检索结果-内蒙古大学图书馆

An Adaptive Scalable Data Pipeline for Multiclass Attack Classification in Large-Scale IoT Networks

Big Data Mining and Analytics 2024年第2期7卷 500-511页

作者： Selvam Saravanan Uma Maheswari Balasubramanian Department of Computer Science and Engineering Amrita School of ComputingAmrita Vishwa VidyapeethamBengaluru 560035India

The current large-scale Internet of Things(IoT)networks typically generate high-velocity network traffic *** use IoT devices to create botnets and launch attacks,such as DDoS,Spamming,Cryptocurrency mining,Phishing,*** service providers of large-scale IoT networks need to set up a data pipeline to collect the vast network traffic data from the IoT devices,store it,analyze it,and report the malicious IoT devices and types of ***,the attacks originating from IoT devices are dynamic,as attackers launch one kind of attack at one time and another kind of attack at another *** number of attacks and benign instances also vary from time to *** phenomenon of change in attack patterns is called concept ***,the attack detection system must learn continuously from the ever-changing real-time attack patterns in large-scale IoT network *** meet this requirement,in this work,we propose a data pipeline with Apache Kafka,Apache Spark structured streaming,and MongoDB that can adapt to the ever-changing attack patterns in real time and classify attacks in large-scale IoT *** concept drift is detected,the proposed system retrains the classifier with the instances that cause the drift and a representative subsample instances from the previous training of the *** proposed approach is evaluated with the latest dataset,IoT23,which consists of benign and several attack instances from various IoT *** classification accuracy is improved from 97.8%to 99.46%by the proposed *** training time of distributed random forest algorithm is also studied by varying the number of cores in Apache Spark environment.

关键词： Internet of Things(IoT) concept drift Apache Spark MongoDB Apache Kafka streaming

来源：评论

学校读者我要写书评

暂无评论

Music genre classification using convolution temporal pooling network

引用

Multimedia Tools and Applications 2024年 1-16页

作者： T. M, Vijayameenakshi T. R, Swapna Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Coimbatore India

Music genre classification is one of the most interesting topics in digital music. Classifying genres is basically subjective, and different listeners may perceive genres in various ways. Furthermore, it might be difficult to classify some songs accurately since they belong to numerous genres. Genres are incredibly wide and ill-defined categories, which makes them problematic. Thus, genre-based measures are inherently inaccurate and coarse. Moreover, not every piece of music cleanly fits into a particular genre. Many papers based on deep neural networks perform sound recognition and classification with input images of audio, which do not affect the time–frequency representation of a signal. The traditional method adds waveform augmentation to the audio signal, thereby increasing the network's training speed. This paper considers music genre classification with the convolution temporal pooling framework and explores the impact of adding the SpecAugment method to augment the spectrogram itself. The augmented spectrogram is then fed into a convolutional temporal pooling network. In this model, the temporal and pooling layers identify the genre pattern and classify the songs based on the genre. It also predicts these duplication that will occur in the given sample. We apply this model to the GTZAN dataset, a widely used dataset for music genre classification. This method improves the identification of Rock and Pop song and also eliminates the replication of the songs. The trained model reports an accuracy of 0.75 for training a 30-s audio file. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Convolutional temporal pooling network Genre classification Music genre classification SpecAugment Spectrogram

来源：评论

学校读者我要写书评

暂无评论

ScanSavant: Malware Detection for Android Applications with Explainable AI

International Journal of Interactive Mobile Technologies

引用

International Journal of Interactive Mobile Technologies 2024年第19期18卷 171-181页

作者： Navaneethan, S. Udhaya Kumar, S. Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Tamil Nadu Chennai India

Mobile devices face SQL injection, malware, and web-based threats. Current solutions lack real-time detection. This paper introduces an Android app with advanced algorithms for real-time threat scanning. During testing, our application detected 94% of SQL injection attempts, outperforming the 86% average detection rate in similar studies. For malware analysis, it achieved a 97% detection accuracy on a dataset of infected files, higher than the industry standard of 93%. Additionally, our app can detect 85 malware variants and assign 15 attributes (***.8, ***, ***, ***, ***, ***, ***-2017-0143, ***, ***, Backdoor. DarkComet, ***, ***, ***, ***32. Necurs, and ***) to some malwares, providing detailed analysis for better threat management. The application effectively scans both EXE and APK files, ensuring com-prehensive protection. When assessing website links, the application identified security risks with 96% accuracy, demonstrating its capability in managing web-based threats. This app detects SQL injections, analyses malware, and assesses website security, bolstering cyber defence with user-friendly features and top-notch threat mitigation. © 2024, International Federation of engineering Education Societies (IFEES). All rights reserved.

关键词： Android malware

来源：评论

学校读者我要写书评

暂无评论

Development of a diacritic-aware large vocabulary automatic speech recognition for Hausa language

引用

International Journal of Speech Technology 2024年第3期27卷 687-700页

作者： Abubakar, Abdulqahar Mukhtar Gupta, Deepa Vekkot, Susmitha Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Bengaluru India Department of Electronics & amp Communication Engineering Amrita School of Engineering Amrita Vishwa Vidyapeetham Bengaluru India

Research on voice recognition for African languages is limited due to the scarcity of digital resources for training and adaptation, despite its broad usefulness. The Hausa language, spoken by almost fifty million inhabitants in West and Central Africa, is an example of a linguistic domain that has not been thoroughly studied. The Hausa language employs diacritics, which are symbols located above alphabetical characters to convey further information. By removing diacritics, the number of homographs increases, making it difficult to distinguish between similar words. This paper presents a study on speech recognition in the Hausa Language, specifically focusing on diacritized words. The study utilises the state-of-the-art wave2vec2.0 and Whisper deep learning architecture models, for transcribing audio signals into corresponding Hausa text. According to the results obtained in the study, the Whisper-large deep model emerged as the best, achieving a word error rate of 4.23% representing a considerable improvement of 43.9% when compared to the existing state-of-the-art model for Hausa language speech recognition. Additionally, the Whsiper-large model demonstrated a diacritic coverage of 92%, precision of 98.87%, with a diacritic error rate of 2.1%. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Discovering Social Media Influencers Using a Deep Regression Analysis

引用

SN computer science 2025年第4期6卷 1-18页

作者： Ramya, G.R. Bagavathi Sivakumar, P. Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Coimbatore India

Highly influential users (IUs) play a vital role in disseminating information on online social networks (OSNs). Recognizing IUs is crucial for brand awareness, strategic marketing and consumer engagement. Researchers have developed the temporal influential model (TIM) and incremental logistic regression (ILLR) to find out the probability of a user being influential on Twitter. However, accurately identifying IUs requires considering multiple user and tweet attributes. This study proposes a model for predicting IUs by analyzing user and trend-based attributes from tweets, calculating a weighted associated influence (WAI) score for each user and enhancing IU prediction through user similarity, influence degree, and trustworthiness metrics. The goal is to combine the TIM with deep regression (TIM-DeepReg) to improve accuracy in predicting high IUs on OSNs while handling large amounts of data and attributes. The TIM-DeepReg model combines convolutional neural network (CNN) with regression to automatically learn user and tweet attributes and enhance IU prediction efficiency. Lastly, the experimental findings show that the TIM-DeepReg model surpasses other models in IU prediction for the Tariff, Service, and Punctuality topics of airline services on Twitter and Facebook, with 87.4%, 88.9%, and 88.8% of the most influential individuals predicted. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Deep regression Influencer on social networks Influencer prediction from Facebook and Twitter Temporal influential model Weighted associated influence score

来源：评论

学校读者我要写书评

暂无评论

HybridViT: An Approach for Alzheimer’s Disease Classification with ADNI Neuroimaging Data

引用

SN computer science 2025年第4期6卷 1-18页

作者： Tekumudi, Vivek Sai Surya Chaitanya Ramanathan, Annamalai Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Chennai India

Alzheimer’s disease (AD) is a prevalent neurological disorder characterized by progressive brain cell degeneration and atrophy, leading to a gradual decline in cognitive and functional abilities. Timely diagnosis is paramount in potentially delaying or preventing the course of this debilitating condition. Magnetic Resonance Imaging presents a non-invasive means for longitudinal monitoring and serves as a crucial biomarker for tracking disease progression. In particular, Structural Magnetic Resonance Imaging (sMRI) enables the quantification of atrophy, which is a reliable indicator for assessing the precise stage and severity of AD-related neurodegeneration. In this research, we employ advanced machine learning techniques to address the challenge of accurate and early diagnosis of AD. We utilize a comprehensive dataset comprising five stages of 2D sMRI Image data, encompassing AD, Cognitively Normal, Mild Cognitive Impairment, Early Mild Cognitive Impairment, and Late Mild Cognitive Impairment classes. To optimize the classification process, we explore a novel approach that combines a Vision Transformer with pre-trained convolutional neural networks. Our study includes binary and multi-class classification tasks, and the performance of 26 pre-trained Keras Deep Learning models is assessed. Notably, The DenseNet121+ViT model achieves 91.26% accuracy in multi-class classification. For binary classification, MobileNetV2+ViT, EfficientNetB4+ViT, and MobileNet+ViT achieve accuracies of 92.33%, 97.82%, and 94.81%, respectively. These results highlight our approach’s potential to improve AD diagnosis accuracy and underscore the importance of deep learning in early detection of neurological disorders. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Alzheimer’s disease (AD) Alzheimer’s disease neuroimaging initiative (ADNI) HybridViT Pre-trained models Vision transformer

来源：评论

学校读者我要写书评

暂无评论

CSG_IACO: an efficient stigmergic-based improved ACO routing strategy to determine the effective path based on traffic density in V2V networks

引用

International Journal of Information Technology (Singapore) 2025年第3期17卷 1883-1908页

作者： Kumari, J. Jesy Janet Thangam, S. Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Bengaluru India

Vehicular ad hoc networks (VANETs) are an essential element and building block of the autonomous vehicle system. VANETs, a subcategory of mobile ad hoc networks (MANETs), stand out due to certain predetermined attributes. In VANETs, optimal path routing offers solutions to address issues like traffic density, node mobility, link failure, and bandwidth. Finding the optimal VANET road path between source and destination conditions is difficult. Extensive paths increase network overhead, communication costs, and failure probability, reducing routing efficiency. We present the new traffic density-based stigmergic ant colony optimization routing (CSG_IACO) to improve routing performance. Improved ant colonies stimulate the proposed CSG_IACO, which measures inter-vehicular density. Ant behavior serves as the inspiration for the improved ACO technique, a meta-heuristic methodology. Phase 1 of the CSG_IACO involves using the traffic density to form a cluster among the vehicles. Phase 2 of the suggested approach is Stigmergy-based ant colony optimization, which guides ants to indirectly communicate with each other to act upon the change in the environment to find a best path. This improves protocol efficiency in all aspects. The simulations using NS 2.35 and SUMO indicate that the proposed system considerably reduces obstacles to finding the optimal path compared to existing systems. The simulations show that our suggested scheme performs better than AODV-DATRLD, ACO-DATRLD, and CaCOIoV in terms of performance including the count of delivered packets, throughput, end-to-end delay, the average count of cluster formations per cluster head, and the number of dropped packets. © Bharati Vidyapeeth's Institute of computer Applications and Management 2025.

关键词： Ant colony optimization Cluster Optimal path QoS Stigmergy Traffic density VANETs

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Approach to Dynamic Thread Allocation Driven by Workload and Task Prioritization

引用

SN computer science 2025年第3期6卷 1-19页

作者： Abirami, K. Harini, N. Department of Computer Science and Engineering Amrita School of Computing Coimbatore Amrita Vishwa Vidyapeetham Coimbatore India

In today’s rapidly changing world, cloud service providers face numerous challenges in managing resources and meeting customer demands. To address these challenges, cloud service providers should prioritize the tasks and allocate threads efficiently. By allocating threads efficiently, service providers can optimize resource utilization and reduce response times and job completion times which in turn improves customer satisfaction. We propose a hybrid model that integrates the threshold-driven thread Balancer (TDTB) for dynamic thread allocation with a Customer Retention Model (CRM) to prioritize tasks based on customer behaviour. Our model initiates the establishment of a thread pool configured with a dynamic threshold based on workload (on-demand requests). Concurrently, a queue to handle incoming customer requests is set up alongside a CRM priority queue designed for task prioritization. Our model calculates Job duration and completion times for n priority classes. Crown consumers experience shorter durations and completion times compared to other priority classes. Tasks are completed in descending order of priority. The results of thorough experimentation clearly show that crown consumer tasks are highly satisfied. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Customer segmentation Response time Service differentiation Thread allocation Throughput Workload

来源：评论

学校读者我要写书评

暂无评论

Enhancing Secure Data Transmission Through Deep Learning-Based Image Steganography and Super-Resolution Generative Adversarial Networks

引用

SN computer science 2024年第8期5卷 1-15页

作者： Ramesh, Sreyas Sukanth, B.N. Jaswanth, Sathyavarapu Sri Singh, Rimjhim Padam Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Bengaluru India

In the era of multimedia technology digital images are essential and keeping them safe from unauthorised access is crucial. To address this issue, the proposed research explores the intersection of image steganography and deep learning for enhanced secure data transmission. It proposes a unique approach that leverages skip-connection-based residual network to extract semantic features and reconstruct images for secure data transmission. This paper introduces a novel steganographic approach combining LSB embedding with sprite image generation instead of directly using image features. Hence, the proposed method effectively hides data within images while maintaining high invisibility. Furthermore, Super-Resolution Generative Adversarial Network (SRGAN) is employed to improve the quality of the reconstructed images. Several other custom and state-of-art convolutional architectures have also been investigated for semantic feature extraction and image reconstruction, including Convolutional Autoencoders (CAE), Deep convolutional network (DCN), DenseNet, InceptionNet and the XceptionNet model, etc. for the comprehensive comparisons against the proposed model. The proposed technique proved to be the best performing model with an SSIM value of 1.00 and PSNR value of 33.04 on the test data. Overall, this research work combines deep learning with steganography to create a secure and hidden communication method. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Data security Least significant bit Sprite image Steganography Super-resolution generative adversarial network

来源：评论

学校读者我要写书评

暂无评论

Convergence of Deep Learning and Forensic Methodologies Using Self-attention Integrated EfficientNet Model for Deep Fake Detection

引用

SN computer science 2024年第8期5卷 1-12页

作者： Singh, Rimjhim Padam Sree, Nichenametla Hima Reddy, Koti Leela Sai Praneeth Jashwanth, Kandukuri Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Bengaluru India

The immensely increasing number of Deepfake technologies poses significant challenges to digital media integrity, leading to the immediate need for effective Deepfake detection methods. In light of the growing threat posed by sophisticated synthetic media, this paper tackles the urgent need for accurate deepfake detection in real-time by proposing a novel classification model employing a fine-tuned EfficientNet-B0 model with an optimally integrated Self-attention mechanism. The study also analyzes various Deepfake generation techniques by utilizing a suite of Convolutional Neural Networks namely, XceptionNet, NasNetLarge, InceptionV3, ConNextv1, ConvNext2 and EfficientB0 model on the Celeb-DF dataset for baseline model selection and performance comparison purposes. The results prove that the proposed fine-tuned EfficientNet-B0 model with integrated self-attention mechanism is the most superior at identifying complex long-range dependencies and relations in various unconnected image regions, highlighting the importance of attention mechanisms in improving deep learning models for multimedia forensics with a remarkable F-score of 97.67%. In an era marked by the proliferation of sophisticated artificial intelligence (AI)-driven synthetic media, these findings establish the foundation for strengthening digital media platform security and protecting information authenticity. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Celeb-DF dataset EfficientNetB0 Multimedia forensics NasNetLarge Self-attention mechanism Xception Deepfake detection

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：