检索结果-内蒙古大学图书馆

iSSL-AL: a deep active learning framework based on self-supervised learning for image classification

Neural Computing and Applications 2024年第28期36卷 17699-17713页

作者： Agha, Rand Mustafa, Ahmad M. Abuein, Qusai Department of Computer Information Systems Jordan University of Science and Technology Irbid22110 Jordan

Deep neural networks have demonstrated exceptional performance across numerous applications. However, DNNs require large amounts of labeled data to avoid overfitting. Unfortunately, the labeled data may not be available;annotating large amounts of data is time-consuming, laborious, and requires human expertise, making it unfeasible to rely on manpower for annotation. One of the solutions to address this limitation is active learning (AL), a technique that utilizes unlabeled data while maintaining high performance. AL reduces the annotation cost of large datasets and enhances the training of models with fewer annotations. Uncertainty sampling has been proven to be one of the most effective strategies in AL;however, it lacks diversity. This research proposes iSSL-AL, a novel active learning framework that utilizes self-supervised learning (SSL) to ensure informative yet diverse samples. Three main aspects categorize the novelty of our work. The first is extending the margin uncertainty sampling by incorporating SSL to select informative and diverse points. The second is employing incremental learning for efficient training of the AL base classifier, where the model is trained incrementally in each AL cycle. The third is addressing the cold start problem, as our framework achieved high results in the early stages of training. Experiments show that iSSL-AL outperforms other state-of-the-art algorithms on the MNIST, FashionMNIST, and CIFAR-10 datasets, with accuracy scores of 99%, 98.9%, and 93.5%, respectively, effectively selecting diverse and informative samples. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Supervised learning

来源：评论

学校读者我要写书评

暂无评论

Transparent and Accurate COVID-19 Diagnosis:Integrating Explainable AI with Advanced Deep Learning in CT Imaging

引用

computer Modeling in engineering & sciences 2024年第6期139卷 3101-3123页

作者： Mohammad Mehedi Hassan Salman A.AlQahtani Mabrook S.AlRakhami Ahmed Zohier Elhendi Department of Information Systems College of Computer and Information SciencesKing Saud UniversityRiyadh11543Saudi Arabia Department of Computer Engineering College of Computer and Information SciencesKing Saud UniversityRiyadh11543Saudi Arabia Science Technology and Innovation Department King Saud UniversityRiyadh11543Saudi Arabia

In the current landscape of the COVID-19 pandemic,the utilization of deep learning in medical imaging,especially in chest computed tomography(CT)scan analysis for virus detection,has become increasingly *** its potential,deep learning’s“black box”nature has been a major impediment to its broader acceptance in clinical environments,where transparency in decision-making is *** bridge this gap,our research integrates Explainable AI(XAI)techniques,specifically the Local Interpretable Model-Agnostic Explanations(LIME)method,with advanced deep learning *** integration forms a sophisticated and transparent framework for COVID-19 identification,enhancing the capability of standard Convolutional Neural Network(CNN)models through transfer learning and data *** approach leverages the refined DenseNet201 architecture for superior feature extraction and employs data augmentation strategies to foster robust model *** pivotal element of our methodology is the use of LIME,which demystifies the AI decision-making process,providing clinicians with clear,interpretable insights into the AI’s *** unique combination of an optimized Deep Neural Network(DNN)with LIME not only elevates the precision in detecting COVID-19 cases but also equips healthcare professionals with a deeper understanding of the diagnostic *** method,validated on the SARS-COV-2 CT-Scan dataset,demonstrates exceptional diagnostic accuracy,with performance metrics that reinforce its potential for seamless integration into modern healthcare *** innovative approach marks a significant advancement in creating explainable and trustworthy AI tools for medical decisionmaking in the ongoing battle against COVID-19.

关键词： Explainable AI COVID-19 CT images deep learning

来源：评论

学校读者我要写书评

暂无评论

Class-conditional domain adaptation for semantic segmentation

引用

Computational Visual Media 2024年第5期10卷 1013-1030页

作者： Yue Wang Yuke Li James H.Elder Runmin Wu Huchuan Lu School of Information and Communication Engineering Dalian University of TechnologyDalian 116024China School of Computer Science Wuhan UniversityWuhan 430072China Department of Electrical Engineering and Computer Science York UniversityToronto M3J 1P3Canada Department of Computer Science the University of Hong KongHong Kong 999077China

Semantic segmentation is an important sub-task for many ***,pixel-level ground-truth labeling is costly,and there is a tendency to overfit to training data,thereby limiting the generalization *** domain adaptation can potentially address these problems by allowing systems trained on labelled datasets from the source domain(including less expensive synthetic domain)to be adapted to a novel target *** conventional approach involves automatic extraction and alignment of the representations of source and target domains *** limitation of this approach is that it tends to neglect the differences between classes:representations of certain classes can be more easily extracted and aligned between the source and target domains than others,limiting the adaptation over all ***,we address:this problem by introducing a Class-Conditional Domain Adaptation(CCDA)*** incorporates a class-conditional multi-scale discriminator and class-conditional losses for both segmentation and ***,they measure the segmentation,shift the domain in a classconditional manner,and equalize the loss over *** results demonstrate that the performance of our CCDA method matches,and in some cases,surpasses that of state-of-the-art methods.

关键词： domain adaptation generative adversarial networks semantic segmentation cityscapes

来源：评论

学校读者我要写书评

暂无评论

Ensemble activation enabled deep classifier for Alzheimer's disease detection in the blockchain-enabled distributed edge environment

引用

International Journal of information technology (Singapore) 2024年 1-18页

作者： Mandawkar, Umakant Diwan, Tausif Department of Computer Science and Engineering Indian Institute of Information Technology Nagpur India Department of Computer Engineering SVKM’s Institute of Technology Maharashtra Dhule India

Alzheimer's disease (AD) is the most well-known cause of dementia that affects memory. Alzheimer's patients have a neurodegenerative disorder that results in the loss of many brain functions. Today’s researchers have determined that the key to enhancing patient care and treatment is early disease diagnosis. Traditional techniques for diagnosing AD are difficult ineffective, and take a lot of time to learn and execute. In this research, a disease detection system is developed using the proposed tawny flamingo-based federated deep CNN (Convolutional Neural Networks) (TF based-deep CNN) which can manage accurate patient diagnosis in dispersed contexts with blockchain support. The collected data is maintained and structured in a blockchain-based database, which upholds a high level of data security and increases the detection accuracy. After the medical professional receives the data, disease detection is done using the proposed federated learning framework, which makes it feasible to manage remote data effectively. The features are then extracted from the segments created by the CUF-GW (Cuttlefish-GreyWolf) algorithm. The detection is then carried out by the proposed ensemble activation-based deep CNN classifier, with the ensemble process being assisted by the TF (Tawny Flamingo) approach. According to AD accomplishments, at TP (Training Percentage) 80, acc (Accuracy), sen (Sensitivity), and spe (Specificity) are respectively 99.19%, 99.99%, and 98.87%;similarly, based on the k-fold 10, acc, sen, and spe are respectively 99.45%, 99.99%, and 98.45%. © Bharati Vidyapeeth's Institute of computer Applications and Management 2024.

关键词： Alzheimer's disease CUF-GW algorithm Deep CNN Federated learning TF

来源：评论

学校读者我要写书评

暂无评论

Enhanced Multi-Scale Object Detection Algorithm for Foggy Traffic Scenarios

引用

computers, Materials & Continua 2025年第2期82卷 2451-2474页

作者： Honglin Wang Zitong Shi Cheng Zhu School of Artificial Intelligence Nanjing University of Information Science and TechnologyNanjing210044China School of Computer Science Nanjing University of Information Science and TechnologyNanjing210044China School of Electrical&Computer Engineering University of Illinois at Urbana ChampaignUrbanaIL 61801USA

In foggy traffic scenarios, existing object detection algorithms face challenges such as low detection accuracy, poor robustness, occlusion, missed detections, and false detections. To address this issue, a multi-scale object detection algorithm based on an improved YOLOv8 has been proposed. Firstly, a lightweight attention mechanism, Triplet Attention, is introduced to enhance the algorithm’s ability to extract multi-dimensional and multi-scale features, thereby improving the receptive capability of the feature maps. Secondly, the Diverse Branch Block (DBB) is integrated into the CSP Bottleneck with two Convolutions (C2F) module to strengthen the fusion of semantic information across different layers. Thirdly, a new decoupled detection head is proposed by redesigning the original network head based on the Diverse Branch Block module to improve detection accuracy and reduce missed and false detections. Finally, the Minimum Point Distance based Intersection-over-Union (MPDIoU) is used to replace the original YOLOv8 Complete Intersection-over-Union (CIoU) to accelerate the network’s training convergence. Comparative experiments and dehazing pre-processing tests were conducted on the RTTS and VOC-Fog datasets. Compared to the baseline YOLOv8 model, the improved algorithm achieved mean Average Precision (mAP) improvements of 4.6% and 3.8%, respectively. After defogging pre-processing, the mAP increased by 5.3% and 4.4%, respectively. The experimental results demonstrate that the improved algorithm exhibits high practicality and effectiveness in foggy traffic scenarios.

关键词： Deep learning object detection foggy scenes traffic detection YOLOv8

来源：评论

学校读者我要写书评

暂无评论

A Lightweight Deep Learning-Based Model for Tomato Leaf Disease Classification

引用

computers, Materials & Continua 2023年第12期77卷 3969-3992页

作者： Naeem Ullah Javed Ali Khan Sultan Almakdi Mohammed S.Alshehri Mimonah Al Qathrady Eman Abdullah Aldakheel Doaa Sami Khafaga Department of Software Engineering University of Engineering and TechnologyTaxila4400Pakistan Department of Computer Science Faculty of PhysicsEngineeringand Computer ScienceUniversity of HertfordshireHatfieldAL109ABUK Department of Computer Science College of Computer Science and Information SystemNajran UniversityNajran55461Saudi Arabia Departments of Information Systems College of Computer Science and Information SystemsNajran UniversityNajran61441Saudi Arabia Department of Computer Sciences College of Computer and Information SciencesPrincess Nourah bint Abdulrahman UniversityRiyadh11671Saudi Arabia

Tomato leaf diseases significantly impact crop production,necessitating early detection for sustainable *** Learning(DL)has recently shown excellent results in identifying and classifying tomato leaf ***,current DL methods often require substantial computational resources,hindering their application on resource-constrained *** propose the Deep Tomato Detection Network(DTomatoDNet),a lightweight DL-based framework comprising 19 learnable layers for efficient tomato leaf disease classification to overcome *** Convn kernels used in the proposed(DTomatoDNet)framework is 1×1,which reduces the number of parameters and helps in more detailed and descriptive feature extraction for *** proposed DTomatoDNet model is trained from scratch to determine the classification success rate.10,000 tomato leaf images(1000 images per class)from the publicly accessible dataset,covering one healthy category and nine disease categories,are utilized in training the proposed DTomatoDNet *** specifically,we classified tomato leaf images into Target Spot(TS),Early Blight(EB),Late Blight(LB),Bacterial Spot(BS),Leaf Mold(LM),Tomato Yellow Leaf Curl Virus(YLCV),Septoria Leaf Spot(SLS),Spider Mites(SM),Tomato Mosaic Virus(MV),and Tomato Healthy(H).The proposed DTomatoDNet approach obtains a classification accuracy of 99.34%,demonstrating excellent accuracy in differentiating between tomato *** model could be used on mobile platforms because it is lightweight and designed with fewer *** farmers can utilize the proposed DTomatoDNet methodology to detect disease more quickly and easily once it has been integrated into mobile platforms by developing a mobile application.

关键词： CNN deep learning DTomatoDNet tomato leaf disease classification smart agriculture

来源：评论

学校读者我要写书评

暂无评论

Video captioning – a survey

引用

Multimedia Tools and Applications 2025年第2期84卷 947-978页

作者： Vaishnavi, J. Narmatha, V. Department of Computer Applications SRM Institute of Science and Technology Trichy India Department of Computer and Information Science Annamalai University Annamalai Nagar Tamilnadu 608002 India

The explosion of the novel phenomenon of the combination of computer vision and Natural language processing is playing a vital role in converting the ordinary world into a more technological pool. Natural language processing and computer vision are the vanguard of Artificial intelligence with the enormous potential currently ruling more fields. The trending Accelerated research of the NLP and CV combo are cancer screening, surgical simulation, visual properties description, visual retrieval, visual description, etc. Video captioning is a task of localizing the event and various features are extracted from the particular event and then the description of the entire event is performed with Natural languages using various techniques. The task of video captioning has some evolution of various techniques from Traditional methods to Deep learning. Currently, Deep learning techniques are ruling the field. Several models are being proposed for the better enhancement of the video captioning task with the default Encoder-Decoder framework and some models with the variance of Attention mechanism and Transfer learning etc., The Attainment of every technique and the accuracy of the generated results are highly dependent on the nature of the problem, diversity, performance of chosen dataset and construction of the technique with various layers, and the amount of data split for train, validate, and test. This survey discusses the captioning types, real-time applications of video captioning, evolution of various methods in video captioning, advanced techniques, different datasets, metrics that are used to evaluate the results, and analysis of the results of the existing models for video captioning. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： computer vision

来源：评论

学校读者我要写书评

暂无评论

A Base Station Deployment Algorithm for Wireless Positioning Considering Dynamic Obstacles

引用

computers, Materials & Continua 2025年第3期82卷 4573-4591页

作者： Aiguo Li Yunfei Jia Department of Network Engineering Xi’an University of Science and TechnologyXi’an710600China College of Computer Science and Technology Xi’an University of Science and TechnologyXi’an710600China

In the context of security systems,adequate signal coverage is paramount for the communication between security personnel and the accurate positioning of *** studies focus on optimizing base station deployment under the assumption of static obstacles,aiming to maximize the perception coverage of wireless RF(Radio Frequency)signals and reduce positioning blind ***,in practical security systems,obstacles are subject to change,necessitating the consideration of base station deployment in dynamic ***,research in this area still needs to be *** paper proposes a Dynamic Indoor Environment Beacon Deployment Algorithm(DIE-BDA)to address this *** algorithm considers the dynamic alterations in obstacle locations within the designated *** determines the requisite number of base stations,the requisite time,and the area’s practical and overall signal coverage *** experimental results demonstrate that the algorithm can calculate the deployment strategy in 0.12 s following a change in obstacle *** results show that the algorithm in this paper requires 0.12 s to compute the deployment strategy after the positions of obstacles *** 13 base stations,it achieves an effective coverage rate of 93.5%and an overall coverage rate of 97.75%.The algorithm can rapidly compute a revised deployment strategy in response to changes in obstacle positions within security systems,thereby ensuring the efficacy of signal coverage.

关键词： Wireless positioning base station deployment dynamic obstacles dynamic obstacle wireless positioning

来源：评论

学校读者我要写书评

暂无评论

An Enhanced Hybrid Model Based on CNN and BiLSTM for Identifying Individuals via Handwriting Analysis

引用

computer Modeling in engineering & sciences 2024年第8期140卷 1689-1710页

作者： Md.Abdur Rahim Fahmid Al Farid Abu Saleh Musa Miah Arpa Kar Puza Md.Nur Alam Md.Najmul Hossain Sarina Mansor Hezerul Abdul Karim Department of Computer Science and Engineering Pabna University of Science and TechnologyPabna6600Bangladesh Faculty of Engineering Multimedia UniversityCyberjaya63100Malaysia Department of Computer Science and Engineering Bangladesh Army University of Science and TechnologySaidpur5311Bangladesh Department of Mathematics Pabna University of Science and TechnologyPabna6600Bangladesh Department of Electrical Electronic and Communication EngineeringPabnaUniversity of Science and TechnologyPabna6600Bangladesh Department of Electrical and Communication Engineering Pabna University of Science and TechnologyPabna6600Bangladesh

Handwriting is a unique and significant human feature that distinguishes them from one *** are many researchers have endeavored to develop writing recognition systems utilizing specific signatures or symbols for person identification through ***,such systems are susceptible to forgery,posing security *** response to these challenges,we propose an innovative hybrid technique for individual identification based on independent handwriting,eliminating the reliance on specific signatures or *** response to these challenges,we propose an innovative hybrid technique for individual identification based on independent handwriting,eliminating the reliance on specific signatures or *** innovative method is intricately designed,encompassing five distinct phases:data collection,preprocessing,feature extraction,significant feature selection,and *** key advancement lies in the creation of a novel dataset specifically tailored for Bengali handwriting(BHW),setting the foundation for our comprehensive ***-preprocessing,we embarked on an exhaustive feature extraction process,encompassing integration with kinematic,statistical,spatial,and composite *** meticulous amalgamation resulted in a robust set of 91 *** enhance the efficiency of our system,we employed an analysis of variance(ANOVA)F test and mutual information scores approach,meticulously selecting the most pertinent *** the identification phase,we harnessed the power of cutting-edge deep learning models,notably the Convolutional Neural Network(CNN)and Bidirectional Long Short-Term Memory(BiLSTM).These models underwent rigorous training and testing to accurately discern individuals based on their handwriting ***,our methodology introduces a groundbreaking hybrid model that synergizes CNN and BiLSTM,capitalizing on fine motor features for enhanced individual ***,our experimental results unde

关键词： Bengali handwriting(BHW) person identification convolutional neural network(CNN) BiLSTM

来源：评论

学校读者我要写书评

暂无评论

Integration of EfficientNetB0 and Machine Learning for Fingerprint Classification

Informatica (Slovenia)

引用

Informatica (Slovenia) 2023年第5期47卷 49-56页

作者： Alhijaj, Jenan A. Khudeyer, Raidah S. Department of Computer Science College of Computer Science and Information Technology University of Basarh Basrah Iraq

A fingerprint is a common form of biometric technology used in human identification. The classification of fingerprints is crucial in identification systems because it significantly reduces the time required to identify a person and allows for the possibility of using fingerprints to distinguish between genders and identify individuals. Fingerprints are the most reliable identifiers because they are unique and impossible to fake. As a method of personal identification, fingerprints remain the best and most trustworthy. Fingerprint classification is crucial in a wide variety of settings, such as airports, banks, and emergencies involving explosives and natural disasters. This study proposes a deep learning strategy for determining whether a fingerprint belongs to a male or female person. With the help of pre-trained convolutional neural networks (CNN) in computer vision and an extremely powerful tool that has achieved significant success in image classification and pattern recognition. This work includes the use of the SOCOFing fingerprint dataset for training and employing a state-of-the-art model for feature extraction called EfficientNetB0, which was trained on the ImageNet dataset. Then feeding the extracted features into a principal component analysis (PCA) to decrease the dimension of these features and random forest RF classifier for fingerprint classification. Lastly, the tests showed that the proposed strategy outperformed the previous categorization methods in terms of accuracy (99.91%), speed for execution time, and efficiency. © 2023 Slovene Society Informatika. All rights reserved.

关键词： Principal component analysis

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：