检索结果-内蒙古大学图书馆

Optimizing hyperparameters for dual-attention network in lung segmentation

Intelligence-Based Medicine 2025年 11卷 100221-100221页

作者： Wahyuningrum, Rima Tri Fadillah, Rizki Abdil Yunita, Indah Satoto, Budi Dwi Muntasa, Arif Sari, Amillia Kartika Rahardjo, Paulus Dewi, Deshinta Arrova Bauravindah, Achmad Department of Informatics Engineering Faculty of Engineering University of Trunojoyo Madura Indonesia Department of Health Faculty of Vocational Studies Universitas Airlangga Indonesia Department of Computer Science and Technology Faculty of Data Science and Information Technology INTI International University Malaysia Department of Informatics Faculty of Industrial Technology Islamic University of Indonesia Indonesia

Medical imaging, particularly chest X-rays (CXR), is a cornerstone in the diagnosis of lung diseases, such as pneumonia, tuberculosis and COVID-19, owing to its accessibility and effectiveness. However, the sheer volume of CXR images, especially during pandemics, combined with the complexity of subtle abnormalities, poses significant challenges for manual analysis. Lung segmentation plays a pivotal role in artificial intelligence-driven CXR analysis by isolating lung fields, which facilitates the detection of disease-affected regions. Recent advances in deep learning, particularly with attention mechanisms, have improved segmentation accuracy, but the performance of these models heavily depends on the selection of appropriate hyperparameters. This study investigates the impact of key hyperparameters—learning rate and number of epochs—on the performance of the dual-attention network (DANet) in lung segmentation tasks. DANet was tested on a CXR dataset from Qatar University and evaluated under four different hyperparameter configurations: 20 epochs with a learning rate of 0.001, 10 epochs with a learning rate of 0.001, 10 epochs with a learning rate of 0.0001 and 20 epochs with a learning rate of 0.0001. The model's performance was assessed using two widely recognised segmentation metrics: the Dice coefficient and Intersection over Union (IoU). The results indicated that higher learning rates and greater numbers of epochs lead to improved segmentation performance. Specifically, the DANet model achieved a Dice coefficient of 97.29 % and an IoU value of 94.74 %, demonstrating its effectiveness compared to other models. These findings highlight the importance of hyperparameter tuning in achieving high segmentation accuracy and demonstrate the potential of the DANet model to improve diagnostic workflows for CXR analysis. © 2025 The Authors

关键词： Chest X-ray images DANet Epoch Hyperparameter Learning rate Lung segmentation Process innovation

来源：评论

学校读者我要写书评

暂无评论

A Review of Extremophile Protein Classification Using Machine Learning 8

A Review of Extremophile Protein Classification Using Machin...

引用

8th International Conference on Wireless and Telematics, ICWT 2022

作者： Susanty, Meredita Rajab, Tati Latifah Erawati Hertadi, Rukman Bandung Institute of Technology School of Electrical Engineering and Informatics Bandung Indonesia Universitas Pertamina School of Computer Science Indonesia Bandung Institute of Technology Faculty of Math and Natural Science Bandung Indonesia

ISBN: (数字)9781665451222

ISBN: (纸本)9781665451222

Proteins can be denatured under extreme conditions due to breaking non-covalent bonds between protein side chains, causing conformational changes and loss of function. However, some proteins can withstand extreme conditions, known as extremophile proteins. The stability of this extremophile protein has attracted the interest of researchers to be used in industrial applications. Over the last decade, both traditional methods involve experimental and systematic comparison and computational methods such as machine learning are developed to classify extremophilic proteins and their counterparts and identify their essential properties as the basis for carrying out protein design and engineering. This review compares various studies to classify some extremophile proteins in terms of datasets, feature extractions, and classification algorithms used in each study. It also includes some challenges in this research area, such as data limitations, handcrafted features, limited sequence method utilized for classification, and model interpretability. The studies presented in this review are based on our unique review and by no means complete. Nevertheless, hopefully, the review will provide insight for new research in developing the more effective, interpretable and improved performance of predictions. © 2022 IEEE.

关键词： Proteins

来源：评论

学校读者我要写书评

暂无评论

Multicluster Kernel Intuitionistic Fuzzy C-Means and State Transition Algorithm : Framework for Low Light Segmentation Imaging of Tuberculosis Bacilli Base of Semisupervised Approach

Multicluster Kernel Intuitionistic Fuzzy C-Means and State T...

引用

IEEE International Conference on Fuzzy Systems (FUZZ-IEEE)

作者： Sari Ayu Wulandari I Ketut Eddy Purnama Eko Mulyanto Yuniarno Mauridhi Hery Purnomo Dept. of Electrical Engineering Faculty of Intelligent Electrical and Informatics Technology (F-Electics) Institut Teknologi Sepuluh Nopember Surabaya Indonesia Dept. of Biomedical Engineering Faculty of Electrical Engineering Universitas Dian Nuswantoro Semarang Indonesia Dept. of Electrical Engineering Dept. of Computer Engineering Faculty of Intelligent Electrical and Informatics Technology (F-Electics) Institut Teknologi Sepuluh Nopember Surabaya Indonesia

ISBN: (数字)9798350319545

ISBN: (纸本)9798350319552

Annotated images are required as ground truth for deep learning. Bacterial segmentation is requires a lot of time and effort when done manually. The autosegmentation task gets more challenging since bacterial images contain low light properties, which have an important effect on auto annotating tasks. In order to solve this issue, we present a system that includes a fuzzy-based clustering method that enhances bacterial object segmentation performance by utilizing the multicluster idea. The State Transform Algorithm (STA) is used to obtain starting centroids in order to increase stability, because the Kernel of Intuitionistic Fuzzy C- Means (KIFCM) is sensitive to starting centroids and hence sensitive to being stuck in local optima. The accuracy of KIFCM-STA with bicluster is poor in low-light images. To boost performance, the multicluster technique (MKIFCM-STA) is presented as a continuation hybrid of KIFCM-STA. This framework allows for the ideal amount of clusters (Silhouette) and cluster ranking to provide clusters containing bacterial objects (Topsis). In order to compare our method against four prior approaches (IFCM, KFCM, KIFCM, and KIFCM-STA), we compare its qualitative aspects (visualization of images) and quantitative aspects (average IOU, Dice, HD, ASD, and Accuracy). In low light image clustering tasks, our model significantly improves and achieves great results in terms of accuracy, with a score of 89.438%. This accomplishment highlights how crucial it is that our framework tackles the problem of low-light image clustering in images of bacteria, eventually improving the image auto-annotation procedure.

关键词： Deep learning Image segmentation Visualization Microorganisms Accuracy Clustering methods Clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

Offline Handwriting Signature Verification: A Transfer Learning and Feature Selection Approach

arXiv

引用

arXiv 2024年

作者： Özyurt, Fatih Majidpour, Jafar Rashid, Tarik A. Koç, Canan Department of Software Engineering Faculty of Engineering Firat University Elazig23119 Turkey Department of Software and Informatics Engineering University of Raparin Kurdistan Region Rania Iraq Department of Computer Science Faculty of Science Soran University Kurdistan Region Soran Iraq Computer Science and Engineering Department University of Kurdistan Hewlêr Erbil Iraq

Handwritten signature verification poses a formidable challenge in biometrics and document authenticity. The objective is to ascertain the authenticity of a provided handwritten signature, distinguishing between genuine and forged ones. This issue has many applications in sectors such as finance, legal documentation, and security. Currently, the field of computer vision and machine learning has made significant progress in the domain of handwritten signature verification. The outcomes, however, may be enhanced depending on the acquired findings, the structure of the datasets, and the used models. Four stages make up our suggested strategy. First, we collected a large dataset of 12600 images from 420 distinct individuals, and each individual has 30 signatures of a certain kind (All authors' signatures are genuine). In the subsequent stage, the best features from each image were extracted using a deep learning model named MobileNetV2. During the feature selection step, three selectors—neighborhood component analysis (NCA), Chi2, and mutual_info (MI)—were used to pull out 200, 300, 400, and 500 features, giving a total of 12 feature vectors. Finally, 12 results have been obtained by applying machine learning techniques such as SVM with kernels (rbf, poly, and linear), KNN, DT, Linear Discriminant Analysis, and Naïve Bayes. Without employing feature selection techniques, our suggested offline signature verification achieved a classification accuracy of 91.3%, whereas using the NCA feature selection approach with just 300 features it achieved a classification accuracy of 97.7%. High classification accuracy was achieved using the designed and suggested model, which also has the benefit of being a self-organized framework. Consequently, using the optimum minimally chosen features, the proposed method could identify the best model performance and result validation prediction vectors. © 2024, CC BY.

关键词： Feature Selection

来源：评论

学校读者我要写书评

暂无评论

Fine-Tuned MobileNetV2 and VGG16 Algorithm for Fish Image Classification 4

Fine-Tuned MobileNetV2 and VGG16 Algorithm for Fish Image Cl...

引用

4th International Conference on informatics, Multimedia, Cyber and Information System, ICIMCIS 2022

作者： Fitrianah, Devi Suryaningrum, Kristien Margi Sagala, Noviyanti Tri Maretta Ayumi, Vina Lim, Siew Mooi School of Computer Science Bina Nusantara University Computer Science Department Jakarta Indonesia School of Computer Science Bina Nusantara University Mathematics Statistics Department Jakarta Indonesia Universitas Mercu Buana Informatics Engineering Program Faculty of Computer Science Jakarta Indonesia Tunku Abdul Rahman University College Faculty of Computing and Information Technology Computer Science Department Kuala Lumpur Malaysia

ISBN: (纸本)9781665473279

Study on the identification and classification of fish is challenging and valuable because of its role in advancing the marine and agricultural fields. This research has benefits interms of monitoring fish populations and ecosystems in a particular area. Furthermore, this research helps monitor fish that are considered threatened or endangered so that it makes iteasier to map prohibited areas for fishing. This research aims to know performance of MobileNetV2 and VGG16 with parameter tuning process by identifying the value of batch size, epoch, learning rate, and optimizer for fish image dataset. The proposed research phase consists of five main stages, including experimental setup, dataset construction, dataset preprocessing, dataset training and modelling and evaluation. As the result, VGG16 obtained the highest accuracy value. For VGG16 without fine-tuning, the testing accuracy is 98.07%. For VGG16 with fine-tuning, the testing accuracy is 96.56%. © 2022 IEEE.

关键词： Fish

来源：评论

学校读者我要写书评

暂无评论

Web Application for Screening Malignant Melanoma in Digital Images Using Deep Learning

Web Application for Screening Malignant Melanoma in Digital ...

引用

International Conference on Big Data and Smart Computing (BIGCOMP)

作者： Kharittha Jangsamsi Kasidet Samosorn Puttimate Hirunuran Prapawin Sakdapetchsiri Wiriya Mahikul Rithee Smithrithee Krisana Chinnasarn Computer Engineering Department Faculty of Engineering King Mongkut's University of Technology Thonburi Bangkok Thailand Princess Srisavangavadhana college of medicine Chulabhorn Royal Academy Bangkok Thailand Faculty of Informatics Burapha University Chonburi Thailand

Malignant Melanoma (MM) is a highly lethal skin cancer and detecting it early is crucial, as it has a mortality rate of 48.08%. To address this issue, a web application was developed using deep learning models and image processing techniques. The application serves as a self-screening tool for individuals who are suspicious of any skin lesions. The study evaluated three deep learning models, namely AlexNet, MobileNet-V2, and ResNet-18, by using a comprehensive dataset from the MED-NODE dermatology database. The results showed that the ResNet-18 model, when combined with other approaches, demonstrated superior performance, with 86.11% accuracy, 88.10% precision, 88.10% sensitivity, and 83.33% specificity. The research highlights the potential of deep learning models and median filters in image processing techniques to facilitate effective melanoma risk screening, promote early detection, and improve patient outcomes.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Mathematical Modelling of Nonlinear Heat Conduction with Relaxing Boundary Conditions 1

引用

10th International Conference on Numerical Methods and Applications, NMA 2022

作者： Filipov, Stefan M. Faragó, István Avdzhieva, Ana Department of Computer Science Faculty of Chemical and System Engineering University of Chemical Technology and Metallurgy Sofia Bulgaria Department of Differential Equations Institute of Mathematics Budapest University of Technology and Economics & MTA-ELTE Research Group Budapest Hungary Faculty of Mathematics and Informatics "St. Kl. Ohridski" University of Sofia Sofia Bulgaria

ISBN: (数字)9783031324123

ISBN: (纸本)9783031324116

This paper considers one dimensional unsteady heat condition in a media with temperature dependent thermal conductivity. When the thermal conductivity depends on the temperature, the corresponding heat equation is nonlinear. At one or both boundaries, a relaxing boundary condition is applied. It is a time dependent condition that approaches continuously, as time increases, a certain time independent condition. Such behavior at the boundaries arises naturally in some physical systems. As an example, we propose a simple model system that can give rise to either Dirichlet or convective relaxing boundary condition. Due to the dependence of the thermal conductivity on the temperature, the convective condition is nonlinear. For the solution of the problem, we propose a new numerical approach that first discretizes the heat equation in time, whereby a sequence of two-point boundary value problems (TPBVPs) is obtained. We use implicit time discretization, which provides for unconditional stability of the method. If the initial condition is given, we can solve consecutively the TPBVPs and get approximations of the temperature at the different time levels. For the solution of the TPBVPs, we apply the finite difference method. The arising nonlinear systems are solved by Newton’s method. A number of example problems are solved demonstrating the efficiency of the proposed approach. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Boundary conditions

来源：评论

学校读者我要写书评

暂无评论

Revolutionizing Neurological Diagnostics: Integrating 6G Technology with Deep Learning for Enhanced Detection of Multiple Sclerosis and Myelitis

Revolutionizing Neurological Diagnostics: Integrating 6G Tec...

引用

Telecommunications Conference (ITC-Egypt), International

作者： Rana M. NourEldeen Rahma A. Elshshtawy Omnia A. Elgohary Fatma Alzahraa A. Wahba Marco K. Attia Zeinab B. Abo elazm Mohamed G. Khattap Hend Galal Eldeen Mohamed Ali Hassan Mariam Mostafa Nayra Ibrahim Reem Ramadan Yomna Refaat Mennatallah Khaled Ali Maher Mohamed Abd Elaziz Ahmed Gamal Abdellatif Ibrahim Radiology and Medical Imaging Faculty of Applied Health Sciences Galala University Suez Egypt Biomedical Informatics Faculty of Computer Science and Engineering Galala University Suez Egypt Artificial Intelligence Science Program Faculty of Computer Science and Engineering Galala University Suez Egypt Department of Electronics and Electrical Communications Engineering Air Defense College Alexandria Egypt

ISBN: (数字)9798350351408

ISBN: (纸本)9798350351415

This paper proposes an alternative detection frame-work for multiple sclerosis (MS) and idiopathic acute transverse myelitis (ATM) within the 6G-enabled Internet of Medical Things (IoMT) environment. The developed framework relies on the implementation of a deep learning technique known as Dense Convolutional Networks (DenseNets) in the 6G-enabled IoMT to enhance prediction performance. To validate the performance of DenseNets, we compared it with other deep learning techniques, including Convolutional Neural Networks (CNN) and MobileNet, using real-world datasets. The experimental results show the high performance of DenseNets in predicting MS and ATM compared to other methods, achieving an accuracy of nearly 90 %.

关键词： 6G mobile communication Deep learning Multiple sclerosis Accuracy Internet of Medical Things Telecommunications Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Grid Search Hyperparameters Tuning with Supervised Machine Learning for Awngi Language Named Entity Recognition 2

Grid Search Hyperparameters Tuning with Supervised Machine L...

引用

2nd International Conference Computational and Characterization Techniques in engineering and Sciences, IC3TES 2024

作者： Ayalew, Amogne Andualem Bitew, Gizachew Salau, Ayodeji Olalekan Tegegne, Melaku Lake Gupta, Shashi Kant Israr, Mohammad Tegegne, Tesfa Assegie, Tsehay Admassu Mizan-Tepi University School of Computing and Informatics Department of Information Technology Tepi Ethiopia Mizan Tepi University School of Computing Computer Science Department Mizan-Tepi Ethiopia Afe Babalola University Department of Electrical/Electronics and Computer Engineering Ado-Ekiti Nigeria Injibara University College of Engineering and Technology Department of Information Technology Injibara Ethiopia Eudoxia Research University Department of Computer Science and Engineering United States Maryam Abacha American University of Nigeria Hotoro Gra Kano State Nigeria Bahir Dar University Faculty of Computing Bahir Dar Institute of Technology Bahir Dara Ethiopia Injibara University College of Engineering and Technology Department of Computer Science Injibara Ethiopia

ISBN: (纸本)9798350364699

Named entity recognition has emerged as a critical step in recognizing, classifying, and extracting the most significant information from unstructured text without human intervention. It is used in information retrieval, conversation systems, machine translation, data mining, and information extraction. However, the Awngi language lacks a NER system due to inadequate datasets. This study combines feature extraction techniques with supervised machine learning algorithms, including support vector machines, Naïve Bayes, and conditional random fields with a grid search algorithm parameter optimization. These supervised machine-learning models often have various hyperparameters that must be tuned to optimize the model's performance by automatically selecting the parameter. Furthermore, the dataset used for training and testing consists of 20,193 annotated Awngi tokens, collected from Amhara Media Corporation, Awngi Telegram, and Facebook pages. The proposed SVM with grid search algorithm approach provided a promising result in ANER with an accuracy of 92% compared with Naïve Bayes, conditional random field, and decision-tree supervised machine learning models. Finally, the experiment results show that the proposed parameter tuning-based model can detect and categorize Awingi-named entities and outperforms the performance accuracy without parameter adjustment. © 2024 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Spatio-temporal Network Analysis for Detecting Traffic Accident Clusters 14

Spatio-temporal Network Analysis for Detecting Traffic Accid...

引用

14th IIAI International Congress on Advanced Applied informatics, IIAI-AAI 2023

作者： Ando, Keisuke Kuniyoshi, Yusuke Shimizu, Kazuhiko Uchitane, Takeshi Mukai, Naoto Iwata, Kazunori Ito, Nobuhiro Jiang, Yong Graduate School of Business Administration and Computer Science Aichi Institute of Technology Toyota Japan Graduate School of Informatics and Engineering The University of Electro-Communications Tokyo Japan Faculty of Information Science Aichi Institute of Technology Toyota Japan Sugiyama Jogakuen University Faculty of Culture-Information Studies Nagoya Japan Aichi University Faculty of Business Administration Nagoya Japan Aichi University Faculty of Regional Policy Toyohashi Japan

ISBN: (纸本)9798350324228

Various methods have been proposed for analyzing traffic accident hotspots. One of these methods is to detect traffic accident hotspots on the road network using a hypothesis testing method. However, this method does not consider the time of occurrence of a traffic accident. In other words, this method provides information on high-risk locations for traffic accidents but not on the corresponding time. This paper proposes a new method for detecting traffic accident hotspots considering not only location but also time. We therefore extended the previous hypothesis testing method to consider the time of occurrence of a traffic accident. First, we check for changes in spatial properties over time in the target area using local indicators of spatial autocorrelation (LISA) cluster maps. Next, we estimate the probability density function of traffic accidents in the spatio-temporal network using the spatio-temporal network kernel density estimation (STNKDE) method. Finally, we detect clusters where the probability density of traffic accidents is significantly higher through hypothesis testing based on the estimation results. We detect clusters under four conditions while testing different hypotheses and parameters. The results show that the proposed method can detect traffic accident clusters. Therefore, when considering traffic safety measures, the method can be adopted to detect the location and time that require attention and allocate the necessary resources. © 2023 IEEE.

关键词： Roads and streets

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：