检索结果-内蒙古大学图书馆

Target Detection Algorithm in Foggy Scenes Based on Dual Subnets

computers, Materials & Continua 2024年第2期78卷 1915-1931页

作者： Yuecheng Yu Liming Cai Anqi Ning Jinlong Shi Xudong Chen Shixin Huang School of Computer Science Jiangsu University of Science and TechnologyZhenjiang212100China Information Department China Merchants Heavy IndustryNantong226116China

Under the influence of air humidity,dust,aerosols,etc.,in real scenes,haze presents an uneven *** this way,the image quality and contrast will *** this case,It is difficult to detect the target in the image by the universal detection ***,a dual subnet based on multi-task collaborative training(DSMCT)is proposed in this ***,in the training phase,the Gated Context Aggregation Network(GCANet)is used as the supervisory network of YOLOX to promote the extraction of clean information in foggy *** the test phase,only the YOLOX branch needs to be activated to ensure the detection speed of the ***,the deformable convolution module is used to improve GCANet to enhance the model’s ability to capture details of non-homogeneous ***,the Coordinate Attention mechanism is introduced into the Vision Transformer and the backbone network of YOLOX is *** this way,the feature extraction ability of the network for deep-level information can be *** experimental results on artificial fog data set FOG_VOC and real fog data set RTTS show that the map value of DSMCT reached 86.56%and 62.39%,respectively,which was 2.27%and 4.41%higher than the current most advanced detection *** DSMCT network has high practicality and effectiveness for target detection in real foggy scenes.

关键词： Target detection fog target detection YOLOX twin network multi-task learning

来源：评论

学校读者我要写书评

暂无评论

MediaPipe’s Pose-Based Human Activity Recognition with LSTM 6

MediaPipe’s Pose-Based Human Activity Recognition with LSTM

引用

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

作者： Muliawan, Nicholas Hans Irmawan, Frederick Nathan Angky, Edbert Valencio Prabowo, Abram Setyo Computer Science Department School of Computer Science Bina Nusantara University Jakarta Indonesia

ISBN: (纸本)9798331539603

It is difficult to comprehend and categorize human movements in Human Activity Recognition. Despite being widely utilized, convolutional neural networks (CNN) might not adequately represent the intricacy of motions. This work investigates an alternative method for classifying human motion: the LSTM model combined with MediaPipe. Sequential processing is used by the LSTM model for keypoint sequences, which improves its accuracy in identifying complex human movement features. Although OpenPose provides precise body mapping, MediaPipe is used in this research because of its computational efficiency and ease of integration. Using CV2 libraries, preprocessed video data is read at 15 frames per second. MediaPipe software is then used to extract 33 keypoints coordinates from each frame. Then, in order to make these keypoints compatible with the LSTM model, they are compressed into single-dimensional arrays. The keypoints from each picture frame are taken out and combined with other frames for a total of 15 frames into a single sequence. The aim of this study is to evaluate the usefulness of the LSTM model in human action classification by contrasting its performance with more conventional sequence modeling methods. Following evaluation, we employed the Recognition of Human Action dataset from Kungliga Tekniska Hogskolan (KTH) University, which is used by LSTM models. Over 99% of the metricsAccuracy, Precision, Recall, and F1-scorewere attained by the model. © 2024 IEEE.

关键词： computer vision

来源：评论

学校读者我要写书评

暂无评论

Improving DNS Server Resilience Against DDoS Attacks Through Anycast Routing 6

Improving DNS Server Resilience Against DDoS Attacks Through...

引用

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

作者： Dzaky, Julian Daffa Saputra, Jason Adrianto Singale, Nicodemus Nathanael Saputri, Hanis Amalia Sari, Azani Cempaka Computer Science Department School of Computer Science Bina Nusantara University Jakarta Indonesia

ISBN: (纸本)9798331539603

Domain Name System (DNS) is the critical backbone of the internet, translating human-readable domain names into machine-readable IP addresses. DDoS attacks are a type of distributed denial-of-service (DDoS) attacks where the attacker attempts to make a victim's resources unavailable by flooding the server with bots that have high bandwidth which results in a slow server or outright a server shutdown. Anycast routing is a crucial technology in internet services such as DNS servers. This research paper aims to test the features of anycast routing in DNS to evaluate its resilience against DDoS attacks. A quantitative and Black-box technique approach is used, using experiments that simulate DDoS attacks on DNS servers in various scenarios. data was collected from nine popular websites in Indonesia that use DNS servers with and without Anycast routing. The results showed that Anycast routing, especially with Cloudflare's WARP feature, improved DNS resilience against DDoS attacks by maintaining service availability with an average response time of 172.44 ms for before the attack and 117.66 ms after the DDoS attack. However, this experiment also found the need to use additional protections, such as WARP, to effectively manage and mitigate DDoS attacks. This research also emphasizes the importance of combining Anycast routing with additional protection to improve the resilience of DNS servers against external threats such as DDoS attacks. © 2024 IEEE.

关键词： Denial-of-service attack

来源：评论

学校读者我要写书评

暂无评论

Web-Based Application for GUI programming Learning Through Reverse Engineering Method

Web-Based Application for GUI Programming Learning Through R...

引用

2024 International Conference on Information Technology Systems and Innovation, ICITSI 2024

作者： Tunardi, Yovita Yulianto, Budi Layona, Rita Bina Nusantara University School of Computer Science Computer Science Department Jakarta Indonesia

ISBN: (纸本)9798331511470

GUI (Graphical User Interface) programming is the process of developing (coding) a graphical interface application. GUI-based application, which replaces TUI (Textual User Interface), allows users to interact with a program interactively. computer science students are required to master in creating GUI-based applications because of industry needs. Despite the importance of mastering GUI programming, students still have obstacles when learning it. Students who fail the GUI programming course will potentially experience difficulties in two things. First, it is difficult to finish the final project. Second, it is difficult to get a programmer job after graduation. This research aims to develop a web-based application for GUI programming learning using a reverse engineering method that can produce sourcecode based on GUI designed by users. The application development method used is the Waterfall Method with the stages of Requirement Collection, Analysis & Design, Development & Testing, Implementation, and Maintenance. The conclusion of this research is that the application helps users understand GUI programming learning through the reverse engineering method. © 2024 IEEE.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

Exploring the Ethical Landscape: Safeguarding data Privacy in the Realm of Virtual Reality

Exploring the Ethical Landscape: Safeguarding Data Privacy i...

引用

2024 International Conference on Informatics Electrical and Electronics, ICIEE 2024

作者： Tionanda, Leandro Satyawira Sin, Bun Jak Achmad, Said Sutoyo, Rhio School of Computer Science Bina Nusantara University Computer Science Department Jakarta Indonesia

ISBN: (纸本)9798331519216

As VR becomes more immersive and widespread, the fact that it collects vast amounts of personal and sensitive data from users, have raised concerns about how this information is managed and protected. This paper delves into the complex landscape of VR, focusing on the collection, management, and protection of user data. Through a systematic literature review using the PRISMA framework filtered by inclusion and exclusion criteria which are based on our research questions, this research identifies key privacy concerns, evaluates current security measures, and examines the ethical implications of VR usage. Our findings reveal several critical issues: the potential for misuse of personal data, insufficient transparency in data handling practices, and the lack of robust security protocols to safeguard user information. We also explore the ethical dimensions of VR, such as the potential for manipulation of user behavior and the responsibilities of VR companies to protect user welfare. To address these challenges, we recommend the development of comprehensive privacy policies, implementation of more advanced security measures, and establishment of clear ethical guidelines. Additionally, it is essential for VR companies to prioritize privacy and ethic in their products. By prioritizing these aspects, VR industry can create a safer and more trustworthy virtual environment for users which would in turn ensure a secure and ethical future for this technology. This paper aims to provide insights and practical recommendations for policymakers, developers, and researchers, encouraging a collaborative effort to enhance data privacy and ethical standards in the rapidly evolving field of Virtual Reality. © 2024 IEEE.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of KNN and SVM Algorithms for Diabetes Prediction 6

Comparative Analysis of KNN and SVM Algorithms for Diabetes ...

引用

6th International Conference on Informatics, Multimedia, Cyber and Information System, ICIMCIS 2024

作者： Oktovianus, Louis Linardi, Alvin Makalew, Brilly Andro Wairooy, Irma Kartika School of Computer Science Bina Nusantara University Computer Science department Tangerang Indonesia

ISBN: (纸本)9798331533243

Diabetes Mellitus is a chronic, multifactorial syndrome related to defective insulin secretion or its action. This can be considered a disease of all people in the world;hence, early detection of this disease will provide a better outcome in terms of prevention and improvement of patients. This also follows the comparative analysis between K-Nearest Neighbors and the Support Vector Machine on the Pima Indians Diabetes dataset, consisting of 768 instances with eight quantitative attributes along with a binary target attribute. Cleaning the data, handling the class imbalance by using SMOTE, and scaling the features are the preprocessing steps involved in the exploration of data. After the training and evaluation of the model, it was found that 80% of the data was used for training, and the remaining 20% were used for testing. From these, the results proved very dominantly that SVM holds a very strong position over KNN in nearly all performance metrics. Here, the accuracy of SVM is about 89.45%, while that of KNN is approximately 78.57%. The confusion matrix graph also showed very minimum false positives and false negatives, hence developed more accurate predictions about diabetes. To this end, it would be desirable that future research investigate the inclusion of ensemble techniques for more advanced and heuristic diagnostic tools in healthcare that may further assist in real-time and cost-effective solutions for improved global diabetes management or good health outcomes. © 2024 IEEE.

关键词： Diabetes Prediction Diagnostic Accuracy kNearest Neighbors Machine Learning Support Vector Machine

来源：评论

学校读者我要写书评

暂无评论

Improving Robustness of Time-Aware Shaper in Time-Sensitive Networking

引用

IEEE Internet of Things Journal 2025年第13期12卷 25211-25223页

作者： Ma, Yanying Zhang, Tong Feng, XiaoQin Liu, Ping Yang, Hao Ren, Fengyuan Lanzhou University School of Information Science and Engineering Lanzhou730000 China Nanjing University of Aeronautics and Astronautics College of Computer Science and Technology Nanjing211106 China School of Information Science and Engineering Lanzhou China Tsinghua University Department of Computer Science and Technology Lanzhou Beijing730000 and 100084 China

Deterministic delivery of scheduled traffic (ST) is critical in time-sensitive networking (TSN). The time-aware shaper (TAS) defined by IEEE 802.1Qbv is the enabler to ensure deterministic end-to-end delays of ST flows. However, TAS does not consider the emergency sporadic flows that commonly exist in automotive and industrial control applications. Therefore, TAS cannot resist the interference of emergency sporadic traffic on normal scheduled traffic. In this paper, we improve the TAS’s time slot allocation and queue occupancy rule and propose a combined scheduling strategy composed of dynamic local regulation and static global planning. Dynamic local regulation executes the earliest deadline first (EDF) discipline at switching nodes to adjust ST frames’ transmission dynamically. Static global planning calculates the offset of ST flows at the source and guarantees bounded delay and jitter. Furthermore, we formalize the offset optimizing problem with considering the EDF, delay, and jitter constraints. Finally, we implement our prototype on OMNet++ 6.0.1. Simulation results further demonstrate that the combined strategy has bounded latency and jitter in the presence of emergency sporadic flows. Besides, the comparison experiments show that our design is superior to eTAS in terms of the jitter and end-to-end delay. © 2014 IEEE.

关键词： Timing jitter

来源：评论

学校读者我要写书评

暂无评论

Comparison of Different Machine Learning Algorithms for Predicting Loan Risk Categories

Comparison of Different Machine Learning Algorithms for Pred...

引用

2023 International Conference on computer science, Information Technology and Engineering, ICCoSITE 2023

作者： Sudjono, Salomo Hendrian Adrian, Favian H.N. Sunarya, Caroline Angelina Ariyanto, Gabrielle Felicia Sagala, Noviyanti T M Bina Nusantara University Data Science Program School of Computer Science Jakarta11480 Indonesia School of Computer Science Bina Nusantara University Statistics Department Jakarta11480 Indonesia

ISBN: (纸本)9798350320954

There are many cases where borrowed money by debtors is not returned. It is because the company misjudged in determining the risk of lending. Thus, debtors cannot repay their debts and end up in losses on the company's side. Thus, this research aims to create an accurate classification model for determining loan risk categories to minimize business losses. We used the public dataset from kaggle website. Then we perform several data transformation and analysis methods are performed to obtain information from the data. We built 2 classification models to predict the original label. Unfortunately, the desired results did not achieve based on the f1-score. Therefore, the clustering method using K-Means is applied to make a new label for the dataset and determine the behavior of each cluster. Each cluster is labeled then build a model to perform cluster classification. We created several classification models and found that gradient tree boosting is the most accurate predictor of the loan risk category based on the accuracy. We used clustering to create a classification model that can have high accuracy. The classification model we built can have high accuracy because it uses clustering. The empirical results of this study and others pave the way for further in-depth future work to guarantee that the algorithm performs acceptably when confronted with more complicated real data. © 2023 IEEE.

关键词： K-means clustering

来源：评论

学校读者我要写书评

暂无评论

Unlocking the Potential of IndoBERT for Classification of Indonesian Thesis 6

Unlocking the Potential of IndoBERT for Classification of In...

引用

6th International Conference on Informatics, Multimedia, Cyber and Information System, ICIMCIS 2024

作者： Kurniawan, Marcel Juarto, Budi School of Computer Science Bina Nusantara University Computer Science Department Jakarta Indonesia

ISBN: (纸本)9798331533243

Nowadays, the development of digital information volume is growing rapidly, thus creating new challenges in data processing in educational institutions. One of the data that will always increase every year is thesis, a scientific work prepared by final year students as a graduation requirement. The purpose of this study is to explore the use of IndoBERT in classifying Indonesian thesis which has never been done before. This study also compares the performance of IndoBERT with traditional machine learning models such as XGBoost, LightGBM, and Gradient Boosting Machine (GBM). The stages of this research include data collection followed by text preprocessing, text representation, model implementation, and result evaluation. The results of this study show IndoBERT's high performance in classifying thesis documents by almost classifying all documents with 96% accuracy which proves that IndoBERT can classify Indonesian thesis documents accurately. The high performance of IndoBERT in classifying Indonesian theses is because IndoBERT is a pre-trained model that has been trained from various Indonesian language sources such as news and the internet. © 2024 IEEE.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

Performance Analysis of Regression Models on Predicting Movie Ratings 2

Performance Analysis of Regression Models on Predicting Movi...

引用

2nd International Conference on Technology Innovation and Its Applications, ICTIIA 2024

作者： Kennard, Ignatius Hartono, Hansel Lucky, Henry Suhartono, Derwin School of Computer Science Bina Nusantara University Computer Science Department Jakarta Indonesia

ISBN: (纸本)9798350351613

This research explores the application and performance comparison of various machine learning for movie ratings. This study aims to make a comprehensive comparison in performance between various machine learning models such as XGBoost, Random Forest Regressor, K Nearest Neighbor Regression, Support Vector Regressor and Linear Regression. The dataset used is MovieLens 20M which includes a rating dataset and a movie dataset. We employ a content-based recommendation system in conjunction to applying machine learning models to predict movie ratings based on user meta data. Our methodology involves splitting the data into 70:30 train-test data and evaluating them using Mean Absolute Error (MAE), Mean Square Error (MSE), and Root Mean Squared Error (RMSE). The findings of our research conclude that XGBoost is the superior model which is shown from the RMSE value of XGBoost being the lowest (0.8727). © 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

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：