检索结果-内蒙古大学图书馆

A fine-tuned vision transformer-based on limited dataset for facial expression recognition

Procedia computer science 2024年 245卷 574-582页

作者： Rio Febrian Ronald Richie Huang Nicholas Setiono Dimas Ramdhan Andry Chowanda Computer Science Department School of Computer Science Bina Nusantara University Jakarta Indonesia 11480 Game Application & Technology Program Computer Science Department School of Computer Science Bina Nusantara University Jakarta Indonesia 11480

With the increase of interest in Facial Expression Recognition (FER) in the past few decades. Several challenges surfaced with the invention of many different FER models which are often based on Convolution Neural Network (CNN) architectures. Recently, an attention-based transformer model has been presented to address FER. One of the major issues with Transformers is the need for a large data quantity for training. Therefore, in this paper, we propose to learn how to fine-tune a vision transformer-based (ViT) model using a limited dataset. We will be using the JAFFE Dataset, which consists of only 213 images containing seven different emotions. The proposed method is evaluated using several fine-tuning methods, such as adding dropout, data augmentation, and layer freezing. We compared the models implemented with 5% dropout regularization, augmented dataset (up to 5000 images), and freezing the initial model's layers, fine-tuning around a fourth of the last layers. The best model was achieved by fine-tuning ViT L-16 with 96.06% accuracy, trained with 5% dropout in the augmented dataset, and freezing the initial 21st layers. We also compared our model to the other previous work model and the results showed that our model reached the state-of-the art for the JAFFE dataset.

关键词： Facial Expression Recognition Vision Transformer Attention Deep Learning

来源：评论

学校读者我要写书评

暂无评论

Mental illness detection using sentiment analysis in social media

引用

Procedia computer science 2024年 245卷 971-978页

作者： Kasimirus Derryl Odja Jasson Widiarta Eko Setyo Purwanto Muhamad Keenan Ario Computer Science Department School of Computer Science Bina Nusantara University Jakarta Indonesia 11480 Mobile Application & Technology Program Computer Science Department School of Computer Science Bina Nusantara University Jakarta Indonesia 11480

This research tries to detect mental illness using sentiment analysis on Reddit data, as well as comparing the performance of the k-Nearest Neighbors (k-NN), Random Forest, and Neural Network models. Using text post data from the pre-pandemic and post-pandemic periods, we concluded that the Random Forest model had the highest overall performance with an F1 Score, accuracy, recall and precision of 80.6%, making it quite effective in detecting depression. Even though the Neural Network model shows slightly lower accuracy, namely 79%, in fact this model has the lowest error rate, namely 0.06496. The k-NN model showed the lowest accuracy and higher error rate. These findings highlight the potential of sentiment analysis and machine learning in identifying mental health issues on social media and suggest that better models can improve early detection and intervention efforts.

关键词： Sentiment Analysis Mental Illness k-NN Random Forest Neural Network

来源：评论

学校读者我要写书评

暂无评论

The Prospect of Combining NFC Technology with Hologram Projection for Early Childhood Education 2

The Prospect of Combining NFC Technology with Hologram Proje...

引用

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

作者： Astriani, Maria Seraphina School of Computing and Creative Arts Bina Nusantara University Computer Science Department Jakarta11480 Indonesia

ISBN: (纸本)9798350351613

Near Field Communication or NFC is a subset of Radio Frequency Identification which enables information storage and retrieval. Its application in education remains limited. Researcher aims to combine NFC technology with hologram projection to enhance engagement and interactivity in learning. Since children often struggle to remain still during conventional learning process, there is a hope that holograms will make the learning process more interesting. The researcher uses the toy as a medium because the target is for children in early childhood education. Students can interact with smart toys while viewing educational material through hologram projection. This research contributes to Information and Communication Technology by integrating NFC to create interactive smart toys and holograms projection. Tests have demonstrated the feasibility of combining NFC technology with hologram projection by using NFC tag, reader, smartphone, and box with transparent panel. © 2024 IEEE.

关键词： Holograms

来源：评论

学校读者我要写书评

暂无评论

Using Machine Learning for Shaping the Future of Advertising for Telco Industry in Indonesia 11

Using Machine Learning for Shaping the Future of Advertising...

引用

11th International Conference on ICT for Smart Society, ICISS 2024

作者： Neilson, Davin Widagda, I Kadek Gita Pradnya Agatsya, Gde Ngurah Randy Qomariyah, Nunung Nurul Bahana, Raymond School of Computing and Creative Arts Bina Nusantara University Computer Science Department Jakarta11480 Indonesia

ISBN: (纸本)9798350389654

This paper investigates the enhancement of adver-tising efficiency through data science and machine learning techniques in dynamic business environments. It emphasizes effective advertisement planning and deployment for varied audi-ences using supervised learning methods from machine learning. Analyzing IT sector datasets, the study shows data-driven models guide in optimizing advertisement strategies to bolster market presence and growth. Key results include identifying 'Success Billing' and 'Advertising Networks' as revenue drivers, with the CatBoost regressors providing precise predictions, evidenced by an R2 Score of 0.82 and the lowest RMSE of 11966 (USD), respectively. This underscores targeted ad spend's impact on revenue and the lesser influence of subscription cycle length. The research also highlights the importance of strategic ad placement and varying effectiveness of ad networks by region, particularly in Indonesia, where a broader ad placement strategy is more effective than relying on a single popular network. © 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

Lighter Student Engagement Recognition in a Classroom Environment Using Skeletal Keypoints

IAENG International Journal of Computer Science

引用

IAENG International Journal of computer science 2025年第6期52卷 1997-2014页

作者： Tarigan, Gabriel Asael Elwirehardja, Gregorius Natanael Nugroho, Kuncahyo Setyo Pardamean, Bens Computer Science Department Binus Graduate Program – Master of Computer Science Program Bina Nusantara University Jakarta11480 Indonesia Computer Science Department School of Computer Science Bina Nusantara University Jakarta1140 Indonesia Bioinformatics and Data Science Research Center Bina Nusantara University Jakarta11480 Indonesia BINUS Graduate Program – Master of Computer Science Program Bina Nusantara University Jakarta11480 Indonesia

Student retention is crucial for educational institutions, influencing reputation, finances, and ranking metrics. Engagement, reflecting a student’s connection, interest, and effort, plays a vital role in learning, fostering critical thinking, and supporting retention. Recent advancements use students’ poses to predict engagement, providing valuable insights without disrupting the teaching-learning dynamic. The prevailing research trend leans toward employing multi-modal approaches, such as a combination of pose detection with object detection. However, current methods use out-of-date object detection methods and manual dataset creation, which is cumbersome, requiring thousands of manually annotated data. This problem is then addressed by proposing a novel method of student engagement detection, using a combination of You Only Look Once Version 8 Mini (YOLOv8m) and MediaPipe as state-of-the-art alternatives to improve both object detection and human pose estimation. The (YOLOv8m + MediaPipe) method surpasses the baseline (YOLOv4 + OpenPose) with higher accuracy (0.70 vs. 0.41) and lower cross-entropy loss (0.40 vs. 0.60) on the test set, confirmed by a statistically significant paired t-test. It also exhibits a remarkable speed advantage, around 16 times faster than the baseline in pose detection data collection rates. Despite not being designed for it, the proposed method achieves multiple keypoint detection, matching the baseline’s amount. © (2025), (International Association of Engineers). All rights reserved.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Study on Object-Relational Mapping (ORM) Data Model Performance Effects in the Oil and Gas Industry 2

Study on Object-Relational Mapping (ORM) Data Model Performa...

引用

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

作者： Pangesa, Darren Mannuela Astriani, Maria Seraphina Manuaba, Ida Arts Bina Nusantara University School of Computing and Creative Computer Science Department Jakarta11480 Indonesia

ISBN: (纸本)9798350351613

This paper discusses the performance of Object Relational Model data models in managing data for the oil and gas industry within Indonesia. The paper derives from a regulation made by the Indonesian government to standardize the format for data that falls under the category of upstream oil and gas affecting a total of 45 different data types. The paper aims to simulate ORM data models to identify their performance in managing rapidly growing data. The comparisons conducted within the paper are file space, data redundancy, and average operation time. The comparisons will be between normalized and non-normalized data models that are tested over 1000, 10,000, and 100,000 iterations of CRUD operations. A two-way ANOVA will analyze the mean differences between ORM models and database operations (Create, Read, Update, Delete) concerning average operation time. Results indicate that normalized data models perform better in average operation time compared to non-normalized models. © 2024 IEEE.

关键词： Gas industry

来源：评论

学校读者我要写书评

暂无评论

Comprehensive Analysis and Remediation of Insecure Direct Object References (IDOR) Vulnerabilities in Android APIs 1

Comprehensive Analysis and Remediation of Insecure Direct Ob...

引用

1st IEEE International Conference on Cryptography, Informatics, and Cybersecurity, ICoCICs 2023

作者： Yulianto, Semi Abdullah, Roni Reza Soewito, Benfano Bina Nusantara University Binus Graduate Program Computer Science Computer Science Department Jakarta11480 Indonesia National Institute of Science and Technology Faculty of Informatics Engineering Jakarta12630 Indonesia Bina Nusantara University Binus Graduate Program. Master of Computer Science Computer Science Department Jakarta11480 Indonesia

ISBN: (纸本)9798350339437

The escalating visibility of secure direct object reference (IDOR) vulnerabilities in API security, as indicated in the compilation of OWASP Top 10 API Security Risks, highlights a noteworthy peril to sensitive data. This study explores IDOR vulnerabilities found within Android APIs, intending to clarify their inception while evaluating their implications for application security. This study combined the qualitative and quantitative approaches. Insights were obtained from an actual penetration test on an Android app into the primary reasons for IDOR vulnerabilities, underscoring insufficient input validation and weak authorization methods. We stress the frequent occurrence of IDOR vulnerabilities in the OWASP Top 10 API vulnerability list, highlighting the necessity to prioritize them in security evaluations. There are mitigation recommendations available for developers, which recognize its limitations involving a possibly small and homogeneous selection of tested Android applications, the testing environment that could cause some inaccuracies, and the impact of time constraints. Additionally, the study noted insufficient threat modeling and root cause analysis, affecting its generalizability and real-world relevance. However, comprehending and controlling IDOR dangers can enhance Android API security, protect user data, and bolster application resilience. © 2023 IEEE.

关键词： Sensitive data

来源：评论

学校读者我要写书评

暂无评论

Mobile-Apps Gender Equality Prototype to Improve Women's Soft Skills 3

Mobile-Apps Gender Equality Prototype to Improve Women's Sof...

引用

3rd International Conference on Creative Communication and Innovative Technology, ICCIT 2024

作者： Maryani Fitriawati, Nora Sia, Chow Siing Yulistiani, Risma Zega, Grisella Meirisia Bina Nusantara University School of Information System Information System Department Jakarta Indonesia Bina Nusantara University Binus Online Learning Information System Department Jakarta Indonesia Cardiff School of Technologies Cardiff United Kingdom Bina Nusantara University School of Computer Science Computer Science Department Jakarta Indonesia

ISBN: (纸本)9798350367492

The high level of gender inequality in the world of work poses serious challenges, especially regarding participation and career development between men and women. Data shows the comparison of the Labor Force Participation Rate (LFPR) between women and men in Indonesia. One of the factors causing the low level of labor force participation rate (LFPR)."for women in Indonesia is the existence of cultural norms that are still strong in most societies. By considering obstacles such as the low level of education among women, it is clear the impact that arises from gender inequality in the structure of society. As a result, women often experience limitations in accessing training and skills development, making it difficult for them to compete equally in the world of work. Awareness of the importance of transformational initiatives in changing the gender equality paradigm in the world of work is the key to producing a positive impact. By empowering women, a positive potential is opened that not only advances women individually but also makes a significant contribution to overall economic and social progress. Through understanding that transformational initiatives and skills can strengthen gender equality in the world of work, it is also necessary to emphasize the role of technology as a catalyst in achieving this goal. Increased use of technology has been key in providing solutions to gender equality challenges across various levels of Society. Research objectives: To improve women's abilities through training provided in video form;To encourage the ability to discuss gender equality on certain topics through discussion forums;To increase access to equal employment opportunities for women through the Job Vacancy feature;To provide a consultation space regarding gender equality for a woman;To measure someone's understanding by giving an interactive quiz. The online counselling development method can be carried out using the Rapid Application Development (RAD) method and the

关键词： Black-box testing

来源：评论

学校读者我要写书评

暂无评论

Analyzing the Performance of Golang Web Frameworks Utilizing GORM in the Oil and Gas Industry 2

Analyzing the Performance of Golang Web Frameworks Utilizing...

引用

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

作者： Lie, Karsten Eugene Astriani, Maria Seraphina Manuaba, Ida Bagus Kerthyayana School of Computing and Creative Arts Bina Nusantara University Computer Science Department Jakarta11480 Indonesia

ISBN: (纸本)9798350351613

This research evaluates the request completion time, average latency, error rate, and memory usage of four Go web frameworks: Fiber, Gin, Beego, and Echo, under different concurrent connections to find a fast, reliable, and scalable framework. These frameworks utilize GORM alongside a MySQL database. The test uses appropriate data from the oil and gas industry and it is done by hitting the endpoint that initializes a double inner join query using Grafana K6. The test is done on a Windows 11 machine with an Intel Core i5-12400F and 16GB of RAM, running Go 1.19. The results show that Echo outperforms the other frameworks with fastest request completion time, least average latency, error rate, and memory usage under heavier load, showing that Echo is fast, reliable, uses minimal resources, and has great scalability. However, under lighter load, the results obtained show similar speeds and error rates for all the frameworks but Gin showed a significantly higher memory usage. © 2024 IEEE.

关键词： Gas industry

来源：评论

学校读者我要写书评

暂无评论

E-Commerce Website: A Systematic Literature Review

E-Commerce Website: A Systematic Literature Review

引用

2023 International Conference on Informatics, Multimedia, Cyber and Information Systems, ICIMCIS 2023

作者： Wihardja, Caitlyn Eugenia Widianto, Mochammad Haldi Binus University School of Computer Science Computer Science Department Bandung Campus Jakarta11480 Indonesia Bina Nusantara University School of Computer Science Computer Science Department Bandung Campus Jakarta11480 Indonesia

ISBN: (纸本)9798350307306

E-commerce websites are now an important platform for e-commerce companies to connect with their target audience and conduct online transactions. As e-commerce grows increasingly popular, the quality of a company's website has emerged as a key sign of success. This study focuses on a specific research question: How does website design affect customer trust? With that, an SLR is performed by collecting and reviewing papers on website design, consumer trust, and e-commerce from Google Scholar and using Kitchenham's methodology to investigate the relationship between website design and customer trust in e-commerce platforms. Upon completing the SLR, The study emphasizes the importance of website design in terms of user behavior, experience, and trust. It underlines the significance of e-commerce design in simplicity, ease of use, and accurate information. It encourages future web development to promote modern, user-friendly, and visually appealing designs. © 2023 IEEE.

关键词： Web Design

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：