciences Saveetha University Saveetha School of Engineering Department of Electronics and Communication Engineering Chennai India Saveetha Engineering College Department of Information Technology Tamilnadu Chennai India

ISBN: (纸本)9798331505790

This study presents a sophisticated framework for automatically identifying lung cancer nodules from CT scans using a unique DeepResNet model. Our approach combines the segmentation abilities of UNet with the feature extraction capabilities of ResNet, resulting in accurate identification and classification of pulmonary nodules. The model's performance is assessed using the LIDC-IDRI dataset, showcasing exceptional metrics: an accuracy of 9 9. 8 9 %, an F1 score of 9 9. 8 3 2 %, a sensitivity of 100 %, a precision of 99.832 %, and a specificity of 9 9. 7 8 8 3 %. This blend deep learning architecture improves both the accuracy of diagnoses and the ability to detect subtle characteristics of nodules that are important for diagnosing early-stage lung cancer. © 2024 IEEE.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

CLIP4Video-Sampling: Global Semantics-Guided Multi-Granularity Frame Sampling for Video-Text Retrieval

引用

Journal of computer and Communications 2024年第11期12卷 26-36页

作者： Tao Zhang Yu Zhang School of Computer Science and Engineering The Key Lab of Computer Network and Information Integration (Ministry of Education) Southeast University Nanjing China

Video-text retrieval (VTR) is an essential task in multimodal learning, aiming to bridge the semantic gap between visual and textual data. Effective video frame sampling plays a crucial role in improving retrieval performance, as it determines the quality of the visual content representation. Traditional sampling methods, such as uniform sampling and optical flow-based techniques, often fail to capture the full semantic range of videos, leading to redundancy and inefficiencies. In this work, we propose CLIP4Video-Sampling: Global Semantics-Guided Multi-Granularity Frame Sampling for Video-Text Retrieval, a global semantics-guided multi-granularity frame sampling strategy designed to optimize both computational efficiency and retrieval accuracy. By integrating multi-scale global and local temporal sampling and leveraging the CLIP (Contrastive Language-Image Pre-training) model’s powerful feature extraction capabilities, our method significantly outperforms existing approaches in both zero-shot and fine-tuned video-text retrieval tasks on popular datasets. CLIP4Video-Sampling reduces redundancy, ensures keyframe coverage, and serves as an adaptable pre-processing module for multimodal models.

关键词： Video Sampling Multimodal Large Language Model Text-Video Retrieval CLIP Model

来源：评论

学校读者我要写书评

暂无评论

Secure beamforming designs for maximizing secrecy sum rate in MISO-NOMA networks

引用

Digital Communications and Networks 2025年第1期11卷 83-91页

作者： Yanbo Zhang Zheng Yang Jingjing Cui Xianfu Lei Yi Wu Jun Zhang Chao Fang Zhiguo Ding College of Photonic and Electronic Engineering Fujian Normal UniversityFuzhou 350007China School of Electronics and Computer Science University of SouthamptonSouthampton SO171BJUK Institute of Mobile Communications Southwest Jiaotong UniversityChengdu 610031China Jiangsu Key Laboratory of Wireless Communications Nanjing University of Posts and TelecommunicationsNanjing 210003China Faculty of Information Technology Beijing University of TechnologyBeijingChina School of Electrical and Electronics Engineering The University of ManchesterManchester M139PLUK

In this paper,the application of Non-Orthogonal Multiple Access(NOMA)is investigated in a multiple-input single-output network consisting of multiple legitimate users and a potential *** support secure transmissions from legitimate users,two NOMA Secrecy Sum Rate Transmit Beam Forming(NOMA-SSR-TBF)schemes are proposed to maximise the SSR of a Base Station(BS)with sufficient and insufficient transmit *** BS with sufficient transmit power,an artificial jamming beamforming design scheme is proposed to disrupt the potential eavesdropping without impacting the legitimate *** addition,for BS with insufficient transmit power,a modified successive interference cancellation decoding sequence is used to reduce the impact of artificial jamming on legitimate *** specifically,iterative algorithm for the successive convex approximation are provided to jointly optimise the vectors of transmit beamforming and artificial *** results demonstrate that the proposed NOMA-SSR-TBF schemes outperforms the existing works,such as the maximized artificial jamming power scheme,the maximized artificial jamming power scheme with artificial jamming beamforming design and maximized secrecy sum rate scheme without artificial jamming beamforming design.

关键词： NOMA Secure transmission Artificial jamming Beamforming design

来源：评论

学校读者我要写书评

暂无评论

Ensemble Learning for Multi-Class Classification in Cervical Cancer Risk Factor Analysis: An Empirical Computational Approach 3

Ensemble Learning for Multi-Class Classification in Cervical...

引用

3rd IEEE International Conference on ICT in Business Industry and Government, ICTBIG 2023

作者： Chaudhuri, Sumantra Suneetha, K. Yadav, Rakesh Kumar Presidency University Department of Electronics and Communication Engineering Karnataka Bangalore India Department of Computer Science and Information Technology Bangalore India Maharishi School of Engineering & Technology Maharishi University of Information Technology Uttar Pradesh India

ISBN: (纸本)9798350343274

One cancer that affects women of all ages frequently and is also particularly fatal is cervical cancer. It is fatal because it cannot be anticipated in the early stages when the patient could be treated properly. When cervical cancer cells begin to migrate to nearby organs including the liver, lymph nodes, and lungs, they block the tube that takes urine from the bladder and cause symptoms. Leg discomfort, exhaustion, weight loss, vaginal bleeding, bone fractures, and pelvic pain are among the symptoms in this latter stage. Human papillomavirus (HPV), smoking, numerous pregnancies, using hormonal contraception, and other risk factors include cervical cancer. Screening is a crucial step in the detection of cervical cancer. This study suggests a technique for categorizing cervical cancer risk variables into four separate classes using a multi-class approach. The ensemble learning approach aims to increase the suggested model's accuracy. This technique aids in enhancing the study of risk factors used to determine the likelihood of developing cervical cancer. A multi-layer feed-forward network is used to examine the suggested Artificial Bee colony (ABC) along with a deep learning method known as the LST Model for various screening methods. The gradient boosting method with Multi-Layer Perceptron (MLP) is found to produce a higher % for all performance measures using the suggested model. © 2023 IEEE.

关键词： Learning systems

来源：评论

学校读者我要写书评

暂无评论

EEG-Based Joint Semi-supervised Learning for Major Depressive Disorder Detection 26th

EEG-Based Joint Semi-supervised Learning for Major Depressi...

引用

26th International Conference on Pattern Recognition, ICPR 2022

作者： Chen, Tao Zheng, Tong Shi, Jinlong Guo, Yanrong Key Laboratory of Knowledge Engineering with Big Data Hefei University of Technology Hefei China School of Computer Science and Information Engineering Hefei University of Technology Hefei China

ISBN: (纸本)9783031376597

Major Depressive Disorder (MDD) has been a major mental disease in recent years, imposing huge negative impacts on both our society and individuals. The current clinical MDD detection methods, such as self-report scales and interviews, often suffer from their subjectiveness. In contrast, electroencephalography (EEG) signals reflect brain activity and can be used as an effective information source for MDD detection. However, the learning-based MDD detection methods with EEG data usually neglect the valuable information contained in the unlabeled data, therefore possibly suffering from the label scarcity issue. In this context, we propose a simple EEG-based Joint Semi-supervised Learning (JSL) model for the MDD detection task, which fully leverages the information from unlabeled data. Specifically, by equipping the unlabeled samples with the intermediately predicted pseudo labels, the accurate position of the class-specific centroids in the feature space can be gradually refined. Based on that, the auxiliary labels based on a heuristic centroid-based classifier can be obtained, which plays an effective complementary role in predicting the sample labels. The experiments on a public dataset demonstrate the effectivenes s of our method, showing an obvious performance gain via using the proposed semi-supervised strategy. © 2023, Springer Nature Switzerland AG.

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

Enhancing UAV Coverage Path Planning via Deep Reinforcement learning with Intrinsic Curiosity 10

Enhancing UAV Coverage Path Planning via Deep Reinforcement ...

引用

10th IEEE Smart World Congress, SWC 2024

作者： Zhou, Yinxin Chen, Yanxi Wang, Renyu Huang, Haiping Liu, Shenghao Deng, Xianjun Nanjing University of Posts and Telecommunications School of Computer Science Nanjing China Huazhong University of Science and Technology School of Electronic Information and Communications Wuhan China Huazhong University of Science and Technology School of Cyber Science and Engineering Wuhan China

ISBN: (纸本)9798331520861

Unmanned Aerial Vehicles (UAVs) are widely used in various fields due to their agility and versatility, but their limited energy supply necessitates efficient Coverage Path Planning (CPP). Traditional CPP methods are effective in structured environments but often fall short in more complex scenarios. This paper proposes a novel approach integrating advanced reinforcement learning techniques to enhance UAV path planning. Our method combines Double Deep Q-Networks (DDQN) with Long Short-Term Memory (LSTM) networks to improve decision-making and temporal information processing capabilities. Additionally, we incorporate an Intrinsic Curiosity Module (ICM) to drive exploration and Prioritized Experience Replay (PER) to focus on high-value experiences, accelerating learning and enhancing performance. The proposed framework demonstrates superior coverage efficiency and adaptability in simulations, achieving higher coverage ratios and reduced redundant paths compared to traditional methods. These innovations provide a robust and flexible solution for autonomous UAV navigation, offering a solid foundation for future research in CPP optimization. © 2024 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Deep Learning Method for Forecasting Reservoir Water Level from Sentinel-2 Satellite Images

引用

computers, Materials & Continua 2025年第6期83卷 4915-4937页

作者： Hoang Thi Minh Chau Tran Thi Ngan Nguyen Long Giang Tran Manh Tuan Tran Kim Chau Graduate University of Science and Technology Vietnam Academy of Science and TechnologyHanoi100000Vietnam Faculty of Information Technology University of Economics Technology for IndustriesHanoi100000Vietnam Artificial Intelligence Research Center VNU Information Technology InstituteVietnam National UniversityHanoi100000Vietnam International School Vietnam National UniversityHanoi100000Vietnam Institute of Information Technology Vietnam Academy of Science and TechnologyHanoi100000Vietnam Faculty of Computer Science and Engineering Thuyloi UniversityHanoi100000Vietnam Faculty of Water Resources Engineering Thuyloi UniversityHanoi100000Vietnam

Global climate change,along with the rapid increase of the population,has put significant pressure on water security.A water reservoir is an effective solution for adjusting and ensuring water *** particular,the reservoir water level is an essential physical indicator for the *** the reservoir water level effectively assists the managers in making decisions and plans related to reservoir management *** recent years,deep learning models have been widely applied to solve forecasting *** this study,we propose a novel hybrid deep learning model namely the YOLOv9_ConvLSTM that integrates YOLOv9,ConvLSTM,and linear interpolation to predict reservoir water *** utilizes data from Sentinel-2 satellite images,generated from visible spectrum bands(Red-Blue-Green)to reconstruct true-color reservoir *** is used as the optimization algorithm with the loss function being MSE(Mean Squared Error)to evaluate the model’s error during *** implemented and validated the proposed model using Sentinel-2 satellite imagery for the An Khe reservoir in *** assess its performance,we also conducted comparative experiments with other related models,including SegNet_ConvLSTM and UNet_ConvLSTM,on the same *** model performances were validated using k-fold cross-validation and ANOVA *** experimental results demonstrate that the YOLOv9_ConvLSTM model outperforms the compared *** has been seen that the proposed approach serves as a valuable tool for reservoir water level forecasting using satellite imagery that contributes to effective water resource management.

关键词： YOLOv9 ConvLSTM reservoir water level forecasting satellite images

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：