检索结果-内蒙古大学图书馆

Video classification for efficient data storage using deep learning: a comparison of sequential and simultaneous feature extraction methods

引用

Multimedia Tools and Applications 2025年第6期84卷 3071-3094页

作者： Dang Thi, Phuc Nguyen Dang, Hoang Thi Doan Huu, Phuc Dau Sy, Hieu Department of Computer Science Faculty of Information Technology Industrial University of Ho Chi Minh City 12 Nguyen Van Bao Ho Chi Minh City7000 Viet Nam Department of Applied Physics Faculty of Applied Science University of Technology - Viet Nam National University HCMC To Hien Thanh Ho Chi Minh City7000 Viet Nam

Optimizing camera information storage is a critical issue due to the increasing data volume and a large number of daily surveillance videos. In this study, we propose a deep learning-based system for efficient data storage. Videos captured by cameras are classified into four categories: no action, normal action, human action, and dangerous action. Videos without action or with normal action are stored temporarily and then deleted to save storage space. Videos with human action are stored for easy retrieval, while videos with dangerous action are promptly alerted to users. In the paper, we propose two approaches using deep learning models to address the video classification problem. The first approach is a separate approach, where pretrained CNN models extract features from video frame images. These features are then passed through RNN, Transformer models to extract relationships between them. The goal of this approach is to delve into extracting features of objects in the video. The proposed models include VGG16, InceptionV3 combined with LSTM, BiLSTM, Attention, and Vision Transformer. The next approach combines CNN and LSTM layers simultaneously through models like ConvLSTM and LRCN. This approach aims to help the model simultaneously extract object features and their relationships, with the goal of reducing model size, accelerating the training process, and increasing object recognition speed when deployed in the system. In Approach 1, we construct and refine network architectures such as VGG16+LSTM, VGG16+Attention+LSTM, VGG16+BiLSTM, VGG16+ViT, InceptionV3+LSTM, InceptionV3+Attention+LSTM, InceptionV3+BiLSTM. In Approach 2, we build a new network architecture based on the ConvLSTM and LRCN model. The training dataset, collected from real surveillance cameras, comprises 3315 videos labeled into four classes: no action (1018 videos), actions involving people (832 videos), dangerous actions (751 videos), and normal actions (714 videos). Experimental results show t

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Dynamic generative R-CNN

引用

Neural Computing and Applications 2025年第10期37卷 7107-7120页

作者： Saffarini, Rasha Khamayseh, Faisal Awwad, Yousef Sabha, Muath Eleyan, Derar IT Engineering Arab American University Jenin Palestine Department of Computer Science Palestine Polytechnic University Hebron Palestine Department of Computer System Engineering Palestine Technical University Tulkarm Palestine Department of Multimedia Technology Arab American University Jenin Palestine Department of Applied Computing Palestine Technical University Tulkarm Palestine

Different techniques have been developed for object detection and recognition. These techniques can be divided into single-shot and two-shot methods. Single-shot methods focus on real-time applications, while two-shot methods are used in applications requiring higher accuracy. However, different versions of the two-shot techniques produce limited results in terms of accuracy and speed, or both. Therefore, this study proposes a novel model called dynamic generative R-CNN (DGR-CNN) that reduces the number of proposed regions using a dynamic programming model that applies the graph similarity method over graph-based image segmentation. Additionally, the proposed model employs DCGAN technique to improve detection performance. DGR-CNN reduces the overall detection and classification time and enhances the detection accuracy. The PASCAL VOC2007 and MS COCO datasets were utilized to evaluate the model. The results showed that DGR-CNN significantly reduces the number of candidate regions compared to the selective search algorithm employed in R-CNN and fast R-CNN. Although fast R-CNN utilizes 2000 regions and faster R-CNN utilizes 300 regions, DGR-CNN reduces the number of regions to approximately 130. The mean average precision of the proposed method was 75.1% on the PASCAL VOC2007, while fast and faster R-CNN scored 66.9% and 69.9%, respectively. Moreover, the DGR-CNN model significantly improved the classification accuracy when tested on the MS COCO dataset, achieving an MAP of 68.76%, compared with 32.64% and 42.3% for fast and faster R-CNN. This increase in accuracy was achieved without significantly compromising the speed compared with faster R-CNN. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Automated Radiology Report Generation from Chest X-ray Scans Using Deep Learning 6th

Automated Radiology Report Generation from Chest X-ray Scans...

引用

6th International Conference on Recent Innovations in Computing, ICRIC 2023

作者： Veer, V. Druva Priya, S. Baghavathi Tamilselvi, M. Department of Computer Science and Engineering Amrita School of Computing AmritaVishwaVidyapeetham Chennai India Information Technology Department University of Technology and Applied Sciences Nizwa Oman

ISBN: (纸本)9789819778614

A chest X-ray is a common diagnostic tool for many thoracic illnesses. Interpreting these images and coming up with accurate diagnostic results is a difficult and time-consuming task for radiologists. Recent results using deep learning approaches to automate the analysis of chest X-rays are promising. This paper provides a comprehensive description of deep learning techniques used to chest X-ray processing for the diagnosis of several diseases, including TB, pneumonia, and COVID-19. This paper provide a summary of the primary deep learning architectures, such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), and attention models, that are used for disease classification, localization, segmentation, and report generation from chest X-rays. This paper also discuss explainability techniques that are used to offer transparency and increase trust in model predictions. Along with existing issues, future directions are taken into consideration, such as multimodal learning, few-shot learning, and model evaluation. This paper explains the model which takes less time and give the optimal results. In order to facilitate future research into developing AI systems that are more accurate, understandable, and advantageous from a medical standpoint, this article provides an overview of the most current developments in deep learning techniques for automated chest X-ray diagnosis. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Utilization of Machine Learning in Breast Cancer Detection 25

Utilization of Machine Learning in Breast Cancer Detection

引用

25th International Arab Conference on Information technology, ACIT 2024

作者： Momani, Mutaz Yousef, Nidal Aljawazneh, Qotadeh Issa, A.T. Abuowaida, Suhaila Alazaidah, Raed Khouj, Mohammed Talat Batyha, Radwan Elsoud, Esraa Abu Zarqa University Faculty of Information Technology Department of Computer Science Zarqa Jordan Al-Albayt University Faculty of Information Technology Department of Computer Science Mafraq Jordan Zarqa University Department of Data Science and Ai Faculty of Information Technology Zarqa Jordan University of Business and Technology Department of Electrical Engineering Jeddah21432 Saudi Arabia Private University Faculty of It Applied Science Department of Computer Science Amman Jordan

ISBN: (纸本)9798331540012

Breast cancer is the most prevalent cancer among women and can be deadly, necessitating early detection to enhance patient outcomes and treatment effectiveness. Recently, Machine Learning (ML) techniques have shown potential in medical diagnostics, especially in breast cancer screening. ML algorithms examine extensive medical databases, including patient records, to identify subtle patterns and anomalies indicative of cancer. These algorithms excel at recognizing intricate patterns that may be missed by human analysis. Thereby, increasing the precision of breast cancer diagnoses. By evaluating and analyzing medical data, ML algorithms can detect minute anomalies, facilitating early identification of potential cancer cases. A significant advantage of ML-based breast cancer detection is its capacity for continuous learning and adaptation. In this study, we utilized 34 classifiers and five feature selection methods on three datasets. The first dataset comprises 286 instances with 10 integer-type features. The second dataset includes 699 instances with 10 integer-type features, and the third dataset consists of 286 instances with 13 integer-type features. For Dataset 1, Filtered Classifier and J-48 yielded the best results, while the Bayesnet classifier performed best on Dataset 2. In Dataset 3, SimpleLogistic and LMT achieved the best outcomes. Regarding feature selection, the study showed that the GainRatioAttributeEval, InfoGainAttributeEval, and CorrelationAttributeEval techniques are the best among the feature selection methods tested in the study. © 2024 IEEE.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

Boosting Adversarial Training with Learnable Distribution

引用

computers, Materials & Continua 2024年第3期78卷 3247-3265页

作者： Kai Chen Jinwei Wang James Msughter Adeke Guangjie Liu Yuewei Dai School of Electronics and Information Engineering Nanjing University of Information Science and TechnologyNanjing210044China Key Laboratory of Intelligent Support Technology for Complex Environments Ministry of EducationNanjing210044China School of Computer and Software Nanjing University of Information Science and TechnologyNanjing210044China Nanjing Center for Applied Mathematics Nanjing211135China

In recent years,various adversarial defense methods have been proposed to improve the robustness of deep neural *** training is one of the most potent methods to defend against adversarial ***,the difference in the feature space between natural and adversarial examples hinders the accuracy and robustness of the model in adversarial *** paper proposes a learnable distribution adversarial training method,aiming to construct the same distribution for training data utilizing the Gaussian mixture *** distribution centroid is built to classify samples and constrain the distribution of the sample *** natural and adversarial examples are pushed to the same distribution centroid to improve the accuracy and robustness of the *** proposed method generates adversarial examples to close the distribution gap between the natural and adversarial examples through an attack algorithm explicitly designed for adversarial *** algorithm gradually increases the accuracy and robustness of the model by scaling ***,the proposed method outputs the predicted labels and the distance between the sample and the distribution *** distribution characteristics of the samples can be utilized to detect adversarial cases that can potentially evade the model *** effectiveness of the proposed method is demonstrated through comprehensive experiments.

关键词： Adversarial training feature space learnable distribution distribution centroid

来源：评论

学校读者我要写书评

暂无评论

AI-Powered Healthcare with Advanced Graph Convolutional Neural Networks and SVM for Diagnosis 3

AI-Powered Healthcare with Advanced Graph Convolutional Neur...

引用

3rd International Conference on Integrated Circuits and Communication Systems, ICICACS 2025

作者： Mahendran, A. Nalla, Sridhar Babu Trivedi, Lalit Mohan Mohankumar, S. Williams, D.R. Mohan Renuka, Manne Sona College of Arts and Science Department of Computer Science Salem India Sr University Department of Computer Science and Engineering Hyderabad India Moradabad Institute of Technology Department of Ash Mathematics Moradabad India Erode Sengunthar Engineering College Department of Computer Science and Engineering Perundurai India College of Computing and Information Sciences University of Technology and Applied Science-Sohar Department of Information Technology Sohar Oman Vidya Jyothi Institute of Technology Department of Electronics and Communication Engineering Hyderabad India

ISBN: (纸本)9798331508456

Medical treatment costs are rising for many reasons. These include an increasing world population, an aging population, greater healthcare needs, more diseases, and inflation. With AI's emergence in navigation, voice recognition, image lesion segmentation, ride-sharing, and smartphone personal assistants, the healthcare industry has quickly adopted it. Initial preprocessing includes standardization, missing values, and discarding anomalies. Next, the model is trained using the obtained features. Spearman correlation, a rank-based nonparametric statistic, can quantify monotonic correlation strength and direction. Even with missing values, the Graph Convolutional Network (GCN) performs well in model training. The suggested model outperformed state-of-the-art alternatives with 93.48% accuracy. It outperformed TransLSTM and GraphSAGE on complex datasets. This research stresses the system's ability to improve healthcare forecast accuracy through preprocessing, feature extraction, and GCN model training. Results show that AI can improve healthcare analytics despite lacking data. © 2025 IEEE.

关键词： Patient treatment

来源：评论

学校读者我要写书评

暂无评论

An Anonymous IoT-Based E-Health Monitoring System Using Blockchain technology

引用

IEEE Systems Journal 2023年第2期17卷 2422-2433页

作者： Samuel, Omaji Omojo, Akogwu Blessing Mohsin, Syed Muhammad Tiwari, Prayag Gupta, Deepak Band, Shahab S. Department of Computer Science Osara264103 Nigeria Edo State University Uzairue300281 Nigeria Shestco Applied Mathematics and Simulation Advanced Research Centre Abuja186 Nigeria Comsats University Islamabad Department of Computer Science Islamabad45550 Pakistan Aalto University Department of Computer Science Espoo02150 Finland Department of Computer Science and Engineering Maharaja Agrasen Institute of Technology Delhi110086 India Research Advisor Ucrd Chandigarh University GharuanPunjab Mohali140413 India National Yunlin University of Science and Technology Future Technology Research Center Douliou64002 Taiwan

The Internet of Things (IoT) has made it possible for health institutions to have remote diagnosis, reliable, preventive, and real-time decision-making. However, the anonymity and privacy of patients are not considered in IoT. Therefore, this article proposes a blockchain-based anonymous system, known as GarliMediChain, for providing anonymity and privacy during COVID-19 information sharing. In GarliMediChain, garlic routing and blockchain are integrated to provide low-latency communication, privacy, anonymity, trust, and security. Also, COVID-19 information is encrypted multiple times before transmitting to a series of nodes in the network. To ensure that COVID-19 information is successfully shared, a blockchain-based coalition system is proposed. The coalition system enables health institutions to share information while maximizing their payoffs. In addition, each institution uses the proposed fictitious play to study the strategies of others in order to update its belief by selecting the best responses from them. Furthermore, simulation results show that the proposed system is resistant to security-related attacks and is robust, efficient, and adaptive. From the results, the proposed proof-of-epidemiology-of-interest consensus protocol has 15.93% less computational cost than 26.30% of proof-of-work and 57.77% proof-of-authority consensus protocol, respectively. Nonetheless, the proposed GarliMediChain system promotes global collaborations by combining existing anonymity and trust solutions with the support of blockchain technology. © 2007-2012 IEEE.

关键词： COVID-19

来源：评论

学校读者我要写书评

暂无评论

SkinMarkNet: an automated approach for prediction of monkeyPox using image data augmentation with deep ensemble learning models

引用

Multimedia Tools and Applications 2024年第18期84卷 1-17页

作者： Akram, Aqsa Jamjoom, Arwa A. Innab, Nisreen Almujally, Nouf Abdullah Umer, Muhammad Alsubai, Shtwai Fimiani, Gianluca Bahawalpur Pakistan Department of Information Systems Faculty of Computing and Information Technology King Abdulaziz University Jeddah21589 Saudi Arabia Department of Computer Science and Information Systems College of Applied Sciences AlMaarefa University Riyadh Diriyah13713 Saudi Arabia Department of Information Systems College of Computer and Information Sciences Princess Nourah bint Abdulrahman University P.O. Box 84428 Riyadh11671 Saudi Arabia Department of Computer Science The Islamia University of Bahawalpur Bahawalpur Pakistan Department of Computer Science College of Computer Engineering and Sciences Prince Sattam bin Abdulaziz University P.O. Box 151 Al-Kharj11942 Saudi Arabia Department of Computer Science University of Salerno Fisciano84084 Italy

Monkeypox, a rare but potentially fatal viral disease, poses a significant public health challenge due to its potential for outbreaks and complications. Detecting monkeypox lesions early and accurately is vital to effectively manage and control the disease. This study introduces a novel method for classifying monkeypox lesions, employing data augmentation methods and a framework based on an ensemble of three transfer learning models called "SkinMarkNet". The dataset used in this research consists of skin lesion images collected from the Kaggle data repository, encompassing diverse demographics and lesion characteristics. This research uses image data augmentation techniques to tackle the scarcity of annotated data. This augmentation enriches the training dataset, thereby improving the model’s ability to perform effectively. Moreover, the novelty of this research work lies in the usage of three popular transfer learning models(Inception, Xception, and ResNet) for feature extraction and ensemble learning. The SkinMarkNet achieves promising results showing an accuracy of 90.615% for monkeypox lesion classification, outperforming traditional machine learning and deep learning methods utilized in recent research works. In addition, thorough comparative analysis is done with machine learning models and contemporary approaches to validate the efficacy of the proposed method. Overall, the findings underscore the potential of leveraging advanced deep learning architectures and data augmentation strategies for improving monkeypox lesion classification, thereby facilitating early diagnosis and intervention in clinical settings and public health surveillance efforts. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Learning systems

来源：评论

学校读者我要写书评

暂无评论

Use of Polygram Substitution in Spotting the Swindler in Secret Sharing 11th

Use of Polygram Substitution in Spotting the Swindler in Sec...

引用

11th International Conference on Intelligent Computing and Applications, ICRTC 2023

作者： Yadav, Poonam Satheesh Kumar, S. Muthukumaran, V. Sandeep Kumar, Mathivanan Radhika, E.K. Department of Computer Science School of Applied Sciences Reva University Bangalore India Department of Mathematics College of Engineering and Technology SRM Institute of Science and Technology Tamil Nadu Kattankulathur 603203 India School of Computer Science and Engineering Galgotias University Uttar Pradesh Greater Noida203201 India Department of Computer Science and Applications Sindhi College Bangalore560024 India

ISBN: (纸本)9789819717231

In the polygram substitution, one set of characters is replaced by another set based on algorithm. The model is proposed for calculating and estimating secure online secret shares, where dynamic modification in any of the shares will be identified if number of the shares are more than specified threshold value. The security is based on the intractability of factoring as it uses Lagrange polynomial function. This is further enhanced by a graphical masking and unmasking technique;it includes two phases: encryption of the individual secrets and sharing of the data. The generation of the shares is done by masking using Polygram substitution, and the shares are regenerated and unmasking of the qualified shares is done to find original secret. Masking is done on original secret, then shares are generated. Each encrypted share contains information which is different than original secret. Masking by Polygram substitution gives additional protection of secret data on the Internet. Moreover, we achieve efficient recovery of the original secret data. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Polynomials

来源：评论

学校读者我要写书评

暂无评论

Proposed Biometric Security System Based on Deep Learning and Chaos Algorithms

引用

computers, Materials & Continua 2023年第2期74卷 3515-3537页

作者： Iman Almomani Walid El-Shafai Aala AlKhayer Albandari Alsumayt Sumayh S.Aljameel Khalid Alissa Security Engineering Lab Computer Science DepartmentPrince Sultan UniversityRiyadh11586Saudi Arabia Computer Science Department King Abdullah II School of Information TechnologyThe University of Jordan11942Jordan Department of Electronics and Electrical Communications Engineering Faculty of Electronic EngineeringMenoufia UniversityMenouf32952Egypt Computer Science Department Applied CollegeImam Abdulrahman Bin Faisal UniversityP.O.Box 1982Dammam31441Saudi Arabia Computer Science Department College of Computer Science and Information TechnologyImam Abdulrahman Bin Faisal UniversityP.O.Box 1982Dammam31441Saudi Arabia SAUDI ARAMCO Cybersecurity Chair Networks and Communications DepartmentCollege of Computer Science and Information TechnologyImam Abdulrahman Bin Faisal UniversityP.O.Box 1982Dammam31441Saudi Arabia

Nowadays,there is tremendous growth in biometric authentication and cybersecurity ***,the efficient way of storing and securing personal biometric patterns is mandatory in most governmental and private ***,designing and implementing robust security algorithms for users’biometrics is still a hot research area to be *** work presents a powerful biometric security system(BSS)to protect different biometric modalities such as faces,iris,and *** proposed BSSmodel is based on hybridizing auto-encoder(AE)network and a chaos-based ciphering algorithm to cipher the details of the stored biometric patterns and ensures their *** employed AE network is unsupervised deep learning(DL)structure used in the proposed BSS model to extract main biometric *** obtained features are utilized to generate two random chaos *** first random chaos matrix is used to permute the pixels of biometric *** contrast,the second random matrix is used to further cipher and confuse the resulting permuted biometric pixels using a two-dimensional(2D)chaotic logisticmap(CLM)*** assess the efficiency of the proposed BSS,(1)different standardized color and grayscale images of the examined fingerprint,faces,and iris biometrics were used(2)comprehensive security and recognition evaluation metrics were *** assessment results have proven the authentication and robustness superiority of the proposed BSSmodel compared to other existing *** example,the proposed BSS succeeds in getting a high area under the receiver operating characteristic(AROC)value that reached 99.97%and low rates of 0.00137,0.00148,and 3516 CMC,2023,vol.74,no.20.00157 for equal error rate(EER),false reject rate(FRR),and a false accept rate(FAR),respectively.

关键词： Biometric security deep learning AE network 2D CLM cybersecurity and authentication applications feature extraction unsupervised 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：