检索结果-内蒙古大学图书馆

4th International Conference on Ubiquitous Computing and Intelligent Information Systems, ICUIS 2024

作者： Uke, Shailaja Junghare, Pranay Kenjale, Shrishti Korade, Srushti Kothwade, Aniket Vishwakarma Institute of Technology Dept of Computer Engineering Pune India

ISBN: (纸本)9798331529635

The objective of this project is to develop an advanced quality human intrusion detection system, integrating IoT hardware with advanced software technologies. This will be done by relying on real-time video footage and using image processing techniques in the identification process of human presence and other potential threats, such as weapons. A face recognition module will be attached, which will let the entry of only authorized people and immediately detect unauthorized intrusions. The system includes live alerts for efficient monitoring and protection of secure places. This system, with high algorithms used and strong hardware, is scalable and reliable, proving to be fully comprehensive in relation to the potential threats against the security offered. © 2024 IEEE.

关键词： Network intrusion

来源：评论

学校读者我要写书评

暂无评论

Sherlock in OSS: A Novel Approach of Content-Based Searching in Object Storage System

引用

IEEE ACCESS 2024年 12卷 69456-69474页

作者： Noor, Jannatun Sakif, Md. Sadiqul Islam Mondal, Joyanta Jyoti Rownak Ali Uday, Mir Haque Ratul, Rizwanul Chellappan, Sriram Alim Al Islam, A. B. M. BRAC Univ Sch Data & Sci Comp Sustainabil & Social Good C2SG Res Grp Dhaka 1212 Bangladesh Bangladesh Univ Engn & Technol Dept CSE Next Generat Comp Res Grp Dhaka 1000 Bangladesh Univ Alabama Birmingham Dept Comp Sci Birmingham AL 35294 USA Univ S Florida Dept Comp Sci Tampa FL 33620 USA

Cloud-based Object Storage Systems (OSS) are known for their scalability, durability, availability, and concurrency. However, there is a noticable vaccum in open-source OSS for a straightforward way for users and administrators to conduct data searches within object storage without fully utilizing the cloud infrastructure. In our research, we present Sherlock, a novel Content-Based Searching (CoBS) framework. Sherlock enhances search capabilities by using extra information from images and documents, incorporating this information into an Elasticsearch-powered database to enable content-driven searches. The framework operates through a two-stage process. First, it classifies the incoming data by type, directing images to an object detection model and processing documents for keyword extraction. Then, Elasticsearch catalogs the extracted data, facilitating searches based on content. The effectiveness of our searches is largely dependent on the precision of these models, which we improve by training them on large-scale datasets: the Microsoft COCO Dataset for multimedia content and the SemEval2017 Dataset for text documents. We further test our system's performance by integrating it with the open-source OSS, OpenStack Swift, and conducting real-world experiments with image uploads to evaluate how our model performs within Swift's object storage environments.

关键词： Content-based searching (CoBS) content-based image retrieval (CBIR) deep learning OpenStack Swift object storage system (OSS) distributed systems

来源：评论

学校读者我要写书评

暂无评论

SpecReFlow: an algorithm for specular reflection restoration using flow-guided video completion

引用

JOURNAL OF MEDICAL IMAGING 2024年第2期11卷 024012页

作者： Yin, Haoli Eimen, Rachel Moyer, Daniel Bowden, Audrey K. Vanderbilt Univ Dept Comp Sci Nashville TN USA Vanderbilt Univ Vanderbilt Biophoton Ctr Nashville TN 37235 USA Vanderbilt Univ Dept Biomed Engn Nashville TN 37235 USA Vanderbilt Univ Dept Elect & Comp Engn Nashville TN 37235 USA

Purpose Specular reflections (SRs) are highlight artifacts commonly found in endoscopy videos that can severely disrupt a surgeon's observation and judgment. Despite numerous attempts to restore SR, existing methods are inefficient and time consuming and can lead to false clinical interpretations. Therefore, we propose the first complete deep-learning solution, SpecReFlow, to detect and restore SR regions from endoscopy video with spatial and temporal coherence. Approach SpecReFlow consists of three stages: (1) an image preprocessing stage to enhance contrast, (2) a detection stage to indicate where the SR region is present, and (3) a restoration stage in which we replace SR pixels with an accurate underlying tissue structure. Our restoration approach uses optical flow to seamlessly propagate color and structure from other frames of the endoscopy video. Results Comprehensive quantitative and qualitative tests for each stage reveal that our SpecReFlow solution performs better than previous detection and restoration methods. Our detection stage achieves a Dice score of 82.8% and a sensitivity of 94.6%, and our restoration stage successfully incorporates temporal information with spatial information for more accurate restorations than existing techniques. Conclusions SpecReFlow is a first-of-its-kind solution that combines temporal and spatial information for effective detection and restoration of SR regions, surpassing previous methods relying on single-frame spatial information. Future work will look to optimizing SpecReFlow for real-time applications. SpecReFlow is a software-only solution for restoring image content lost due to SR, making it readily deployable in existing clinical settings to improve endoscopy video quality for accurate diagnosis and treatment.

关键词： specular reflection image artifacts image restoration optical flow multiview restoration endoscopy

来源：评论

学校读者我要写书评

暂无评论

Intelligent Surface Plasmon Resonance Sensor for Refractive Index Substance Identification With Convolutional Neural Network image-Based Model

引用

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT 2025年 74卷

作者： Oliveira, Leiva C. Sales, Adna Q. Lopes, Mateus V. L. Lima, Antonio M. N. Univ Fed Rural Semi Arido Dept Comp Elect Engn Grad Program BR-59625900 Mossoro RN Brazil Univ Fed Rural Semi Arido Dept Comp Comp Sci Grad Program BR-59625900 Mossoro RN Brazil Univ Fed Rural Semiaido UFERSA Elect Engn Grad Program Mossoro RN Brazil Univ Fed Rural Semi Arido Comp Sci Grad Program Mossoro RN Brazil Univ Fed Campina Grande UFCG Dept Elect Engn Campina Grande PB Brazil

The surface plasmon resonance (SPR) sensors are technologically attractive for applications that demand quick and accurate biological substance monitoring. Through its typical SPR image response, the resonance condition indicated by the minimum reflectivity values works like an optical signature for changes in the refractive index (RI) of the substance under analysis. Recently, the incorporation of machine and deep learning methods (MDLMs) on SPR sensors to create intelligent tasks along the signal processing chain employed in SPR biosensing was witnessed. One possible intelligent application is substance identification based on the analysis of SPR responses. Occasionally, this problem is addressed with data from SPR curves, requiring a prior SPR image manipulation for the respective curve generation, leading to extra process steps and time consumption. This article presents the design of an intelligent SPR sensor with analyte identification capabilities directly from its SPR image, offering guidance on the precise moment for substance switching during injection routines. An image-based prediction model with convolutional neural networks (CNNs) was fine-tuned to directly identify individually aqueous solutions with different refractive indices. A new approach was described to generate SPR images (Fresnel images) from calculation with the Fresnel analysis (FA) framework. The proposed CNN architecture was evaluated and compared with seven state-of-the-art CNN architectures. The models were integrated into the experimental setup for real-time identification. The experimental tests demonstrate the viability of the overall pipeline for the model conception, being able to reach more than 96% accuracy in performing the identification task.

关键词： Surface plasmon resonance Convolutional neural networks Feature extraction Training Sensors Resonance Optical imaging Optical reflection Refractive index Predictive models Convolutional neural network (CNN) intelligent sensor surface plasmon resonance (SPR) image substance identification

来源：评论

学校读者我要写书评

暂无评论

real-time neural-network-based denoising for intraoperative 4D-OCT 27

Real-time neural-network-based denoising for intraoperative ...

引用

Conference on Optical Coherence Tomography and Coherence Domain Optical Methods in Biomedicine XXVII

作者： Nienhaus, Jonas Matten, Philipp Britten, Anja Schlegl, Thomas Hoeck, Eva Freytag, Alexander Everett, Matt Hecker-Denschlag, Nancy Drexler, Wolfgang Leitgeb, Rainer A. Schmoll, Tilman Med Univ Vienna Ctr Med Phys & Biomed Engn Vienna Austria Carl Zeiss Jena Germany Carl Zeiss Meditec Inc Dublin CA USA Carl Zeiss Meditec AG Jena Germany

ISBN: (纸本)9781510658394;9781510658400

Noise decreases image quality in optical coherence tomography (OCT) and can obscure important features in real-time visualizations. In this work, we show that a neural network can be applied to denoise volumetric OCT data for intra-surgical visualization in real-time. We adapt a self-supervised training approach, not requiring any paired data for training. Several optimizations and trade-offs in deployment are required, with which we achieved processing times of only few milliseconds. While still being limited by the real-time requirements, denoising in this scenario can enhance surface visibility, and therefore allow guidance for more precise intra-surgical maneuvers.

关键词： Optical coherence tomography denoising deep learning self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Semantic Segmentation of Brain Tumors: A Performance Evaluation Using deepLabV3+, UNet, and Intel's OpenVINO Toolkit 10

Semantic Segmentation of Brain Tumors: A Performance Evaluat...

引用

10th International Conference on Control, Decision and Information Technologies (CoDIT)

作者： Kapo, Medina Akagic, Amila Buza, Emir Univ Sarajevo Fac Elect Engn Dept Comp Sci & Informat Sarajevo 71000 Bosnia & Herceg

ISBN: (纸本)9798350373981;9798350373974

Artificial intelligence, Machine learning, and deep learning are increasingly making significant contributions to the field of medicine. Individual patient conditions, disease localization, and various influencing factors underscore the complexity of disease diagnosis and treatment planning. Introducing new technologies can revolutionize medical diagnostics, facilitating swift and accurate assessments. Among the noninvasive diagnostic methods, Magnetic Resonance Imaging (MRI) stands out, particularly in tumor diagnosis. UNet, renowned for its effectiveness in medical image analysis, serves as a robust model for semantic segmentation, as does deepLabV3+. However, these models are inherently complex, and their inference process can be time-consuming. By leveraging the OpenVINO toolkit, the inference process is significantly reduced. In this study, nearly a 2-fold acceleration is achieved in inference time with the deepLabV3+ model and a roughly 1.2-fold improvement with the UNet model on CPU. Moreover, when employing GPU with FP16 precision, the acceleration reached almost 2.5-fold for UNet and nearly 3-fold for deepLabV3+, showcasing the substantial performance enhancements attainable through optimized hardware utilization.

关键词： Artificial Intelligence Neural Networks image processing

来源：评论

学校读者我要写书评

暂无评论

Robustness of ConvNet to High-Frequency image Corruptions 8th

Robustness of ConvNet to High-Frequency Image Corruptions

引用

8th International Conference on Computer Vision and image processing (CVIP)

作者： Banerjee, Arnab Zoom Video Commun India Pvt Ltd Mumbai Maharashtra India

ISBN: (纸本)9783031581731;9783031581748

image processing techniques such as blurring, JPEG compression are applied to natural images to meet different objectives. Additionally, corruptions such as Gaussian and shot noise appear on images due to digital fluctuations. Unfortunately, standard vision models tend to perform quite poorly under such unavoidable corruptions, i.e., these models are not robust to the distribution shifts induced by these corruptions at test time. The standard approach for overcoming this issue for a known corruption is by augmenting the training data with images perturbed using the corruption of interest. Motivated by settings where the corruption might not be known during training, Gaussian noise is used as an augmentation strategy to gain robustness to high-frequency corruptions. In this paper, we try to understand its properties from a Fourier lens. However, we show that Gaussian augmentation fails to maintain robustness to few high-frequency corruptions at high severity levels. Analyzing the Fourier signature of those corruptions reveal a change in behavior - at high severity they corrupt low frequencies as well. A Gaussian-trained model loses its performance due to this change. Current augmentation strategies for low-frequency corruptions are discussed at the end.

关键词： Fourier transform synthetic corruptions deep learning data augmentation

来源：评论

学校读者我要写书评

暂无评论

Optimizing real-time Freshness: deep Joint Source-Channel Coding Based AoI in Wireless Networks

Optimizing Real-Time Freshness: Deep Joint Source-Channel Co...

引用

2024 IEEE Global Communications Conference, GLOBECOM 2024

作者： Basnayaka, Chathuranga M. Wijerathna Jayakody, Dushantha Nalin K. Beko, Marko Universidade Autónoma de Lisboa Centro de Investigaçäo em Tecnologias Autónoma TechLab Portugal Copelabs Lusófona University Lisbon1700-097 Portugal Instituto de Telecomunicações Instituto Superior Técnico Universidade de Lisboa Lisbon1049-001 Portugal

ISBN: (纸本)9798350351255

This paper proposes a deep joint source-channel coding (DJSCC) to minimize the age of information (AoI) for image transmission. A new content-based AoI metric called age of misclassified information (AoMI) is introduced to estimate the freshness of the information in an image classification system. AoMI is a critical metric in timely information delivery, measuring the age of the most recently received and correctly classified image at the receiver. The proposed system leverages a deep neural network at the transmitter to map image pixels directly to channel input symbols, eliminating the need for separate source and channel coding. At the receiver, the channel output is processed to perform image classification. To analyze the AoMI performance of the system, a stochastic hybrid systems (SHS) approach is employed. Closed-form expressions for the average AoMI (AAoMI) are derived, providing insights into the impact of system parameters on the AoMI. Simulation results demonstrate the effectiveness of the proposed DJSCC-based system in achieving lower AoMI compared to traditional separate source and channel coding schemes. The findings highlight the potential of deep learning techniques to maintain the freshness of the information in wireless communication systems. This work paves the way for the design of wireless communication systems that prioritize the freshness of delivered information - this is crucial in applications such as real-time monitoring, surveillance, and control systems. © 2024 IEEE.

关键词： Stochastic control systems

来源：评论

学校读者我要写书评

暂无评论

A deep learning SYSTEM TO QUANTIFY AND PREDICT THE CHEWING PROCESS OF FOODS

A DEEP LEARNING SYSTEM TO QUANTIFY AND PREDICT THE CHEWING P...

引用

ASME International Mechanical Engineering Congress and Exposition (IMECE)

作者： Ren, Shiyi Chen, Bangxiang Dhupia, Jaspreet Stommel, Martin Xu, Weiliang Univ Auckland Dept Mech & Mechatron Engn Auckland New Zealand Auckland Univ Technol Sch Engn Comp & Math Sci Auckland New Zealand

ISBN: (纸本)9780791887639

The real-time assessment and prognostication of food texture during mastication are paramount for comprehending and emulating the chewing process, bearing significant ramifications for the food industry. This study unveils a deep learning-oriented recognition framework utilizing Mask R-CNN to analyze food texture throughout the mastication process. We create an extensive food bolus image compendium and train a deep learning model proficient in texture discrimination and processing parameter estimation. A vision system, incorporating an Intel realSense D435i camera, is harnessed to capture high-resolution images of the food bolus. The dataset encompasses images of masticated peanut samples, with the number of chews varying from 1 to 12 cycles. Although the model's predictions occasionally diverge from the actual data points, with a maximum discrepancy of +/- 2 chewing instances chews except for predicting peanuts chewed 11 cycles, the proposed system lays the groundwork for real-time evaluation and prediction of food texture during mastication. Future inquiries could concentrate on ameliorating the model's precision, broadening its applicability to diverse food categories, and refining the training dataset. This investigation holds the potential to influence the advancement of delectable and nourishing food products, bestowing benefits upon both the food industry and consumers.

关键词： Chewing Vision system Mask R-CNN real time evaluation Oral food process

来源：评论

学校读者我要写书评

暂无评论

Enhancing Fault Diagnosis Accuracy in Electric Motors: A Digital Twin Approach with Transformer Model 14

Enhancing Fault Diagnosis Accuracy in Electric Motors: A Dig...

引用

14th Asian Control Conference (ASCC)

作者： Rosyadi, Imron Nazaruddin, Yul Yunazwin Siregar, Parsaulian Ishaya Inst Teknol Bandung Doctoral Program Engn Phys Fac Ind Technol Bandung Indonesia Univ Jenderal Soedirman Elect Engn Dept Purwokerto Indonesia Inst Teknol Bandung Dept Engn Phys Instrumentat & Control Res Grp Bandung Indonesia Inst Teknol Bandung Natl Ctr Sustainable Transporat Technol CRCS Bldg2nd Floor Bandung Indonesia

ISBN: (纸本)9798331540845;9789887581598

This paper explores the integration of Digital Twin (DT) technology and machine learning (ML) techniques for predictive maintenance in electric vehicles (EVs), particularly focusing on fault diagnosis in electric motors (EMs). DT technology, which creates a digital replica of a physical entity, combined with real-time sensor data analytics, offers significant potential for improving the reliability and operational efficiency of EMs under variable and intense conditions. By incorporating advanced ML models like Transformers, which are adept at capturing long-range dependencies, into the DT framework, the study addresses critical challenges in fault diagnosis under non-stationary conditions typical in EVs. The research introduces a novel approach that combines Transformers with convolutional layers to enhance the fault diagnosis capabilities by effectively capturing complex patterns in time-series data. Furthermore, incremental learning is proposed within the DT framework to adapt to evolving environments and maintain high diagnostic performance, aiming for domain invariance in fault diagnosis models. The paper evaluates various deep learning models, including Transformers, FCNs, CNNs, Inception, ResNet, and MLP, across metrics such as accuracy, training duration, and inference time, identifying CNNs and FCNs as optimal due to their balance between performance and processing time. This study not only demonstrates the feasibility of advanced ML techniques in enhancing DT-based fault diagnosis but also provides insights into selecting appropriate models for predictive maintenance applications in EVs, highlighting the potential for future research in adapting these technologies to different datasets and operational scenarios.

关键词： digital twin fault diagnosis electric motors deep learning

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：