检索结果-内蒙古大学图书馆

2025 IEEE/CVF Winter Conference on Applications of computer Vision Workshops, WACVW 2025

作者： Perez, Aaron Prasad, Saurabh University of Houston Dept. of Electrical and Computer Engineering HoustonTX United States

ISBN: (纸本)9798331536626

Hyperspectral imaging (HSI) captures detailed spectral data across numerous contiguous bands, offering critical insights for applications such as environmental monitoring, agriculture, and urban planning. However, the high dimensionality of HSI data poses significant challenges for tra-ditional deep learning models, necessitating more efficient solutions. In this paper, we propose the Layer-Optimized Spatial-Spectral Transformer (LO-SST), a refined version of the Spatial-Spectral Transformer (SST) that incorporates structured layer pruning to reduce computational complexity while maintaining robust performance. LO-SST lever-ages self-supervised pretraining with a Masked Autoen-coder (MAE) framework, enabling the model to effectively learn spatial and spectral dependencies even in scenarios with limited labeled data. The use of separate spa-tial and spectral positional embeddings further enhances the model's ability to capture intricate relationships within hyperspectral data. Our experiments show that LO-SST achieves competitive segmentation accuracy while signifi-cantly reducing computational demands compared to traditional models. The effectiveness of random masking over alternative strategies during pretraining is also demonstrated, underscoring its ability to preserve critical image features. These results highlight the potential of LO-SST as an efficient and scalable solution for hyperspectral image segmen-tation, particularly in resource-constrained applications. © 2025 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

A Novel 3D Object Classification Paradigm through Reconfigurable Intelligent Surfaces

A Novel 3D Object Classification Paradigm through Reconfigur...

引用

IEEE Conference on Standards for Communications and Networking (CSCN)

作者： Stavros Tsimpoukis Dimitrios Tyrovolas George Karagiannidis Sotirios Ioannidis Christos Liaskos Institute of Computer Science – Foundation for Research and Technology Hellas Greece Dept. of Electrical and Computer Engineering Aristotle University of Thessaloniki Greece Dept. of Electrical Engineering Technical University of Crete Greece Dept. of Computer Science and Engineering University of Ioannina Greece

Programmable Wireless Environments (PWEs) leverage Reconfigurable Intelligent Surfaces (RIS) to convert the wireless propagation into a deterministic process. Recently, PWEs have shown promising results in boosting the efficiency of Radio-Frequency (RF) imaging, creating a novel, lightweight object detection and visualization approach for Extended Reality (XR). As a first step towards optimizing the PWE-XR synergy, this work proposes and compares a set of four PWE configuration policies. The goal is to deduce which policy yields the optimal classification of a set of arbitrary 3D objects present within the PWE. The rationale is that the PWE configuration policy that yields optimal object classification, will also yield the best XR quality in subsequent studies. Evaluation results regarding the performance of the proposed policies, based on ray-tracing, are demonstrated and discussed.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Novel Single-Stage Single-Phase Transformerless Grid-Connected Photovoltaic Inverter

A Novel Single-Stage Single-Phase Transformerless Grid-Conne...

引用

Annual Conference of Industrial Electronics Society

作者： Yuhao Liang Weimin Wu Houqing Wang Frede Blaabjerg Mohamed Orabi Liang Yuan Dept. of Electrical Engineering Shanghai Maritime University Shanghai China Dept. of Electrical Engineering Anhui University of Science and Technology Huainan China Dept. of Electrical Engineering University of Arkansas Arkansas United States Dept. of Energy Technology Aalborg University Aalborg Denmark Dept. of Electrical Engineering Aswan University Aswan Egypt Dept. of Safety Science and Engineering Anhui University of Science and Technology Huainan China

ISBN: (数字)9781665464543

ISBN: (纸本)9781665464550

This paper proposes a novel single-stage single-phase transformerless topology based on a buck-boost converter for grid-connected photovoltaic (PV) inverters. The proposed inverter has a wide input voltage range, low total harmonic distortion of current, and effective suppression of common-mode leakage current. The single-input structure ensures that it does not suffer from energy imbalance problems. The new inverter utilizes dead beat control and refines the method of controlling the inverter when the input energy is insufficient. The simulation results verify that the proposed grid-connected PV inverter maintains high grid-connected power quality both during normal operation under conventional conditions and when operating under discontinuous conduction mode (DCM) during energy insufficiency. The simulation models the operating mode of the inverter under abnormal environments, such as load changes and grid voltage transients, which proves that the proposed inverter under dead beat control has a better dynamic performance.

关键词： Photovoltaic systems Total harmonic distortion Time-frequency analysis Simulation Transformers Inverters Topology Leakage currents Voltage control Transient analysis

来源：评论

学校读者我要写书评

暂无评论

BlockShield: A TPM-Integrated Blockchain-Based Framework for Shielding Against Deepfakes 32

BlockShield: A TPM-Integrated Blockchain-Based Framework for...

引用

32nd IFIP/IEEE International Conference on Very Large Scale Integration, VLSI-SoC 2024

作者： Bathalapalli, Venkata K. V. V. Kumar, Aakarshan Mohanty, Saraju P. Kougianos, Elias Yanambaka, Venkata P. University of North Texas. Dept. of Computer Science and Engineering United States University of North Texas. Texas Academy of Mathematics and Science United States University of North Texas Dept. of Electrical Engineering United States School of the Sciences Texas Woman's University United States

ISBN: (纸本)9798331539672

The increasing threat to individual privacy and personalized digital content on social media posed by Deepfakes has highlighted the importance for a secure and reliable multimedia content integrity mechanism. In this paper a novel Blockchain driven hardware secure video attestation scheme, BlockShield is proposed for Deepfake mitigation. The proposed system includes a novel approach that ensures digital content traceability and privacy using Blockchain smart contracts and TPM's digital signature mechanism. The proposed work explores the scope of hardware-assisted security for Deepfake mitigation through a hardware TPM working together with Blockchain for enhanced digital media protection and sharing. The proposed work is experimentally validated and presented which validates the scope of hardware-assisted and blockchain integrated Deepfake mitigation framework. © 2024 IEEE.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

Analysis of Soiling Impacts on 100 kWp Rooftop Solar PV Plant

Analysis of Soiling Impacts on 100 kWp Rooftop Solar PV Plan...

引用

IEEE Students' Conference on electrical, Electronics and computer Science (SCEECS)

作者： Ankit Pal V. Atchaya Niranjana G. Nair Subhajit Ghosh L. Manivannan G. Saravana Ilango Dept. of Electrical and Electronics Engineering National Institute of Technology Tiruchirappalli India Dept. of Electrical and Electronics Engineering Amrita Vishwa Vidyapeetham Amritapuri India

ISBN: (数字)9798331529833

ISBN: (纸本)9798331529840

This paper gives a comprehensive analysis of the impact of soiling on the performance of a 100 kWP rooftop photovoltaic (PV) plant using a 2.5-year dataset from NIT Trichy. Traditional soiling measurement methods often rely on specialized equipment, which can limit their accessibility and scalability. This study eliminates the need for special sensors by using historical and real-time operational data, such as irradiance, panel temperature, and power output. Seasonal patterns in soiling losses are thus revealed, allowing optimization of cleaning schedules to maintain energy output at expected levels. Notably, the analysis shows that 34.2% of the plant's capacity remains unutilized due to soiling. This work contributes to the design of robust PV maintenance strategies by demonstrating a scalable data-driven approach to soiling assessment, making it specifically suitable for medium-sized installations.

关键词： Temperature sensors Temperature measurement Energy loss Schedules Soil measurements Estimation Production Cleaning Maintenance Optimization

来源：评论

学校读者我要写书评

暂无评论

LLM-Powered Grapheme-to-Phoneme Conversion: Benchmark and Case Study

arXiv

引用

arXiv 2024年

作者： Qharabagh, Mahta Fetrat Dehghanian, Zahra Rabiee, Hamid R. Dept. of Computer Engineering Sharif University of Technology Iran

Grapheme-to-phoneme (G2P) conversion is critical in speech processing, particularly for applications like speech synthesis. G2P systems must possess linguistic understanding and contextual awareness of languages with polyphone words and context-dependent phonemes. Large language models (LLMs) have recently demonstrated significant potential in various language tasks, suggesting that their phonetic knowledge could be leveraged for G2P. In this paper, we evaluate the performance of LLMs in G2P conversion and introduce prompting and post-processing methods that enhance LLM outputs without additional training or labeled data. We also present a benchmarking dataset designed to assess G2P performance on sentence-level phonetic challenges of the Persian language. Our results show that by applying the proposed methods, LLMs can outperform traditional G2P tools, even in an underrepresented language like Persian, highlighting the potential of developing LLM-aided G2P systems. Copyright © 2024, The Authors. All rights reserved.

关键词： Linguistics

来源：评论

学校读者我要写书评

暂无评论

Automation of FSM Verification Using Formal Tools

Automation of FSM Verification Using Formal Tools

引用

2023 International Conference on Communication, Circuits, and Systems, IC3S 2023

作者： Shambu, Amith Dalal, Vishal Naik, Basavaraj Sanjay, M. National Institute of Technology Calicut Dept. of Electrical Engg. India Intern DES Infineon Technologies Bangalore India Des Infineon Technologies Bangalore India Des Infineon Technologies Munich Germany National Institute of Technology Calicut Dept. of Electrical Engineering India

ISBN: (纸本)9798350325904

In today's design Finite state Machine (FSM) complexity is increasing tremendously. Exhaustive Design Verification (DV) of these FSM's using simulation at Register Transfer Level (RTL) is challenging due to possible combinations of stimuli. Formal Verification (FV) can efficiently verify them exhaustively. These FSM's undergo multiple design changes during development cycle and manual DV efforts must be repeated. An automated flow is needed to reduce this manual *** paper talks about a novel strategy and a flow called Property Generator (PropGen) to automatically generate System Verilog Assertions (SVA) directly from the FSM design specification captured in a Visio format. It also details various types of assertions generated by this flow. The paper also presents the results of applying the flow to a medium complex design for AI/ML acceleration. The verification effort in terms of time taken to create properties, the time required for the first formal verification report, and property re-generation for design changes is compared between manually written properties. The DUT FSM was verified exhaustively using FV and overall results show significant reduction in FSM DV efforts. © 2023 IEEE.

关键词： Formal verification

来源：评论

学校读者我要写书评

暂无评论

Implementation of a Camera Sensor for Helmet Detection in Two-Wheelers and Its Integration with the Electronic Control Unit (ECU)

Implementation of a Camera Sensor for Helmet Detection in Tw...

引用

IEEE Industrial Electronics and Applications Conference (IEACon)

作者： Akul P Abhijna MB Dhruvi T Avlani Dept. of Computer Science and Engineering Global Academy of Technology

ISBN: (数字)9798350382976

ISBN: (纸本)9798350382983

The governing bodies set across different regions have several rules and regulations for the safety of motorcyclists, which they must adhere to. Accidents are undesirable events that result in injuries, or sometimes even worse, deaths. These accidents could be averted by wearing helmets while riding. The idea of integrating the output of the image processing algorithms with the ignition system stems from the increased fatality rate of motorcyclists on the road. Therefore, this project aims to reduce such accidents by using a camera tracker as a sensor, where it is further implemented with image processing hardware, such as the Raspberry Pi Camera Module, which runs image processing algorithm to determine if helmet is being worn by the rider or not. Furthermore, Raspberry Pi Camera Module is connected to the ECU using appropriate communication protocols, and the ECU sends outputs that control the ignition system. For the helmet detection, control logics on microcontroller are developed in a way to send signals to the ECU regarding helmet detection status. This project is focused on improving safety and reducing accidents.

关键词： Visualization Accuracy Image processing Motorcycles Ignition Helmets Cameras Turning Real-time systems Accidents

来源：评论

学校读者我要写书评

暂无评论

XGBClassifier and Explainable AI for Disease Classification from Patient Symptom and Demographic Data with Late Fusion

XGBClassifier and Explainable AI for Disease Classification ...

引用

International Conference on electrical, computer and Communication engineering (ECCE)

作者： Md. Ismiel Hossen Abir Nurul Bashar Md. Shaek Ahmed Hasibul Islam Peyal Dept. of Computer Science & Engineering International Standard University Dhaka Bangladesh Dept. of Computer Science & Engineering Rajshahi University of Engineering & Technology Rajshahi Bangladesh

ISBN: (数字)9798350357509

ISBN: (纸本)9798350357516

The accurate identification of diseases based on patient symptoms and demographic data is a critical area of healthcare research, with the potential to significantly improve patient outcomes. In this study, we employed machine learning algorithms, specifically the XGBClassifier, to classify diseases based on a dataset containing patient symptoms and demographic information such as age, gender, fever, cough, fatigue, and other health indicators. To ensure model interpretability and transparency, we incorporated Explainable AI (XAI) techniques like SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations), which allow us to understand and interpret the feature contributions to the model’s predictions. The results demonstrated that the XGBClassifier achieved an accuracy of 81.42%, outperforming other machine learning models tested. This study emphasizes the importance of combining XAI with machine learning for disease classification, offering greater transparency in the decision-making process, which is vital in healthcare settings.

关键词： Support vector machines Electric potential Accuracy Machine learning algorithms Explainable AI Medical services Predictive models Fatigue Convolutional neural networks Diseases

来源：评论

学校读者我要写书评

暂无评论

Face Sketch Construction and Recognition

Face Sketch Construction and Recognition

引用

Cognitive Computing in engineering, Communications, Sciences and Biomedical Health Informatics (IC3ECSBHI), International Conference on

作者： Shaik Yacoob B. Sujatha Tamarana Anudeep Bozzam Eesha Sharvi Gande Veera Venkata Siva Sai Koppula Charles Antony Dept. of Computer Science and Engineering Godavari Global University Rajamahendravaram India Dept. of Computer Science and Engineering Godavari Institute of Engineering & Technology (A) Rajamahendravaram India

ISBN: (数字)9798331518523

ISBN: (纸本)9798331518530

As the crime rate increases rapidly, law enforcement organizations are facing the urgent task of quickly and properly identifying suspects. This paper proposes an innovative way to improve crime detection using new facial recognition technologies by describing a newly designed technology that allows users to create precise facial drawings using a user- friendly drag-and-drop interface. In which it will completely eliminate the need for expert forensic artists. This paper allows to create face drawings that can be effortlessly matched to huge police databases by utilizing the powerful deep learning algorithms and cloud infrastructure. By combining these modern technologies this tool speeds up the identification process by overcoming the problems and issues of typical handdrawn drawings. It will not only enhance the speed and accuracy of suspect identification, but it will also give law enforcement officers an efficient resource thereby, it will enhance the overall investigative effectiveness and responsiveness.

关键词： Deep learning computer vision Accuracy Law enforcement Databases Face recognition Forensics Organizations Cognitive systems Bioinformatics

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：