检索结果-内蒙古大学图书馆

Informatica (Slovenia) 2025年第17期49卷 173-188页

作者： Mezzoudj, Saliha Khelifa, Meriem Saadna, Yasmina Department of Computer Science University of Algiers Algiers Algeria Artificial Intelligence of Information Technologies Department of Computer Science and Information Technologies University of Kasdi Merbah Ouargla Algeria Labstic laboratory Batna 2 University Batna Algeria

The COVID-19 epidemic has been a critical global challenge due to its high mortality rate and rapid spread. Initial diagnostic methods, such as chest X-rays and reverse transcriptase polymerase chain reaction (RT-PCR), can be time-consuming and require enhanced efficiency, especially in scenarios where large-scale testing is necessary. Combining deep learning (DL) and big data analytics tools is crucial for developing accurate and fast systems to detect COVID-19 infections and prevent further transmission. In this study, we introduce a distributed Deep Convolutional Neural Network (DCNN) framework designed for the detection of COVID-19, utilizing a comprehensive dataset of chest X-ray images. Our approach leverages the spark-tensorflow-distributor, which integrates TensorFlow 2.x with Apache Spark 3.x, enabling distributed training within the Spark framework. The system utilizes Spark’s barrier mode to execute distributed training tasks, significantly enhancing the speed of both training and testing phases. We classify chest X-ray images into two categories: COVID-19 and non-COVID-19. The experimental setup involves processing the large-scale dataset of chest X-ray images in a parallelized manner using the Spark framework. Our results demonstrate that the proposed DCNN model achieves a classification accuracy of 95.53%, with a testing time of 0.06 seconds and a training time of 18,909 seconds. Furthermore, when compared to other state-of-the-art methods for COVID-19 classification, our approach proves to be more efficient and effective. This study highlights the potential of combining distributed computing frameworks with deep learning techniques to address largescale medical image analysis challenges, particularly in the context of COVID-19 detection. © 2025 Slovene Society Informatika. All rights reserved.

关键词： COVID-19

来源：评论

学校读者我要写书评

暂无评论

SSVEP-DAN: Cross-Domain Data Alignment for SSVEP-Based Brain-computer Interfaces

引用

IEEE Transactions on Neural Systems and Rehabilitation Engineering 2024年 32卷 2027-2037页

作者： Chen, Sung-Yu Chang, Chi-Min Chiang, Kuan-Jung Wei, Chun-Shu Department of Computer Science Hsinchu30010 Taiwan Arctop Inc. La Jolla CA92093 United States Department of Computer Science The Institute of Education The Institute of Biomedical Engineering Hsinchu30010 Taiwan

Steady-state visual-evoked potential (SSVEP)-based brain-computer interfaces (BCIs) offer a non-invasive means of communication through high-speed speller systems. However, their efficiency is highly dependent on individual training data acquired during time-consuming calibration sessions. To address the challenge of data insufficiency in SSVEP-based BCIs, we introduce SSVEP-DAN, the first dedicated neural network model designed to align SSVEP data across different domains, encompassing various sessions, subjects, or devices. Our experimental results demonstrate the ability of SSVEP-DAN to transform existing source SSVEP data into supplementary calibration data. This results in a significant improvement in SSVEP decoding accuracy while reducing the calibration time. We envision SSVEP-DAN playing a crucial role in future applications of high-performance SSVEP-based BCIs. The source code for this work is available at: https://***/CECNL/SSVEP-DAN. © 2001-2011 IEEE.

关键词： Brain computer interface

来源：评论

学校读者我要写书评

暂无评论

Identifying Abnormal Eating Behavior Patterns with Machine Learning for Early Detection of Eating Disorders: A Systematic Literature Review 9

Identifying Abnormal Eating Behavior Patterns with Machine L...

引用

9th International Conference on Information Management and technology, ICIMTech 2024

作者： Santoso, Kezia Angeline Maharani, Mawar Sidharta, Sidharta School of Computer Science Bina Nusantara Univeristy Computer Science Department Jakarta11480 Indonesia

ISBN: (纸本)9798350390025

In the field of health, there are common diseases that can have adverse effects on the body and last a lifetime, one of which is eating disorders (ED), which is a mental illness with features of varying degrees related to eating behavior and body image. As such, prediction and early detection of eating disorders are critical to a patient's recovery. To date, machine learning has been implemented to predict and identify eating disorders using a variety of different models and datasets such as interviews/surveys, social media posts and observations. This paper is written according to the PRISMA 2020 guidelines, and aims to answer four research questions, which are Whether ML is an effective way to detect eating disorders, What scales and data types are utilized in the ML methods used to analyze the datasets, What factors should be considered in determining the ML method used, and Between social media and interview datasets, which dataset give more accuracy in predicting eating disorders. There are 30 papers that are material for analysis in the literature review. The results of the analysis indicate that the methods used by the researchers are random forest, support vector machine and logistic regression. In addition, the accuracy achieved by machine learning in predicting eating disorders is very promising and effective. This paper also concluded that ML models used on social media and interview/survey datasets have an almost identical accuracy © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Bilingual Approach: Leveraging Deep Neural Network Techniques for Handwritten Signature Authentication

Bilingual Approach: Leveraging Deep Neural Network Technique...

引用

2024 IEEE International Conference on Information technology, Electronics and Intelligent Communication Systems, ICITEICS 2024

作者： Mohanraj, P. Janit, R. Sophia Hindustan Institute of Technology and Science Department of Computer Applications Chennai India

ISBN: (纸本)9798350382693

In an era defined by technological innovation, the importance of robust authentication mechanisms cannot be overstated, particularly in the realm of security management. Handwritten signature recognition stands as a pivotal component in identity verification processes. Our system harnesses the power of Deep Neural Networks (DNNs) to extract intricate hierarchical features from signature images, effectively capturing both global and local patterns with remarkable precision. To assess and confirm the effectiveness of our algorithm in distinguishing genuine signatures from fraudulent ones, we carefully compile datasets containing both new and existing signatures. Additionally, our work introduces an offline human signature system, underpinned by the sophisticated Histogram Equalization method, further augmenting the robustness of our approach. Our approach, which is based on research utilised in the educational system, is very helpful in protecting and validating academic documents. The addition of bilingual features increases the usefulness of our system by supporting a variety of linguistic signatures that are frequently seen in educational materials. This method guarantees a thorough and flexible solution for handwritten signature authentication in security and educational institutions. © 2024 IEEE.

关键词： Authentication Protocol

来源：评论

学校读者我要写书评

暂无评论

Predicting Age and Gender Across Different Races Using Convolutional Neural Networks: A Deep Learning Approach

Predicting Age and Gender Across Different Races Using Convo...

引用

2024 International Conference on Information technology and Computing, ICITCOM 2024

作者： Ardelia, Elvina Thenando, Jericho Gunawan, Alexander A.S. Syahputra, Muhammad E. School of Computer Science Bina Nusantara University Computer Science Department Jakarta11480 Indonesia

ISBN: (纸本)9798350379839

Age and gender are two important attributes possessed by humans that serve as their identity in social life. Faking age and gender are criminal acts that are harmful especially for bad purposes. Therefore, artificial intelligence systems are needed which can predict the age and gender of humans with high accuracy. The model commonly used to detect images and perform visual-based tasks is CNN. Nowadays, there are many frameworks and other models that are the development of CNN models, one of them is DeepFace. The purpose of this research is to evaluate the CNN model and compare its accuracy with DeepFace using Mean Absolute Error metric for age and Exact Match Ratio metric for gender. The test data used is divided into 5 races, namely Asian, Black, Indian, White, and other races (like Middle Eastern, Latino, and Hispanic). The result of this research is the accuracy value for each race using CNN and DeepFace which states that DeepFace is superior for age prediction and CNN is superior for gender prediction. © 2024 IEEE.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Adversarial Neural Collaborative Filtering with Embedding Dimension Correlations

引用

Data Intelligence 2023年第3期5卷 786-806页

作者： Yi Gao Jianxia Chen Liang Xiao Hongyang Wang Liwei Pan Xuan Wen Zhiwei Ye Xinyun Wu Hubei University of Technology School of Computer ScienceWuhan 430068China

Recently,convolutional neural networks(CNNs)have achieved excellent performance for the recommendation system by extracting deep features and building collaborative filtering ***,CNNs have been verified susceptible to adversarial *** is because adversarial samples are subtle non-random disturbances,which indicates that machine learning models produce incorrect ***,we propose a novel model of Adversarial Neural Collaborative Filtering with Embedding Dimension Correlations,named ANCF in short,to address the adversarial problem of CNN-based recommendation *** particular,the proposed ANCF model adopts the matrix factorization to train the adversarial personalized ranking in the prediction *** is because matrix factorization supposes that the linear interaction of the latent factors,which are captured between the user and the item,can describe the observable feedback,thus the proposed ANCF model can learn more complicated representation of their latent factors to improve the performance of *** addition,the ANCF model utilizes the outer product instead of the inner product or concatenation to learn explicitly pairwise embedding dimensional correlations and obtain the interaction map from which CNNs can utilize its strengths to learn high-order *** a result,the proposed ANCF model can improve the robustness performance by the adversarial personalized ranking,and obtain more information by encoding correlations between different embedding *** results carried out on three public datasets demonstrate that the ANCF model outperforms other existing recommendation models.

关键词： Neural Collaborative Filtering Matrix Factorization Convolutional Neural Networks Adversarial Training Recommendation systems

来源：评论

学校读者我要写书评

暂无评论

Analyzing Views on Presidential Candidates for Election 2024 Based on the Instagram and X Platforms with Text Clustering 9th

Analyzing Views on Presidential Candidates for Election 2024...

引用

9th International Conference on computer science and Computational Intelligence, ICCSCI 2024

作者： Hanifa, Aida Debora, Catheline Hasani, Muhammad Fikri Wicaksono, Pandu Computer Science Department School of Computer Science Bina Nusantara University Jakarta11480 Indonesia

Due to technological advancements, social media has become a major pawn in building people's opinions and actions, especially ahead of important moments such as the 2024 general election in Indonesia. Having a good understanding of the headlines in social media will have a big impact on the people and candidates in terms of votes. The purpose of this research is to group sentences in clusters and find keywords that can be used as parameters to determine topics in each cluster. Data sources will be taken from Instagram and X And will be vectorized with TF-IDF, SBERT, and clustered with the DBSCAN method. Then, complemented by LDA for result exposure. The results obtained in this research found that the best parameters of DBSCAN with TF-IDF are 0.85 epsilon, 2 minPts, 0.594768 score, 2 number of clusters and 20 outliers. The best parameters of DBSCAN with Sentence BERT are 0.1 epsilon, 2 minPts, -0.293426 score, 4 number of clusters and 937 outliers. The results of DBSCAN with each text embedding from topic modeling with LDA show that it is related to the 2024 presidential candidates such as Anies, Prabowo, gibran, etc. From all topics displayed by LDA, the most frequently discussed presidential candidate pair is pair number 01, namely Anies Baswedan and Muhaimin Iskandar with several words related to their traits or actions. Based on the dataset, the topics formed refer to certain sentiments by the community with predominantly positive or supportive responses such as discussing traits that are suitable as role models or ideas that are considered modern (digital approach) displayed at DBSCAN with Sentence-BERT method. Where most of these responses are sourced from the social media platform Instagram. These indices that most Indonesians view pair 01 as ideal presidential candidate both in traits and actions as a presidential candidate, especially the public opinion from Instagram. © 2024 The Authors.

关键词： Statistics

来源：评论

学校读者我要写书评

暂无评论

Recognitions of Bahasa Isyarat Indonesia (BISINDO) Alphabets Using SVM and Mediapipe 6

Recognitions of Bahasa Isyarat Indonesia (BISINDO) Alphabets...

引用

6th International Conference on Cybernetics and Intelligent System, ICORIS 2024

作者： Wiliam, Robert Lufian, Charlie Meiliana Zakiyyah, Alfi Yusrotis Computer Science Department School of Computer Science Bina Nusantara University Jakarta11480 Indonesia

ISBN: (纸本)9798331539603

Bahasa Isyarat Indonesia (BISINDO), Indonesia's national sign language, serves as a primary mode of communication for a large portion of the population. This study explores a novel approach to BISINDO sign language recognition by leveraging Support Vector Machine (SVM) in conjunction with MediaPipe, a framework for real-time machine learning. MediaPipe has demonstrated strong performance on other datasets, and in this application, it extracts features from hand gesture images that are then classified by the SVM algorithm. Our model achieved an impressive average accuracy score of 98.82%, with an average precision of 97.34%, recall of 99.00%, and F1-score of 97.76%. In comparison to previous studies using HMM (which achieved an accuracy of 76.25%) and CNN (with an accuracy of 75.83%), our model demonstrates a significant improvement in recognizing BISINDO. Our model offers several distinct advantages. MediaPipe’s robust feature extraction capabilities are particularly well-suited to capturing complex hand movements, allowing SVM to focus on high-quality input features that enhance classification accuracy. Unlike deep learning models such as CNN, which require substantial computational resources, our model remains lightweight and efficient, making it ideal for deployment on mobile and low-resource devices. This efficiency ensures that BISINDO recognition can be achieved in real-time and in practical applications, enabling a more accessible and reliable communication tool for the BISINDO-used community. © 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

CryptStego: Powerful Blend of Cryptography and Steganography for Securing Communications

CryptStego: Powerful Blend of Cryptography and Steganography...

引用

International Conference on Emerging Trends and Applications in Artificial Intelligence, ICETAI 2023

作者： Pandey, Shraiyash Baniya, Pashupati Nand, Parma Hameed, Alaa Ali Bhushan, Bharat Jamil, Akhtar Department of Computer Science and Engineering School of Engineering and Technology Sharda University Greater Noida India Department of Computer Engineering Istinye University Istanbul Turkey Department of Computer Science National University of Computer and Emerging Sciences Islamabad Pakistan

ISBN: (纸本)9783031567278

In today’s era, security is one of the most critical issues in the development of electronic communications applications, especially when sending private data. The data may be encrypted with several algorithms;however, an extra layer of security can improve protection by a significant amount. Therefore, in this paper, we have developed an application, CryptStego, to secure data using two techniques, cryptography and steganography, to transmit data securely. The encryption of original data is executed using Blowfish algorithm, a cryptographic technique. Additionally, the encrypted data is hidden through using Least Significant Bit (LSB), a steganography technique. The implementation of both techniques offers an extensive level of security, since an intruder must firstly identify the encrypted text within the image to attain the encrypted text, then secondly to decrypt using the algorithm to obtain the original message. Therefore, any intruder must encounter multiple levels of security to obtain the original message from the cipher image. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Steganography

来源：评论

学校读者我要写书评

暂无评论

CoverGAN: cover photo generation from text story using layout guided GAN

引用

Soft Computing 2025年第1期29卷 405-423页

作者： Cheema, Adeel Naeem, M. Asif Department of Data Science and Artificial Intelligence National University of Computer and Emerging Sciences Islamabad Pakistan

Generating cover photos from story text is a non trivial challenge to solve. Existing approaches focus on generating only images from given text prompt. To the best of our knowledge, non of these approaches focus on generating cover photos from a text story. The paper addresses this issue by introducing multi-object image generation with text title from a text story. We split the problem into three steps:- understanding semantics of text story, predicting layout of objects, and generating a cover photo. At start, a semantic relation was encoded between text story objects using a Scene graph, then features from graph neural network were concatenated with single object features from scene graph to create an object layout. All of these features were then passed on to the image generating part. Image generation was further divided into two phases. In the first phase, the image is generated using a scene graph image generation model. While in the second phase, the results of first phase were further enhanced using image translation model conditioned on the object layout. In final phase we generated title of the given story based on generated image. In our experiments, we used custom dataset of text stories with three animal categories along with the COCO dataset. For the image generating part, we evaluated our approach with state of the art models known as scene_gen and sg2im. Our method generated high-resolution informative cover photo with story title by positioning the objects at right locations as specified in the text story. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： Graph neural networks

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：