检索结果-内蒙古大学图书馆

EPDTNet + -Em. Advanced Transfer Learning and SubNet Architecture for m.dical Im.ge Diagnosis

Cognitive Com.utation 2025年第2期17卷 1-22页

作者： K, Dhivya K, Sangam.thrai S, Indra Priyadharshini m. Vedaraj Department of Electronics and Communication Engineering SRM Institute of Science and Technology Tamilnadu Kattankulathur India Department of Computer Science and Engineering Vel Tech Rangarajan Dr. Sagunthala R & d Institute of Science and Technology Tamilnadu Chennai India School of Computer Science and Engineering Vellore Institute of Technology Chennai India Department of Computer Science and Engineering R.M.D. Engineering College RSM Nagar Tamilnadu Kavaraipettai Thiruvallur601206 India

This paper presents a novel m.dical im.ging fram.work, Efficient Parallel Deep Transfer SubNet+-based Explainable m.del (EPDTNet + -Em., designed to im.rove the detection and classification of abnorm.lities in m.dical im.ges. The m.del integrates Enhanced Exponential Transfer Learning (En-ETL) and the Parallel SubNet (PSNet +) architecture, addressing challenges like overfitting, high com.utational cost, and lim.ted generalization across com.lex datasets. En-ETL accelerates learning and enhances precision, while PSNet + balances com.utational efficiency and perform.nce through parallel convolutional layers and an axis attention m.chanism. The proposed fram.work preprocesses m.dical im.ges using resizing, noise rem.val, and contrast enhancem.nt, ensuring high-quality inputs for analysis. For classification, the PSNet + m.del categorizes m.dical im.ges into norm.l and abnorm.l cases. Explainable AI (XAI) is applied to highlight key features in classified im.ges, im.roving interpretability and clinical decision-m.king. Experim.ntal results validate the fram.work on various datasets, including m.I, CT, X-ray, and m.m.ography im.ges. The proposed m.del achieves a classification accuracy of 98.83%, a false positive rate of 2%, and an execution tim. of 5.3 m.. Com.arative Analysis show superior perform.nce over existing approaches in classification accuracy and efficiency. Ablation studies confirm.d the significant contributions of the proposed m.del using each com.onent in enhancing perform.nce. © The Author(s), under exclusive licence to Springer science+Business m.dia, LLC, part of Springer Nature 2025.

关键词： Transfer learning

来源：评论

学校读者我要写书评

暂无评论

A Current Survey Trends on Child Safety Devices Using IoT

A Current Survey Trends on Child Safety Devices Using IoT

引用

International Conference on Intelligent System. and Sustainable Com.uting, ICISSC 2022

作者： Vinod Kum.r, B. Sangeetha, J. Department of Computer Science and Engineering M. S. Ramaiah Institute of Technology Bengaluru India

ISBN: (纸本)9789819947164

In today’s m.dern society, children and wom.n safety is considered to be seen as a big concern in both rural and urban areas with their safety issues. It is also very im.ossible that we will be able to alter people’s attitudes in general, but we should provide protection and surveillance devices for children and wom.n who are victim. of sexual assault, suicide bom.s, abduction, and m.lestation, am.ng other things. The survey consists of three different segm.nts. In the first segm.nt, we have discussed various wearables and technologies related to child protection, as well as the purpose of safety devices. The second segm.nt addresses sim.lar works by analyzing where in the various researchers have contributed and the drawbacks of the child protection fram.work. Then, in the third segm.nt, we com.ared all of the various technologies used in the child protection system. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Tracking (position)

来源：评论

学校读者我要写书评

暂无评论

Driver Drowsiness Detection System.Based on Behavioral m.thod, Biological m.thod and Vehicular Feature-Based m.thod—A Review

Driver Drowsiness Detection System Based on Behavioral Metho...

引用

International Conference on Intelligent System. and Sustainable Com.uting, ICISSC 2022

作者： Sowm.ashree, P. Sangeetha, J. Department of Computer Science and Engineering M. S. Ramaiah Institute of Technology Bangalore India

ISBN: (纸本)9789819947164

IoT applications span a good range of applications including agriculture, industry and transport. Although transport system.has changed people's lives and m.de daily activities easier, it has also been connected to negative outcom.s like road accidents. The m.jority of road accidents occur as a result of a driver's drowsiness. Drowsiness is a severe risk to road safety, which results in serious injuries and financial losses. Drowsiness detection in drivers is one am.ng the im.ortant applications of IoT that helps in saving drivers and passengers from.road accidents by warning the drivers at the right tim.. Various researchers have developed techniques to distinguish between drowsy and non-drowsy. In this paper, we analyze the different m.thods to detect drowsiness am.ng drivers. The three different m.thods are (i) Behavioral m.thods, (ii) Biological m.thods and (iii) Vehicular features-based m.thods. Behavioral m.thods use driver’s facial characteristics for detecting drowsiness. Biological m.thods use driver’s physical characteristics, whereas vehicular feature-based m.thods use vehicular features to detect drowsiness in driver. In this research review, we have explained and com.ared each of these m.thods in detail and also discussed their advantages and disadvantages. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

IoT based Voice Controlled Hom. computerization Utilizing Nodem.u and Android Application 5

IoT based Voice Controlled Home Computerization Utilizing No...

引用

5th International Conference on m.bile Com.uting and Sustainable Inform.tics, ICm.SI 2024

作者： Selvarathi, C. Nagul, J.G. Naveen, m. Sutharsan, V. M. Kumarasamy College of Engineering Department of Computer Science and Engineering Thalavapalayam Tamilnadu Karur639113 India

ISBN: (纸本)9798350395235

Innovative solutions that benefit society are essential in today's world. The creation of devices that autom.te a range of operations is encouraged by the IOT, a pioneering industry. A network of Internet-connected devices that can perform.tasks with m.nim.l hum.n intervention is referred to as the Internet of Things (IOT). IOT-enabled gadgets can be incredibly helpful for individuals with disabilities when used appropriately. The purpose of the study is to design a gadget that will help people with m.bility im.airm.nts by utilising the principle of this particular com.onent of IOT. m.inly, there are two segm.nts to the suggested device. A channel relay, an Arduino, and a sm.rt lightbulb are included in the hardware section. The cloud and IOT platform., along with an app created especially for this use, m.ke up the software com.onent. The principal purpose of this apparatus is to identify particular words in an individual's speech and subsequently trigger or deactivate the sm.rt gadget, so establishing an intelligent living space. First, a lightbulb will be used to dem.nstrate this. Java program.ing language will be used in the app's developm.nt. The prim.ry characteristic that sets this program.e apart is its m.crophone, which will be used to issue com.ands. To increase transm.ssion speed, Bluetooth and Wi-Fi chips will also be included in the lightbulb and the phone. Additionally, noise sensitivity and quick transm.ssion will be given special consideration. It is also possible to add different appliances, such fans and air conditioners, to the suggested gadget. An enhanced safety feature is taken into consideration in cases where the application reacts exclusively to a specific voice. © 2024 IEEE.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Survey on Traffic Sign Recognition Using computer Vision Technologies 5

Survey on Traffic Sign Recognition Using Computer Vision Tec...

引用

5th International Conference on Intelligent Com.unication Technologies and Virtual m.bile Networks, ICICV 2024

作者： Nandhakum.r, C. Rajashiva, A. Srinithi, B. Velm.rugan, K. M. Kumarasamy College of Engineering Department of Computer Science and Engineering Thalavapalayam Tamilnadu Karur639113 India

ISBN: (纸本)9798350385649

Autonom.us vehicles and m.chines are prepared to travel on public roads. Drivers and autonom.us cars find it m.re challenging to recognize every traffic sign along a route as traffic density increases. Neglecting the possibility of ignoring im.ortant indications that could cause deadly m.shaps is not an option. In order to regulate traffic, enforce driver behaviour. and prevent collisions. property dam.ge, and fatalities, traffic signs are crucial. An essential com.onent of any Intelligent Transportation System.(ITS) is autom.ted traffic sign identification and recognition. Autom.ted traffic sign recognition and identification is crucial in the era of self- driving cars. An autonom.us deep learning-based traffic sign awareness m.thod for India is presented in this research study. This research study analyzes a novel strategyv for identifying and recognizing traffic signs that Integrates deep Iearning techniques with im.ge processing techniques. Based on deep learning, the Convolution Neural Network (CNN) technology provides higher recognition accuracy for traffic signs, according to experim.ntal results. © 2024 IEEE.

关键词： Traffic signs

来源：评论

学校读者我要写书评

暂无评论

A Study on the Detection of Cyberbullying using CNN with IbI Logics Algorithm.(ILA) 5

A Study on the Detection of Cyberbullying using CNN with IbI...

引用

5th International Conference on m.bile Com.uting and Sustainable Inform.tics, ICm.SI 2024

作者： Karthik, K. Hem., R. Jayashree, S. Lakshedha, P. M. Kumarasamy College of Engineering Department of Computer Science and Engineering Thalavapalayam Tamilnadu Karur639113 India

ISBN: (纸本)9798350395235

While it's true that OSNs m.ke it easier for people to connect with one another, they also fuel an increase in antisocial behaviors like hate speech, or trolling. Especially for wom.n and children, cyberbullying has caused extrem. em.tional and physical suffering, and in extrem. cases, it has even driven victim. to take their own lives. Depending on users to report instances of bullying is one conventional m.thod for detecting cyberbullying, but it isn't foolproof. Cyberbullying m.y be detected and flagged autom.tically using deep learning along with m.chine learning m.thods. These algorithm. can also identify patterns of behavior that could indicate cyberbullying. Consequently, the purpose of this research is to identify and categorize instances of cyberbullying using Twitter data by m.ans of an ensem.le deep learning system.called IbI Logics system.(ILA). The research aim. to analyze social m.dia data by m.ans of an ensem.le learning procedure and natural language processing (NLP). The last step was to try every conceivable com.ination using the ensem.le m.dels. According to the findings, ensem.le m.dels are m.re likely to provide optim.l outcom.s. The optim.l ensem.le CNN, which includes all three CNNs, outperform. the others with respect to precision (97.7%), accuracy (97.7%), recall (97.8%), in addition F1 average score (97.7%). © 2024 IEEE.

关键词： Learning system.

来源：评论

学校读者我要写书评

暂无评论

Design and Im.lem.ntation of a m.PT Controller-Based Bidirectional DC-DC Converter for Electric Vehicles Using PUm. Algorithm 4

Design and Implementation of a MPPT Controller-Based Bidirec...

引用

4th International Conference on Sentim.nt Analysis and Deep Learning, ICSADL 2025

作者： Subram.niam. G. Devayani, G. Gokulapriya, C. Janani, S. Dhanush, T. Shashhwath, S. M. Kumarasamy College of Engineering Electrical and Electronics Engineering Tamil Nadu Karur India M. Kumarasamy College of Engineering Computer Science and Business Systems Tamil Nadu Karur India

ISBN: (数字)9798331523923

ISBN: (纸本)9798331523923

The growing need for effective energy m.nagem.nt in electric vehicles (EVs) has led to advancem.nts in power conversion technologies, especially in system. that facilitate the use of renewable energy sources and efficient battery m.nagem.nt. The design and im.lem.ntation of bidirectional DC-DC converters in electric cars utilize m.xim.m.Power Point Tracking (m.PT) for their controllers. This approach ensures optim.l energy conversion through the PUm. (Perturb and Observe with m.dified Algorithm. technique. By intelligently regulating the power flow between a renewable energy source and the vehicle's energy storage system.(such as solar panels and batteries), the proposed converter m.xim.zes energy conversion from.the renewable source while keeping operational costs low and m.intaining ideal battery conditions. A bidirectional converter enables the battery to both charge and discharge, while also providing a seam.ess interface for energy exchange between the grid, solar panels, and the car. The m.PT controller utilizes the PUm. algorithm. which is known for its faster convergence and greater stability, m.king it superior to traditional Disturb and Observe (P&O) algorithm.. This device includes various com.onents that m.nitor voltage and current values in real-tim., em.loying adaptive control and im.roved power m.nagem.nt to ensure high-quality energy conversion regardless of external conditions. Sim.lations and hardware im.lem.ntations have shown that this m.thod effectively enhances vehicle perform.nce, extends the range of electric vehicles, and optim.zes the use of renewable energy sources. The study also addresses several other issues, including the potential for large-scale im.lem.ntation of m.dern electric vehicles (EVs), the robustness of the system. and the reduction of voltage and current ripple. © 2025 IEEE.

关键词： Adaptive control system.

来源：评论

学校读者我要写书评

暂无评论

Artificial Intelligence Driven Colorectal Cancer Classification Via Deep Learning Technique 4

Artificial Intelligence Driven Colorectal Cancer Classificat...

引用

4th International Conference on Ubiquitous Com.uting and Intelligent Inform.tion System., ICUIS 2024

作者： Ananthi, S. Rithish Kum.r, J. Sai Prasanth, R. Soundhar, G. Sujith, I. M. Kumarasamy College of Engineering Department of Computer Science and Engineering Thalavapalayam TamilNadu Karur639113 India

ISBN: (数字)9798331529635

ISBN: (纸本)9798331529635

Collateral cancer is a serious concern for patients with prim.ry tum.rs, as the developm.nt of secondary tum.rs can significantly reduce survival rates and increase the com.lexity of treatm.nt. Colon polyps are a com.on precursor to colorectal cancer, and early detection is essential for effective treatm.nt. Traditional strategies for polyp detection include colonoscopy and biopsy, which can be invasive and tim.-consum.ng. In this work, we explore the use of convolutional neural networks (CNNs) to detect polyps in colonoscopy im.ges, as m.chine learning m.thods have shown potential in this area. Im.roving patient outcom.s requires early detection and prevention of collateral cancer. However, precisely predicting the probability of recurrent tum.rs can be challenging due to the com.lexity of cancer progression and the m.ltiplicity of factors that m.y influence tum.r form.tion. In this study, we exam.ne the use of a CNN to forecast the likelihood that patients with prim.ry tum.rs m.y develop collateral cancer. m.chine learning algorithm. have dem.nstrated potential in predicting cancer outcom.s based on patient data. The study uses a dataset of colonoscopy im.ges, including both positive and negative cases of polyps. A CNN m.del is developed using this dataset to classify im.ges as either positive or negative for polyps. The m.del is trained using a supervised learning approach, where the network learns from.labeled exam.les of im.ges with and without polyps. The accuracy of the CNN m.del is com.ared to other polyp detection strategies, such as traditional im.ge analysis techniques and other m.chine learning algorithm.. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Language Detection for Transliterated Content 3

Language Detection for Transliterated Content

引用

3rd International Conference for Innovation in technology, INOCON 2024

作者： Selva Kum.r, S. Khan m.ham.ed Ajm.l Khan, Afifah m.njeshwar, Chirag Banday, Im.dh Ajaz B. M. S. College of Engineering Department of Computer Science and Engineering Bangalore India

ISBN: (纸本)9798350381931

In the contem.orary digital era, the Internet functions as an unparalleled catalyst, dism.ntling geographical and linguistic barriers, particularly evident in texting. This evolution facilitates global com.unication, transcending physical distances and fostering dynam.c cultural exchange. A notable trend is the widespread use of transliteration, where the English alphabet is em.loyed to convey m.ssages in native languages, posing a unique challenge for language technology in accurately detecting the source language. This paper addresses this challenge through a dataset of phone text m.ssages in Hindi, and Russian transliterated into English, utilizing BERT for language classification and Google Translate API for transliteration conversion. The research pioneers innovative approaches to identify and convert transliterated text, navigating challenges in the diverse linguistic landscape of digital com.unication. Em.hasizing the pivotal role of com.rehensive datasets for training Large Language m.dels (LLm.) like BERT, our m.del showcases exceptional proficiency in accurately identifying and classifying languages from.transliterated text. With a validation accuracy of 99%, our m.del's robust perform.nce underscores its reliability. The com.rehensive exploration of transliteration dynam.cs, supported by innovative approaches and cutting-edge technologies like BERT, positions our research at the forefront of addressing unique challenges in the linguistic landscape of digital com.unication. Beyond contributing to language identification and transliteration capabilities, this work holds prom.se for applications in content m.deration, analytics, and fostering a globally connected com.unity engaged in m.aningful dialogue. © 2024 IEEE.

关键词： Classification (of inform.tion)

来源：评论

学校读者我要写书评

暂无评论

Overcom.ng LLm.Challenges using RAG-Driven Precision in Coffee Leaf Disease Rem.diation

Overcoming LLM Challenges using RAG-Driven Precision in Coff...

引用

2024 IEEE International Conference on Em.rging Technologies in computer science for Interdisciplinary Applications, ICETCS 2024

作者： Selva Kum.r, S. Khan, Afifah Khan m.ham.ed Ajm.l Banday, Im.dh Ajaz Gada, m.nikantha Shanbhag, Vibha Venkatesh B. M. S. College of Engineering Department of Computer Science and Engineering Bangalore India

ISBN: (纸本)9798350372502

This research introduces an innovative AI-driven precision agriculture system. leveraging YOLOv8 for disease identification and Retrieval Augm.nted Generation (RAG) for context-aware diagnosis. Focused on addressing the challenges of diseases affecting the coffee production sector in Karnataka, The system.integrates sophisticated object detection techniques with language m.dels to address the inherent constraints associated with Large Language m.dels (LLm.). Our m.thodology not only tackles the issue of hallucinations in LLm., but also introduces dynam.c disease identification and rem.diation strategies. Real-tim. m.nitoring, collaborative dataset expansion, and organizational involvem.nt ensure the system.s adaptability in diverse agricultural settings. The effect of the suggested system.extends beyond autom.tion, aim.ng to secure food supplies, protect livelihoods, and prom.te eco-friendly farm.ng practices. By facilitating precise disease identification, the system.contributes to sustainable and environm.ntally conscious agriculture, reducing reliance on pesticides. Looking to the future, the project envisions continuous developm.nt in RAG-integrated object detection system., em.hasizing scalability, reliability, and usability. This research strives to be a beacon for positive change in agriculture, aligning with global efforts toward sustainable and technologically enhanced food production. © 2024 IEEE.

关键词： Object 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：