• Journal of KIBIM
• /
• v.13 no.3
• /
• pp.21-29
• /
• 2023

ChatGPT, a chatbot based on GPT large language models, has gained immense popularity among the general public as well as domain professionals. To assess its proficiency in specialized fields, ChatGPT was tested on mainstream exams like the bar exam and medical licensing tests. This study evaluated ChatGPT's ability to answer questions related to Building Information Modeling (BIM) by testing it on Korea's BIM expertise exam, focusing primarily on multiple-choice problems. Both GPT-3.5 and GPT-4 were tested by prompting them to provide the correct answers to three years' worth of exams, totaling 150 questions. The results showed that both versions passed the test with average scores of 68 and 85, respectively. GPT-4 performed particularly well in categories related to 'BIM software' and 'Smart Construction technology'. However, it did not fare well in 'BIM applications'. Both versions were more proficient with short-answer choices than with sentence-length answers. Additionally, GPT-4 struggled with questions related to BIM policies and regulations specific to the Korean industry. Such limitations might be addressed by using tools like LangChain, which allow for feeding domain-specific documents to customize ChatGPT's responses. These advancements are anticipated to enhance ChatGPT's utility as a virtual assistant for BIM education and modeling automation.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.5
• /
• pp.501-508
• /
• 2023

This research endeavors to categorize a diverse range of hairstyles for both men and women in South Korea using hair images and subsequently generate 3D hair models based on this classification. The classification process relies on prominent visual features, resulting in the division of hairstyles into 14 distinct categories, including six styles for men and eight styles for women. By accurately matching the most appropriate hairstyle to the given hair image, the study aims to provide recommendations for the necessary hair models required for metaverse authoring tools, thus enabling realistic hair styling. This capability can be effectively utilized on platforms like metaverse, allowing users to seamlessly find and apply the 3D hair model that closely resembles their remotely captured or pre-existing hair images. Through this innovative approach, users can be presented with the most similar 3D hair model, enhancing their virtual hairstyling experience.

• Nutrition Research and Practice
• /
• v.17 no.6
• /
• pp.1185-1200
• /
• 2023

BACKGROUND/OBJECTIVES: In the United States, one in every 5 children is obese with greater likelihood in low-income households. The coronavirus disease 2019 (COVID-19) pandemic may have accelerated disparities in child obesity risk factors, such as poor dietary intake and increased sedentary behaviors, among low-income families because of financial difficulties, social isolation and other struggles. This study reveals insights into nutrition and health challenges among low-income families of young children in West Texas to better understand needs and develop interventions. SUBJECTS/METHODS: In-depth individual interviews were performed via Zoom among 11 families of children under the age of 3. A semi-structured interview guide was developed to explore 3 areas: changes in (1) dietary intake and (2) sedentary behaviors and (3) families' preferences regarding a parent nutrition education program. Each interview was audiorecorded, transcribed, and coded using MaxQDA software. RESULTS: Eating together as a family become challenging because of irregular work schedules during the COVID-19 pandemic. Most parents stated that their children's dietary habits shifted with an increased consumption of processed foods. Many parents are unable to afford healthful foods and have utilized food and nutrition assistance programs to help feed their families. All families reported that their children's screen time substantially increased compared to the pre-pandemic times. Moreover, the majority of parents did not associate child screen time with an obesity risk, so this area could be of particular interest for future interventions. Meal preparation ideas, remote modality, and early timing were identified as key intervention strategies. CONCLUSIONS: Online nutrition interventions that emphasize the guidelines for child screen time and regular meal routines will be effective and promising tools to reach low-income parents for early childhood health promotion and obesity prevention.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.48 no.2
• /
• pp.193-210
• /
• 2024

The portrait featuring Jo Ban(1341-1401), a scholar-official from the late Goryeo and early Joseon period, and his wife is the oldest surviving couple portrait in Korea. It is of great value in uncovering the clothing culture of the period given the limited number of historical artifacts and records. This study examines the historical clothing and textile patterns of Jo Ban's wife and reproduces them using 3D fashion design software program CLO. She wears jokduri, chima, and baeja over layers of jeogori, a mixture of traditional Korean and Chinese styles. Her clothing illustrates eight patterns-one flower, five geometric, and two cloud. Records and relics of similar periods show that flower and geometric patterns in her clothing follow the prevailing styles of Goryeo, while the cloud patterns are representative of early Joseon. These details are used to reproduce six different styles of Jo Ban's wife with CLO tools. Various visualizations of textile patterns are applied to materials, generating a more realistic look than her existing 3D character created with the portrait. Results of this study are expected to help promote the use of Goryeo clothing and patterns in numerous designs and enhance intuitive understanding of Goryeo clothing based on 3D visualization.

• Convergence Security Journal
• /
• v.24 no.1
• /
• pp.43-49
• /
• 2024

The convergence of education and technology has been emphasized, leading to the application of educational technology (EdTech) in the field of education. EdTech provides learner-centered, customized learning environments through various media and learning situations. In this paper, we designed hardware for EdTech-based educational tools for IoT security education in the field of cybersecurity education. The hardware is based on a dual microcontroller unit (MCU) within a single board, allowing for both attack and defense to be performed. To leverage various sensors in the Internet of Things (IoT), the hardware is modularly designed. From an educational perspective, utilizing EdTech in cybersecurity education enhances engagement by incorporating tangible physical teaching aids. The proposed research suggests that the design of IoT security education hardware can serve as a reference for simplifying the creation of a security education environment for embedded hardware, software, sensor networks, and other areas that are challenging to address in traditional education..

• Journal of Advanced Navigation Technology
• /
• v.28 no.1
• /
• pp.27-36
• /
• 2024

In this paper, we analyze the unified requirements of international association of classification societies - cyber resilience of ships, ahead of implementation of the agreement on July 1, 2024, and respond to ship cyber security and resilience programs based on 5 requirements, 17 details, and documents that must be submitted or maintained according to the ship's cyber resilience,. Measures include document management such as classification certification documents and design documents, configuration of a network with enhanced security, establishment of processes for accident response, configuration management using software tools, integrated network management, malware protection, and detection of ship network security threats with security management solutions. proposed a technology capable of real-time response.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.1
• /
• pp.105-112
• /
• 2024

In contemporary society, the expression of emotions and self-reflection are considered pivotal factors with a positive impact on stress management and mental well-being, thereby highlighting the significance of journaling. However, traditional journaling methods have posed challenges for many individuals due to constraints in terms of time and space. Recent rapid advancements in chatbot and emotion analysis technologies have garnered significant attention as essential tools to address these issues. This paper introduces an artificial intelligence chatbot that integrates the GPT-3 model and emotion analysis technology, detailing the development process of a system that automatically generates journals based on users' chat data. Through this system, users can engage in journaling more conveniently and efficiently, fostering a deeper understanding of their emotions and promoting positive emotional experiences.

• Journal of Cardiovascular Imaging
• /
• v.31 no.3
• /
• pp.135-141
• /
• 2023

BACKGROUND: Global longitudinal strain (GLS) is an accurate and reproducible parameter of left ventricular (LV) systolic function which has shown meaningful prognostic value. Fast, user-friendly, and accurate tools are required for its widespread implementation. We aim to compare a novel web-based tool with two established algorithms for strain analysis and test its reproducibility. METHODS: Thirty echocardiographic datasets with focused LV acquisitions were analyzed using three different semi-automated endocardial GLS algorithms by two readers. Analyses were repeated by one reader for the purpose of intra-observer variability. CAAS Qardia (Pie Medical Imaging) was compared with 2DCPA and AutoLV (TomTec). RESULTS: Mean GLS values were -15.0 ± 3.5% from Qardia, -15.3 ± 4.0% from 2DCPA, and -15.2 ± 3.8% from AutoLV. Mean GLS between Qardia and 2DCPA were not statistically different (p = 0.359), with a bias of -0.3%, limits of agreement (LOA) of 3.7%, and an intraclass correlation coefficient (ICC) of 0.88. Mean GLS between Qardia and AutoLV were not statistically different (p = 0.637), with a bias of -0.2%, LOA of 3.4%, and an ICC of 0.89. The coefficient of variation (CV) for intra-observer variability was 4.4% for Qardia, 8.4% 2DCPA, and 7.7% AutoLV. The CV for inter-observer variability was 4.5%, 8.1%, and 8.0%, respectively. CONCLUSIONS: In echocardiographic datasets of good image quality analyzed at an independent core laboratory using a standardized annotation method, a novel web-based tool for GLS analysis showed consistent results when compared with two algorithms of an established platform. Moreover, inter- and intra-observer reproducibility results were excellent.

• Fashion & Textile Research Journal
• /
• v.26 no.1
• /
• pp.78-87
• /
• 2024

This study aimed to propose an innovative teaching pedagogy using a half-scale body form in apparel design education and evaluate its effectiveness in augmenting students' understanding of ergonomic patterns. Constructed in alignment with Phoenix's (2018) study, which used 3D body scanning and digital editing software, the half-scale body form was created through a five-step process, encompassing body measurement, 3D body modeling, fabrication of a physical half-scale body form, pattern making, and evaluation. Implemented in an undergraduate patternmaking course offered at a 4-year university in the metropolitan Seoul, this instructional approach's effectiveness was gauged through students' course projects and exit interviews. The results underscored the positive impact of the proposed teaching pedagogy on students' grasp of ergonomic pattern development, fostering a keen understanding of diverse body shapes and sizes and the relationship between the human body and garments. Furthermore, it played a role in cultivating positive body image and self-endorsement among students. The research contributes meaningfully by presenting a fresh perspective in apparel design education, seamlessly integrating advanced anthropometric and technological tools into a conventional patternmaking classroom. It offers a novel learning experience for students majoring in apparel, creating a fun and interactive teaching environment.

• Safety and Health at Work
• /
• v.14 no.4
• /
• pp.358-367
• /
• 2023

Background: This study aimed to develop and assess the psychometric features of the Physical Activity at Workplace Questionnaire (PAWPQ) based on the Social Cognitive Theory (SCT) to evaluate employees' physical activity (PA) behaviors at the workplace. Methods: This psychometric cross-sectional study was conducted on 455 employees working in one of the gas refineries in Iran. The participants were selected using the proportional stratified sampling method in 2019. The data collection tools were a demographic information questionnaire, the short form of the International Physical Activity Questionnaire (IPAQ), and a questionnaire developed based on the SCT, whose psychometric features were confirmed in terms of validity and reliability. Data were analyzed using SPSS₂₂ and AMOS₂₀ software. Results: The first version of PAWPQ-SCT had 74 items. After evaluating content and face validity, nine items were removed. The results of the content validity index (0.98), content validity ratio (0.86), and impact score (3.62) were acceptable for the whole instrument. In exploratory factor analysis, after removing seven items-58-item final version of the scale-six factors could explain 73.54% of the total variance. The results of structural equation modeling showed the acceptable fit of the model into the data (RMSEA = 0.052, CFI = 0.917, NFI = 0.878, TLI = 0.905, IFI = 0.917, CMIN/DF = 2.818). Cronbach's alpha coefficient and Intraclass Correlation were 0.90 and 0.86, respectively. Conclusion: This study confirmed that the psychometric features of the 58-item final version of PAWPQSCT constructs were acceptable in a sample of Iranian employees. This questionnaire can be used as a valid and reliable tool to evaluate Iranian employees' PA behaviors and develop effective educational interventions for workers and managers.