• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.51-54
• /
• 2023

최근 딥러닝 기술의 발달로 물체 탐지를 위한 객체 인식 분야가 기계학습을 접목한 연구가 급격히 증가하고 있다. 하지만, 탐지하려는 물체가 다른 객체에 가려진 경우와 같이 특수한 상황에 대한 데이터의 수량이 부족하여 성능 저하를 야기한다는 점과, 객체 탐지 수행 과정에서 작은 객체의 탐지가 어렵다는 한계점이 있다. 본 연구는 전술한 문제점을 보완할 방법을 제안한다. 데이터 증강 기법을 이용하여 클래스가 부족한 데이터의 양을 늘려 학습 데이터를 증강시켰다. 한편, SRGAN을 사용하여 작은 객체를 확대시킨 뒤 이미지를 합성시켜 데이터를 구성하였다. 제안된 방법은 PyTorch 환경에서 YOLOv5를 수행한 결과, 객체 탐지 성능이 향상되는 것을 확인할 수 있었다.

• Journal of IKEEE
• /
• v.26 no.2
• /
• pp.176-185
• /
• 2022

Data augmentation is general approach to solve overfitting of machine learning models by applying various data transformations and distortions to dataset. However, when data augmentation is applied in GAN-based model, which is deep learning image generation model, data transformation and distortion are reflected in the generated image, then the generated image quality decrease. To prevent this problem called augmentation leak, we propose a scheme that can prevent augmentation leak regardless of the type and number of augmentations. Specifically, we analyze the conditions of augmentation leak occurrence by type and implement auxiliary augmentation task classifier that can prevent augmentation leak. Through experiments, we show that the proposed technique prevents augmentation leak in the GAN model, and as a result improves the quality of the generated image. We also demonstrate the superiority of the proposed scheme through ablation study and comparison with other representative augmentation leak prevention technique.

• Journal of The Korean Association of Information Education
• /
• v.12 no.1
• /
• pp.57-63
• /
• 2008

This study uses the identification which is one of the critical components of psychological mechanism and enables replacing one's own self because of the needs of self-expression(disclosure) and creation. The study aims to improve educational effects using the realistic by increasing sense of the virtual reality and the attention. After the computer-based contents were developed and converted to be applied into robot, and then the contents were combined the student's photo and the avatar using automatic loading. Finally each one of the contents was applied to the students. The results of the investigation indicated that there were significant effects of the contents based on identification. In other words, the contents effect on student's attention, but not their academic achievement. The study could find the effect of the identification's application using the educational robot. We suggested that improving technical ability of the augmented virtuality as a face-detection and sensitive interaction may lead to the specific suggestions for educational effects for further research.

• Journal of Internet Computing and Services
• /
• v.20 no.6
• /
• pp.65-72
• /
• 2019

Data augmentation is a method that increases the amount of data through various algorithms based on a small amount of sample data. When machine learning and deep learning techniques are used to solve real-world problems, there is often a lack of data sets. The lack of data is at greater risk of underfitting and overfitting, in addition to the poor reflection of the characteristics of the set of data when learning a model. Thus, in this paper, through the layer-wise data augmenting method at each layer of deep neural network, the proposed method produces augmented data that is substantially meaningful and shows that the method presented by the paper through experimentation is effective in the learning of the model by measuring whether the method presented by the paper improves classification accuracy.

• Journal of The Korean Association For Science Education
• /
• v.43 no.2
• /
• pp.125-137
• /
• 2023

This study analyzed the types of knowledge building discourse and knowledge building processes in small group learning using augmented reality. Eight 8th grade students took classes using augmented reality in solubility, boiling and melting points. These classes were carried out twice and all the classes were videotaped and recorded. Every student participated in a semi-structured interview. In the types of knowledge building discourse, the proportion of knowledge sharing and knowledge construction was similar. Beneath the knowledge sharing, the proportion of introductory level discussion was higher than identifying key elements of augmented reality. Recalling existing knowledge rarely appeared. Under the knowledge construction, the proportion of advanced level discussion was the highest and the proportion of sharing and critiquing ideas at a different level and efforts to rise above current levels of explanation was similar. The introductory level discussion and identifying key elements of augmented reality were developed into efforts to rise above current levels of explanation and sharing and critiquing ideas at a different level. Visualized results of knowledge building processes showed all the students' graph drew an upward curve, though cumulative number of impact value was different by each student. As a result of the study, effective ways of improving small group learning using augmented reality are discussed.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.327-330
• /
• 2023

머신러닝과 딥러닝 모델의 사용이 급증함에 따라 충분한 데이터 확보의 중요성이 부각되고 있다. 이에 따라 생성 모델을 통한 데이터 증강 기술이 주목받고 있으나, 증강 데이터를 활용했을 때 학습의 성능 분석은 아직 부족하다. 따라서 본 연구에서는 데이터 증강 시나리오에 따라 증강 비율별 합성 데이터의 유용성을 조사하고자 한다. 본 연구에서는 테이블 데이터를 증강하는 것에 초점을 맞추었으며, 이를 위해 테이블 데이터를 합성할 때 유용한 성능을 보이는 딥러닝 모델 CTGAN을 활용하였다. 실험에서 데이터를 증강하는 두 가지 다른 시나리오를 고려한 결과, 두 시나리오에서 모두 실험에서 설정한 증강 비율까지의 합성 데이터가 유용한 결과를 보임을 확인할 수 있었다.

• The Journal of the Korea Contents Association
• /
• v.22 no.6
• /
• pp.78-91
• /
• 2022

The importance of emotional intelligence education in the development of children's augmented reality education content is overlooked. Therefore, in-depth research is needed to develop children's emotional intelligence. This study was conducted through theoretical consideration and case analysis. The proposal of this paper is that the augmented reality type for children aged 2 to 7 is suitable for indoor activities with marking recognition technology. To promote an understanding of emotions, a large screen is selected, and emoticon dolls or emoticon books are recommended for learning content. Children aged 7 to 11 are suitable for indoor activities of non-marker recognition technology, and can induce emotional control and emotional recognition through active manipulation. For the learning content, "3D art teaching content" and "Online Classic Musical" are recommended. Children after the age of 11 are suitable for non-marker recognition technology outdoor activities and improve each element of emotional intelligence through interaction with nature and society. For the learning content, 'Forest Play Activity through Art' and 'EQ Theater Play' are recommended. Through this paper, we intend to promote the development of children's augmented reality emotional intelligence education.

• Cartoon and Animation Studies
• /
• s.40
• /
• pp.443-476
• /
• 2015

The university e-learning classes give a major focused on practical training in the art and design field are opened such as theory classes. The cyber universities that fully gives on-line classes even open as theory classes. They speak about difficulty and limitation of operating cyber class with practice. So, many of them organize off-line special lecture at the classroom in weekend. In the reason of studying that has the constraints of time and space, the student wastage rates in cyber universities are going higher. This research focused on giving an efficient e-learning practice environment in college e-learning courses, and operated scholastic achievement test between the augmented reality(AR) based practice education and the existing classroom based practice education. The subjects of study were totally 84 people who are the freshmen of the design department in the two-year diploma course college. They were divided 3 groups which one was treated AR based practice e-learning and two others groups were existing classroom based practice. Each of the group took the same treated lecture during 7 weeks. The first of the outcome was the AR based practice e-learning was more effective than existing classroom based practice on the side of scholastic achievement. The second, the AR based practice e-learning aroused increasing in the interest in class on the side of attribute factors. The third, the AR based practice e-learning group made higher level of studying immersion than others. In consideration of this experiment was processed in the regular college academic course, the finding by this research shows AR based major practice e-learning is an alternative lecturing strategy what can be supplemented existing classroom based academic teaching methods.

• Journal of the Korea Society of Computer and Information
• /
• v.19 no.4
• /
• pp.101-109
• /
• 2014

This paper aims to design and implement a collaborative learning system based on the augmented reality. The existing augment reality-based learning systems have just focused on interactivity between a system and learners without consideration of cooperability, thereby leading to an ineffective approach to encouraging an learning system to be more supportive and conducive of and to cooperation among learners. The collaborative learning system is a learning method, with which learners achieve a common objective through critical thinking and cooperative teamwork so as to seek solutions to such fulfillment. This requires positive interdependence, proactive interactions, a sense of responsibility shared by individuals as well as the group, and development of teamwork among learners. Educators and systems assume a critical role in helping the collaborative education be effective. An educator is responsible for defining a project at the outset of learning activities, organizing groups for learners, and providing evaluation criteria applied to a group's project activities. Meanwhile, a system shall support interactions to take place while facilitating learning activities. Furthermore, an educator shall provide a system for managing and evaluating activities involving interactions among learners. This paper suggests and embodies a collaborative learning system based on the augmented reality with consideration of the aforementioned collaborative education.

• The Journal of the Korea Contents Association
• /
• v.17 no.5
• /
• pp.90-99
• /
• 2017

Sports climbing requires repeated movements within a limited space, which may cause climbers to lose their constant interest in it. Furthermore, it is important that coaches should give lessons focusing on demonstrations to make sure that learners can understand the movements on their own, However, in traditional sports climbing lessons, they give instructions on almost every movement of learners' hands and feet. Hence, there have been constant calls for replacing these existing sports climbing lessons and presenting new methods to ensure that learners can observe their coaches' demonstrations in real time and emulate them. An introduction of the image training using spatial augmented reality techniques to solve these problems may have a positive effect on the improvement of learners' motor skills and attitudes toward lessons. This study aims to verify the effectiveness of the Climbing Character Animation (CCA) as a learning tool for sports climbing. To achieve this research objective, it applied it to actual sports climbing to verify its utility. As a result, it was shown that the lessons using spatial augmented reality-based CCA had a higher effectiveness than traditional sports climbing lessons in the degree of interest inducement, emersion and learning effects.