检索结果-内蒙古大学图书馆

CO-ResNet: Optimized ResNet model for COVID-19 diagnosis from X-ray images

学校读者我要写书评

暂无评论

International Journal of Hybrid Intelligent Systems 2021年第1-2期17卷 71-85页

作者： Bharati, Subrato Podder, Prajoy Mondal, M. Rubaiyat Hossain Prasath, V.B. Surya Institute of Information and Communication Technology Bangladesh University of Engineering and Technology Dhaka Bangladesh Division of Biomedical Informatics Cincinnati Children's Hospital Medical Center Cincinnati OH United States Department of Pediatrics University of Cincinnati College of Medicine Cincinnati OH United States Department of Biomedical Informatics College of Medicine University of Cincinnati Cincinnati OH United States Department of Electrical Engineering and Computer Science University of Cincinnati Cincinnati OH United States

This paper focuses on the application of deep learning (DL) based model in the analysis of novel coronavirus disease (COVID-19) from X-ray images. The novelty of this work is in the development of a new DL algorithm termed as optimized residual network (CO-ResNet) for COVID-19. The proposed CO-ResNet is developed by applying hyperparameter tuning to the conventional ResNet 101. CO-ResNet is applied to a novel dataset of 5,935 X-ray images retrieved from two publicly available datasets. By utilizing resizing, augmentation and normalization and testing different epochs our CO-ResNet was optimized for detecting COVID-19 versus pneumonia with normal healthy lung controls. Different evaluation metrics such as the classification accuracy, F1 score, recall, precision, area under the receiver operating characteristics curve (AUC) are used. Our proposed CO-ResNet obtains consistently best performance in the multi-level data classification problem, including health lung, pneumonia affected lung and COVID-19 affected lung samples. In the experimental evaluation, the detection rate accuracy in discerning COVID-19 is 98.74%, and for healthy normal lungs, pneumonia affected lungs are 92.08% and 91.32% respectively for our CO-ResNet with ResNet101 backbone. Further, our model obtained accuracy values of 83.68% and 82% for healthy normal lungs and pneumonia affected lungs with ResNet152 backbone. Experimental results indicate the potential usage of our new DL driven model for classification of COVID-19 and pneumonia. © 2021 - IOS Press. All rights reserved.

关键词： CO-ResNet coronavirus COVID-19 deep learning ResNet X-ray

SEMU5G: an Emulator for Improving Model Accuracy and Performance Evaluation of 5G Network Core Architectures based on SDN

学校读者我要写书评

暂无评论

SEMU5G: an Emulator for Improving Model Accuracy and Perform...

International Conference on Wireless and Telematics (ICWT)

作者： Eki Ahmad Zaki Hamidi Arga Valen Lazuardi Muhammad Litfan Rahmansyah Salman Alfarizi Novel Bajri Rohmat Tulloh Ridha Muldina Negara Syaiful Ahdan Jupriyadi Electrical Engineering Department UIN Sunan Gunung Djati Bandung Bandung Indonesia School of Electrical Engineering Telkom University Bandung Indonesia Telecommunication Technology School of Applied Science Telkom University Bandung Indonesia The University Center of Excellence Intelligent Sensing-IoT Telkom University Indonesia Faculty of Engineering & Computer Science Universitas Teknokrat Indonesia Indonesia School of Electrical Engineering and Informatics Institut Teknologi Bandung Bandung Indonesia Faculty of Engineering and Computer Science Universitas Teknokrat Indonesia Bandarlampung Indonesia

ISBN: (数字)9798350377057

ISBN: (纸本)9798350377064

The rapid advancement of 5G Radio Access Network (RAN) architecture is facilitating the construction of 5G networks, marking a significant milestone in telecommunications evolution. Given the complexity of the 5G core architecture, traditional simulation methods are insufficient, necessitating novel approaches. Emulator systems are crucial for creating dynamic, controlled environments that enable exploration of real-world scenarios without physical constraints. SEMU5G, a 5G core emulator, SDN controller, and RAN simulator, utilizes Open5GS implemented in Docker Container for test system flexibility and isolation. Additionally, an SDN controller is integrated to monitor data flows in the User Plane Function (UPF) and gNB simulated by UERANSIM in Mininet-WiFi. This comprehensive integration facilitates effective and flexible real-world exploration, providing a dynamic and controlled test environment for 5G core research. Scenario testing comprises two stages: firstly, a fixed network topology is employed to compare resource usage and confirm successful SEMU5G integration. Secondly, a mobile network topology is utilized to implement a mobile device scenario and compare the Quality of Service (QoS) of SEMU5G with other available emulators. These stages ensure thorough evaluation of SEMU5G's performance and its comparative advantage over existing solutions.

关键词： Performance evaluation 5G mobile communication Network topology Emulation computer architecture Quality of service Aerodynamics

Multi-platform Process Flow Models and Algorithms for Extraction and Documentation of Digital Forensic Evidence from Mobile Devices

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Ocen, Gilbert Gilibrays Bongomin, Ocident Mugeni, Gilbert Barasa Makau, Mutua Stephen Semwogerere, Twaibu Department of Computer Engineering Faculty of Engineering Busitema University Tororo Uganda Department of Manufacturing Textile and Industrial Engineering School of Engineering Moi University Eldoret Kenya Directorate of Innovation Research and Development Communication Authority of Kenya Nairobi Kenya Department of Computer Science School of Computing and Informatics Meru University of Science and Technology Meru Kenya

The increasing need for the examination of evidence from mobile and portable gadgets increases the essential need to establish dependable measures for the investigation of these gadgets. Many differences exist while detailing the requirement for the examination of each gadget, to help detectives and examiners in guaranteeing that of any kind piece of evidence extracted/ collected from any mobile devices is well documented and the outcomes can be repeatable, a reliable and well-documented investigation process must be implemented if the results of the examination are to be repeatable and defensible in courts of law. In this paper we developed a generic process flow model for the extraction of digital evidence in mobile devices running on android, Windows, iOs and Blackberry operating system. The research adopted survey approach and extensive literature review a s means to collect data. The models developed were validate through expert opinion. Results of this work can guide solution developers in ensuring standardization of evidence extraction tools for mobile devices. © 2022, CC BY.

关键词： Extraction

Mammo-CLIP: Leveraging Contrastive Language-Image Pre-training (CLIP) for Enhanced Breast Cancer Diagnosis with Multi-view Mammography

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Chen, Xuxin Li, Yuheng Hu, Mingzhe Salari, Ella Chen, Xiaoqian Qiu, Richard L.J. Zheng, Bin Yang, Xiaofeng Department of Radiation Oncology Winship Cancer Institute Emory University AtlantaGA30322 United States Department of Biomedical Engineering Emory University Georgia Institute of Technology AtlantaGA30308 United States Department of Biomedical Informatics Emory University AtlantaGA30322 United States School of Electrical and Computer Engineering University of Oklahoma NormanOK73019 United States

Although fusion of information from multiple views of mammograms plays an important role to increase accuracy of breast cancer detection, developing multi-view mammograms-based computer-aided diagnosis (CAD) schemes still faces big challenges and no such CAD schemes have been used in clinical practice. To overcome these challenges, we investigate a new approach based on the concept of Contrastive Language-Image Pre-training (CLIP), which has sparked interest across various medical imaging tasks. By solving the challenges in (1) effectively adapting the single-view CLIP for multi-view feature fusion and (2) efficiently fine-tuning this parameter-dense model with limited samples and computational resources, we introduce a unique Mammo-CLIP, the first multi-modal framework to process multi-view mammograms and corresponding simple texts. Mammo-CLIP uses an early feature fusion strategy to learn multi-view relationships in four mammograms acquired from the craniocaudal (CC) and mediolateral oblique (MLO) views of the left and right breasts. To enhance learning efficiency, plug-and-play adapters are added into CLIP’s image and text encoders for fine-tuning parameters and limiting updates to about 1% of the parameters. For framework evaluation, we assembled two datasets retrospectively. The first dataset, comprising 470 malignant and 479 benign cases, was used for few-shot fine-tuning and internal evaluation of the proposed Mammo-CLIP via 5-fold cross-validation. The second dataset, including 60 malignant and 294 benign cases, was used to test generalizability of Mammo-CLIP. Study results show that Mammo-CLIP outperforms the state-of-art cross-view transformer evaluated using areas under ROC curves (AUC= 0.841±0.017 vs. 0.817±0.012 and 0.837±0.034 vs. 0.807±0.036) on both datasets. It also surpasses previous two CLIP-based methods by 20.3% and 14.3% in AUC. Thus, this study highlights the potential of applying the finetuned vision-language models for developing next-genera

关键词： Mammography

Development of sandbox components with microservices architecture and design patterns in games 6

学校读者我要写书评

暂无评论

Development of sandbox components with microservices archite...

6th Information Systems International Conference, ISICO 2021

作者： Yunanto, Andhik Ampuh Hardiansyah, Fadilah Fahrul Putra, Adhiemas Andira Anantha Rasyid, Maulidan Bagus Afridian Arifiani, Siska Department of Informatics and Computer Engineering Politeknik Elektronika Negeri Surabaya Sukolilo Keputih East Java Surabaya60111 Indonesia PT Maulidan Teknologi Kreatif Klampis Ngasem Sukolilo Surabaya60117 Indonesia Department of Informatics Institut Teknologi Sepuluh Nopember Sukolilo Keputih East Java Surabaya60111 Indonesia

Nowadays, making games is more complicated than typical applications. Based on the survey, most of the architecture applied by the development industry still does not have a pattern. Therefore the game becomes difficult to develop. This study aims to create a sandbox by applying a microservice and design patterns in the game. The development of this research also involves the gaming industry to increase credibility. The results show that the platform developed has run smoothly and received good feedback from users. Therefore, the output of this research can be useful for developers in making games easy and efficient. © 2021 The Authors. Published by Elsevier B.V.

关键词： Architecture

Dynamics of hot random hyperbolic graphs

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Papadopoulos, Fragkiskos Zambirinis, Sofoclis Department of Electrical Engineering Computer Engineering and Informatics Cyprus University of Technology Limassol3036 Cyprus

We derive the most basic dynamical properties of random hyperbolic graphs (the distributions of contact and intercontact durations) in the hot regime (network temperature T > 1). We show that in the thermodynamic limit the contact distribution decays as a power law with exponent 2+T > 3 for durations t > T, while for t thmoment. Otherwise, the contact distribution depends on the expected degree distribution and we show that if the latter is a power law with exponent γ ∈ (2, T+1], then the former decays as a power law with exponent γ+1 > 3. On the other hand, the intercontact distribution exhibits power-law decays with exponent 2-T ∈ (0, 1) for T ∈ (1,2), while for T > 2 it displays linear decays with a slope that depends on the observation interval. This result holds irrespective of the expected degree distribution as long as it has a finite Tthmoment if T ∈ (1,2), or a finite second moment if T > 2. Otherwise, the intercontact distribution depends on the expected degree distribution and if the latter is a power law with exponent γ ∈ (2, 3), then the former decays as a power law with exponent 3-γ ∈ (0,1). Thus, hot random hyperbolic graphs can give rise to contact and intercontact distributions that both decay as power laws. These power laws however are unrealistic for the case of the intercontact distribution, as their exponent is always less than one. These results suggest that hot random hyperbolic graphs are not adequate null models for real temporal networks, in stark contrast to cold random hyperbolic graphs (T Copyright © 2021, The Authors. All rights reserved.

关键词： Graphic methods

Two-dimensional computer Modelling of Contaminant Migration in Unsaturated Porous Media with Nanoparticles

学校读者我要写书评

暂无评论

Two-dimensional Computer Modelling of Contaminant Migration ...

International computer Sciences and Information Technologies (CSIT)

作者： Viktor Zhukovskyy Nataliia Zhukovska Jan Perhac Department of Computer Science and Applied Mathematics The National University of Water and Environmental Engineering Rivne Ukraine Department of Computers and Informatics Technical University of Kosice Kosice Slovak Republic

ISBN: (纸本)9798350334326

The complicated physical problem of computer modeling pollutant migration in the zone of suspended water with nanoparticles to the catcher filter is discussed in this study. The brand-new mathematical model of nonlinearity in two dimensions is offered. The generalized Darcy-Clute laws for mass and moisture movement in various subregions of soil are taken into careful consideration in the model. The finite difference method was used to find a numerical solution to the boundary value problem and to provide an algorithm for computer implementation. Utilizing a newly created Nanosurface program solution, computer modeling was carried out along with the analysis of numerical experiments. The nanoparticle influence was demonstrated.

关键词： Nanoparticles Cloud computing Boundary value problems Computational modeling Soil Predictive models Mathematical models

Time-Domain Channel Estimation for Extremely Large MIMO THz Communications with Beam Squint

学校读者我要写书评

暂无评论

TechRxiv

TechRxiv 2023年

作者： Vlachos, Evangelos Kaushik, Aryan Eldar, Yonina C. Alexandropoulos, George C. Industrial Systems Institute ATHENA Research and Innovation Centre Rio-Patras26504 Greece School of Engineering and Informatics University of Sussex United Kingdom Faculty of Mathematics and Computer Science Weizmann institute of Science Israel Department of Informatics and Telecommunications National Kapodistrian University of Athens Athens15784 Greece

In this paper, we study the problem of extremely large (XL) multiple-input multiple-output (MIMO) channel estimation in the Terahertz (THz) frequency band, considering the presence of propagation delays across the entire array apertures, which leads to frequency selectivity, a problem known as beam squint. Multi-carrier transmission schemes which are usually deployed to address this problem, suffer from high peak-to-average power ratio, which is specifically dominant in THz communications where low transmit power is realized. Diverging from the usual approach, we devise a novel channel estimation problem formulation in the time domain for single-carrier (SC) modulation, which favors transmissions in THz, and incorporate the beam-squint effect in a sparse vector recovery problem that is solved via sparse optimization tools. In particular, the beam squint and the sparse MIMO channel are jointly tracked by using an alternating minimization approach that decomposes the two estimation problems. The presented performance evaluation results validate that the proposed SC technique exhibits superior performance than the conventional one as well as than state-of-the-art multi-carrier approaches. © 2023, CC BY.

关键词： Channel estimation

Adaptive Training of Grid-Dependent Physics-Informed Kolmogorov-Arnold Networks

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Rigas, Spyros Papachristou, Michalis Papadopoulos, Theofilos Anagnostopoulos, Fotios Alexandridis, Georgios Department of Digital Industry Technologies School of Science National and Kapodistrian University of Athens 34400 Greece Department of Physics School of Science National and Kapodistrian University of Athens 15784 Greece School of Electrical & Computer Engineering National Technical University of Athens 15780 Greece Department of Informatics & Telecommunications University of the Peloponnese 22131 Greece

Physics-Informed Neural Networks (PINNs) have emerged as a robust framework for solving Partial Differential Equations (PDEs) by approximating their solutions via neural networks and imposing physics-based constraints on the loss function. Traditionally, Multilayer Perceptrons (MLPs) have been the neural network of choice, with significant progress made in optimizing their training. Recently, Kolmogorov-Arnold Networks (KANs) were introduced as a viable alternative, with the potential of offering better interpretability and efficiency while requiring fewer parameters. In this paper, we present a fast JAX-based implementation of grid-dependent Physics-Informed Kolmogorov-Arnold Networks (PIKANs) for solving PDEs, achieving up to 84 times faster training times than the original KAN implementation. We propose an adaptive training scheme for PIKANs, introducing an adaptive state transition technique to avoid loss function peaks between grid extensions, and a methodology for designing PIKANs with alternative basis functions. Through comparative experiments, we demonstrate that the adaptive features significantly enhance solution accuracy, decreasing the L2 error relative to the reference solution by up to 43.02%. For the studied PDEs, our methodology approaches or surpasses the results obtained from architectures that utilize up to 8.5 times more parameters, highlighting the potential of adaptive, grid-dependent PIKANs as a superior alternative in scientific and engineering applications. © 2024, CC BY.

关键词： Multilayer neural networks