检索结果-内蒙古大学图书馆

Hybrid RMDL-CNN for speech recognition from unclear speech signal

International Journal of Speech Technology 2025年第1期28卷 195-217页

作者： Bhargava, Raja Arivazhagan, N. Babu, Kunchala Suresh Research Scholar Department of Computer Science and Engineering SRM Institute of Science and Technology Kattankulathur Chennai India Department of Computational Intelligence SRM Institute of Science and Technology Kattankulathur Chennai India Department of Computer Science and Engineering Potti Sriramulu Chaluvadi Mallikarjuna Rao college of Engineering and Technology Andhra Pradesh Vijayawada India

ASR is an effectual approach, which converts human speech into computer actions or text format. It involves extracting and determining the noise feature, the audio model, and the language model. The extraction and determination of the noise feature is a crucial aspect of speech recognition, serving as both a process of information compression and signal deconvolution. ASR schemes are mostly employed in smart homes, smart appliances, and biometric schemes. Yet, traditional approaches offer very low performance because of a noisy environment. Moreover, local differences and accents negatively influence the ASR scheme execution during the conversion of the speech signals. This paper introduces a hybrid RMDL-CNN method to address these challenges. At first, the input of unclear speech is carried out by the dataset. Then, signal pre-processing is done by employing a Gaussian filter. After that, voice enhancement is accomplished by employing nonlinear spectral subtraction. Later, the speech word is segmented from the enhanced output based on the Attentional Encoder-Decoder approach and finally, the speech is recognized using the proposed RMDL-CNN. The RMDL-CNN method is devised by the combination of RMDL and CNN. Furthermore, the established RMDL-CNN is accessed for its efficiency based on several values of k-group value, as well as learning data. In addition, the introduced RMDL-CNN approach for speech recognition achieved better accuracy, PPV, as well as NPV of 0.909, 0.947, and 0.917 for dataset 1. Moreover, the RMDL-CNN has achieved the highest accuracy of 0.909, PPV of 0.926 and NPV of 0.888 for dataset 2. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Fraud Detection in Banking using the Kaggle Credit Card Dataset and XGBoost Model 5

Fraud Detection in Banking using the Kaggle Credit Card Data...

引用

5th International Conference on IoT Based Control Networks and Intelligent Systems, ICICNIS 2024

作者： Ango, Rithika Masih, Raj Kumar Reddy, C Kishor Kumar Shuaib, Mohammed Monika Singh, T. Alam, Shadab Stanley College of Engineering & Technology for Women Department of Computer Science and Engineering Hyderabad India Jazan University College of Engineering and Computer Science Department of Computer Science Jazan Saudi Arabia

ISBN: (纸本)9798331518097

The threat posed by credit card fraud, and by extension, online banking, continues to grow with the convenience brought forth by online banking services. Many financial institutions and customers stand at great risk because of this reality. The purpose of the paper is to enhance the security of banking systems through advanced technologies - machine learning. The research employed the well-known dataset from Kaggle Credit Card, which contains numerous records of transactions labeled as either fraud or not, thus presenting an aspect similar to real life imbalanced data. The main focus in the present work is the XGBoost model which involves one of the ensemble learning methods, which is characterized by a high level of precision and fast execution times. As a result of hyperparameter optimization and feature engineering, the paper also works on the performance of the algorithm improving its ability to detect fraudulent transactions in a sea of normal transactions. The results obtained from the experiments show that the XGBoost model is not only proven to be more accurate than earlier approaches to fraud detection, but also greatly reduces the number of false positives. This study not only contributes to the existing literature on detecting financial fraud but also represents the ways in which finance can benefit from machine learning to design secure systems within banking operations. The metrics in which the model suggested outperformed the available approaches are, the accuracy is 98.60%, error rate is 2.10% while its precision reached 95.40, recall of 92.30, F1 score 93.20 and AUC-ROC 99.70. This study not only contributes to the existing literature on detecting financial fraud but also represents the ways in which finance can benefit from machine learning to design secure systems within banking operations. © 2024 IEEE.

关键词： Banking

来源：评论

学校读者我要写书评

暂无评论

Realtime Object Detection Through M-ResNet in Video Surveillance System

引用

Intelligent Automation & Soft Computing 2023年第2期35卷 2257-2271页

作者： S.Prabu J.M.Gnanasekar Department of Computer Science&Engineering Karpaga Vinayaga College of Engineering and TechnologyChengalpattuTamilnadu603308India Department of Computer Science&Engineering Sri Venkateshwara College of Engineering&TechnologyChennai602001TamilnaduIndia

Object detection plays a vital role in the video surveillance *** enhance security,surveillance cameras are now installed in public areas such as trafﬁc signals,roadways,retail malls,train stations,and ***,monitor-ing the video continually at a quicker pace is a challenging *** a consequence,security cameras are useless and need human *** primary difﬁculty with video surveillance is identifying abnormalities such as thefts,accidents,crimes,or other unlawful *** anomalous action does not occur at a high-er rate than usual *** detect the object in a video,ﬁrst we analyze the images pixel by *** digital image processing,segmentation is the process of segregating the individual image parts into *** performance of segmenta-tion is affected by irregular illumination and/or low *** factors highly affect the real-time object detection process in the video surveillance *** this paper,a modiﬁed ResNet model(M-Resnet)is proposed to enhance the image which is affected by insufﬁcient *** results provide the comparison of existing method output and modiﬁcation architecture of the ResNet model shows the considerable amount improvement in detection objects in the video *** proposed model shows better results in the metrics like preci-sion,recall,pixel accuracy,etc.,andﬁnds a reasonable improvement in the object detection.

关键词： Object detection ResNet video survilence image processing object quality

来源：评论

学校读者我要写书评

暂无评论

Cognitive agent based fault tolerance in ubiquitous networks: a machine learning approach

引用

International Journal of Information Technology (Singapore) 2024年第4期16卷 2363-2377页

作者： Bhajantri, Lokesh B. Ayyannavar, Vasudha V. Department of Information Science and Engineering Basaveshwar Engineering College Karnataka Bagalkote India Department of Computer Science and Engineering Basaveshwar Engineering College Bagalkote VTU Karnataka Belagavi India

Ubiquitous Networks play an essential role in accessing ubiquitous computing services at anytime, anywhere, and anyplace through computing nodes of heterogeneous networks. Nowadays, ubiquitous network faces various issues related to fault management or tolerance in a real world environment. In this type of network, fault tolerance is one of the vital issues due to dynamic and uncertain character of ubiquitous networks with various constraints of nodes and networks to maximize network lifetime, availability, and reliability of the networks. In this paper, it is proposed that a cognitive agent based fault tolerance system using reinforcement learning algorithm to provide efficient ubiquitous services to the users over the networks. The main objective of the proposed work is to emphasize the analysis, identification, and recovery of various faults under resource constraints of ubiquitous networks in terms of node, memory, load, and network level faults (link level fault and packet error) using a cognitive agent based machine learning algorithm. Also, the proposed work performs better fault tolerance than conventional methods (without using cognitive agents) by considering fault detection rate, fault recovery rate, fault recovery time, packet delivery ratio, agent computation overhead, and energy consumption for performing the desired set of tasks and services ubiquitously in the ubiquitous computing environment. © The Author(s), under exclusive licence to Bharati Vidyapeeth's Institute of computer Applications and Management 2024.

关键词： Cognitive agents Heterogeneous network systems Reinforcement learning Ubiquitous networks Ubiquitous service systems

来源：评论

学校读者我要写书评

暂无评论

Mood Mapper: Unveiling Emotions in Speech Through Deep Learning 2

Mood Mapper: Unveiling Emotions in Speech Through Deep Learn...

引用

2nd IEEE International Conference on Networks, Multimedia and Information Technology, NMITCON 2024

作者： Sundaram, B. Meenakshi Yakshitha, K.C. New Horizon College of Engineering Department of Computer Science Engineering Bengaluru India New Horizon College of Engineering Computer Science Engineering Bengaluru India

ISBN: (纸本)9798350372892

A key component of human-computer interaction is Speech Emotion Recognition (SER), which allows computers to understand and react to the emotional states expressed through spoken language. Using specialized neural network designs, the model showcases outstanding accuracy in classifying emotions like joy, sadness, anger, and surprise. Trained on a diverse dataset, the methodology involves preprocessing audio data, extracting relevant features, and employing deep neural networks. The model's robust performance across linguistic and cultural variations underscores its potential applications in human-computer interaction, virtual assistants, and affective computing. This makes a substantial contribution to the development of emotionally intelligent machines, advancing human-machine communication through a profound knowledge of feelings conveyed in speech, thus clearing the path for additional empathetic and responsive artificial intelligence systems. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Diabetes Mellitus Prediction and Severity Calculation Using Stacked Ensemble Method

引用

SN computer science 2024年第8期5卷 1001页

作者： Ananthi, G. Santhiya, S. Gokila, V. Department of Computer Science and Engineering Mepco Schlenk Engineering College (Autonomous) Sivakasi India

Millions of individuals throughout the world suffer with diabetes mellitus, a chronic condition that can be effectively managed with early detection and precise prognosis. In this work, a machine learning method for predicting diabetes utilizes stacking ensemble algorithms with hyperparameter tuning. The proposed method uses three fundamental classifiers such as Random Forest, logistic regression, gradient boosting and tunes certain hyperparameters to improve the performance of each algorithm. Another classifier, called support vector machine, with hyperparameter tweaking is used as a meta-classifier to aggregate the predictions of the basic classifiers. It has been demonstrated that the stacking ensemble technique, especially when combined with hyperparameter tuning, enhances the prediction performance of machine learning models. The proposed technique is evaluated using two datasets, namely PIMA Indian Diabetes dataset, and dataset 2. Hierarchical clustering is used to forecast the disease's severity level. Severity levels are assigned based on the model's clustering results. The findings show that the suggested method accurately predicts diabetes and performs better than the individual base classifiers. To evaluate the severity level, a model created using a machine learning classification technique is chosen for further development. The model is finally included with the web application. The efficiency of the model is evaluated using a variety of performance criteria for evaluation such as accuracy, precision, recall, and F1-score. The proposed stacked ensemble method results in the accuracy of 97% which shows the improvement in the state-of-the-art methods. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Accuracy Diabetes Severity Stacked ensemble method Web application

来源：评论

学校读者我要写书评

暂无评论

Hybrid millimeter wave heterogeneous networks with spatially correlated user equipment

引用

Digital Communications and Networks 2024年第4期10卷 904-917页

作者： Arif Ullah Ziaul Haq Abbas Ghulam Abbas Fazal Muhammad Jae-Mo Kang Department of Computer Engineering College of IT ConvergenceChosun UniversityGwangju61452South Korea Faculty of Electrical Engineering GIK Institute of Engineering Sciences and TechnologyTopi23640Pakistan Faculty of Computer Science and Engineering GIK Institute of Engineering Sciences and TechnologyTopi23640Pakistan Department of Electrical Engineering University of Engineering and TechnologyMardan23200Pakistan Department of Artificial Intelligence Kyungpook National UniversityDaegu41566South Korea

In this paper,we analyze a hybrid Heterogeneous Cellular Network(HCNet)framework by deploying millimeter Wave(mmWave)small cells with coexisting traditional sub-6GHz macro cells to achieve improved coverage and high data *** consider randomly-deployed macro base stations throughout the network whereas mmWave Small Base Stations(SBSs)are deployed in the areas with high User Equipment(UE)*** user centric deployment of mmWave SBSs inevitably incurs correlation between UE and *** a realistic scenario where the UEs are distributed according to Poisson cluster process and directional beamforming with line-of-sight and non-line-of-sight transmissions is adopted for mmWave *** using tools from stochastic geometry,we develop an analytical framework to analyze various performance metrics in the downlink hybrid HCNets under biased received power *** UE clustering we considered Thomas cluster process and derive expressions for the association probability,coverage probability,area spectral efficiency,and energy *** also provide Monte Carlo simulation results to validate the accuracy of the derived ***,we analyze the impact of mmWave operating frequency,antenna gain,small cell biasing,and BSs density to get useful engineering insights into the performance of hybrid mmWave *** results show that network performance is significantly improved by deploying millimeter wave SBS instead of microwave BS in hot spots.

关键词： Downlink cell association Heterogeneous cellular networks Integrated sub-6GHz and mmWave networks Millimeter wave communications Poisson cluster process

来源：评论

学校读者我要写书评

暂无评论

Integrating the Internet of Things to Protect Electric Vehicle Control Systems from Cyber Attacks

引用

IAENG International Journal of Applied Mathematics 2024年第3期54卷 433-440页

作者： Bhukya, Baddu Naik Venkataiah, V. Kuchibhatla, S. Mani Koteswari, S. Kumari, R.V.S. Lakshmi Raju, Yallapragada Ravi Department of Electrical and Electronics Engineering Prasad V. Potluri Siddhartha Institute of Technology Andhra Pradesh Vijayawada India Department of Computer Science Engineering CMR College of Engineering & Technology Telangana Hyderabad India Department of Electrical and Electronics Engineering ACE Engineering College Telangana Hyderabad India Department of Electronics and Communication Engineering Pragati Engineering College Kakinada India Andhra Pradesh Surampalem India Department of Electrical and Electronics Engineering Gayatri Vidya Parishad College of Engineering for Women Andhra Pradesh Visakhapatnam India Department of Computer Science and Technology Madanapalle Institute of Technology and Science Andhra Pradesh India

The Internet of Things (IoT) facilitates the delivery of intelligent services by sensing, gathering, processing, and exchanging data from millions of linked smart devices. The Internet of Things (IoT), which is based on communication infrastructure, provides real-time cyber-physical device monitoring, control, and sensing, such as electric autos. Traditional communication infrastructure's vulnerability to cyberattacks hampered IoT's capacity to investigate these potential applications. This study proposes an algorithm for monitoring and controlling electric vehicles through the Internet of Things communication network to prevent fake data injection attacks. A state-space model depicts a fully autonomous electric car equipped with a vision system. Intelligent sensors and actuators from the Internet of Things are utilized to monitor and change system conditions, compensating for the large distance between the electric car and the control center. Sensing data is delivered from the vehicle to the central command center via an insecure, attack-prone communication path. The mean square error approach is used to derive states in the most effective state estimation system for comprehending and showing autos. A semi-definite programming approach is used to create an optimal control algorithm to manage the vehicle states. The simulation results show how precisely and successfully the proposed algorithms can foresee and regulate a vehicle's state. © (2024), (International Association of Engineers). All Rights Reserved.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Analysis of Watermarked Video Optimization and Training Based on Classification Using Deep Learning Techniques

引用

SN computer science 2024年第2期5卷 220页

作者： Muthulakshmi, K. Valarmathi, K. Department of Information Technology Panimalar Engineering College Chennai India Department of Computer Science and Engineering Panimalar Engineering College Chennai India

The exchange of multimedia data is secured using watermarking. Typically, embedding a watermark involves optimising scheme parameters, frequently by using meta-heuristic optimization methods. Despite the fact that DNN have achieved significant advancements in field of multimedia representation, training neural methods takes a lot of time and data. This research propose novel method in analysing watermarked video optimization and training based on DL. Here input is collected as watermarked input video and optimize and train the watermarked input video data with enhanced security using swarm optimization-based clustering fuzzy greedy trust analysis. then the optimized video has been trained using variational temporal-based u-net convolutional encoder neural networks (VTU-net ENN). Experimental analysis has been carried out in terms of accuracy, precision, recall, F-1 score, RMSE and NSE for various watermarked dataset. Accuracy attained by proposed technique is 92%, precision attained by proposed technique is 75%, recall attained is 62%, F-1 score attained 55%, RMSE obtained is 53%, NSE attained is 48% by proposed technique. © 2024, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.

关键词： Deep learning Optimization and training Securing multimedia Watermarked video

来源：评论

学校读者我要写书评

暂无评论

Bernstein polynomials and Bezier curves: a novel modeling approach to secure ECG data transmission

引用

International Journal of Information Technology (Singapore) 2024年第2期16卷 1043-1053页

作者： Rajagopal, Prahalad Premnath, Pooja Arumugam, Chamundeswari Department of Computer Science and Engineering Sri Sivasubramaniya Nadar College of Engineering Chennai India

In the digital era, protecting sensitive medical data, including Electrocardiogram (ECG) signals, is of utmost importance. Maintaining the confidentiality, integrity, and authenticity of ECG signals while safeguarding their privacy is a critical challenge. This paper proposes a novel approach to safeguard ECG waveforms by utilizing polynomial modeling techniques followed by encryption. Polynomial functions, specifically Bernstein and Bezier-Bernstein polynomials, are employed for accurate approximation of the ECG signal, accounting for noise, interference, and channel distortions. The coefficients obtained from polynomial modeling are encrypted using the standard AES and Fernet algorithms. Decryption allows authorized access to the ECG signals, facilitating methodical waveform reconstruction. The proposed methodology’s performance is evaluated using various metrics to assess its effectiveness in protecting sensitive medical data. Metrics include Root Mean Square Error, Peak Signal to Noise Ratio, Correlation Coefficient, and Structural Similarity Index. The proposed method effectively safeguards ECG waveforms by combining polynomial modeling and encryption techniques. The approach offers a robust solution to protect sensitive medical data in the digital realm. © The Author(s), under exclusive licence to Bharati Vidyapeeth's Institute of computer Applications and Management 2023. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s);author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

关键词： Bernstein polynomials Bezier-Bernstein polynomials Electrocardiogram (ECG) Signal decryption Signal encryption

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：