• 산업경영시스템학회지
• /
• 제47권1호
• /
• pp.9-19
• /
• 2024

Deep learning-based computer vision anomaly detection algorithms are widely utilized in various fields. Especially in the manufacturing industry, the difficulty in collecting abnormal data compared to normal data, and the challenge of defining all potential abnormalities in advance, have led to an increasing demand for unsupervised learning methods that rely on normal data. In this study, we conducted a comparative analysis of deep learning-based unsupervised learning algorithms that define and detect abnormalities that can occur when transparent contact lenses are immersed in liquid solution. We validated and applied the unsupervised learning algorithms used in this study to the existing anomaly detection benchmark dataset, MvTecAD. The existing anomaly detection benchmark dataset primarily consists of solid objects, whereas in our study, we compared unsupervised learning-based algorithms in experiments judging the shape and presence of lenses submerged in liquid. Among the algorithms analyzed, EfficientAD showed an AUROC and F1-score of 0.97 in image-level tests. However, the F1-score decreased to 0.18 in pixel-level tests, making it challenging to determine the locations where abnormalities occurred. Despite this, EfficientAD demonstrated excellent performance in image-level tests classifying normal and abnormal instances, suggesting that with the collection and training of large-scale data in real industrial settings, it is expected to exhibit even better performance.

• 한국멀티미디어학회논문지
• /
• 제21권12호
• /
• pp.1439-1447
• /
• 2018

Palm image acquisition without contact has advantages in user convenience and hygienic issues, but such images generally display more image variations than those acquired employing a contact plate or pegs. Therefore, it is necessary to develop a palmprint identification method which is robust to affine variations. This study proposes a deep learning approach which can effectively identify contactless palmprints. In general, it is very difficult to collect enough volume of palmprint images for training a deep convolutional neural network(DCNN). So we adopted an approach to use a pretrained DCNN. We designed two new DCNNs based on the VGGNet. One combines the VGGNet with SVM. The other add a shallow network on the middle-level of the VGGNet. The experimental results with two public palmprint databases show that the proposed method performs well not only contact-based palmprints but also contactless palmprints.

• 문화기술의 융합
• /
• 제10권2호
• /
• pp.505-510
• /
• 2024

철도 레일은 차륜과 레일의 접촉면인 레일 표면에서 구름 접촉 피로 균열이 상시 발생할 수 있는 조건이기 때문에 균열의 상태를 철저히 점검하고 절손을 방지하기 위한 정밀한 점검 및 진단이 필요하다. 최근 궤도 시설의 성능 평가에 대한 세부 지침에서는 궤도 성능평가를 위한 방법과 절차에 관한 필요사항을 제시하고 있다. 그러나 레일 표면 손상을 진단하고 등급을 산정하는 것은 주로 외관 조사(육안 조사)에 의존하며, 이는 점검자의 주관적인 판단에 따른 정성적인 평가에 의존할 수밖에 없는 실정이다. 따라서 본 연구에서는 Fast R-CNN을 사용하여 레일 표면 결함 검출에 대한 딥러닝 모델 연구를 수행하였다. 레일 표면 결함 이미지의 데이터 세트를 구축한 후, 모델을 테스트하였다. 딥러닝 모델의 성능평가 결과에서 mAP가 94.9%로 나타났다. Fast R-CNN의 균열 검출 효과가 높기 때문에 이 모델을 사용하면 레일표면 결함을 효율적으로 식별할 수 있을 것으로 판단된다.

• 산업융합연구
• /
• 제20권2호
• /
• pp.71-78
• /
• 2022

본 연구는 치위생과 학생들의 전공 만족도를 높일 수 있는 방안을 마련하여 비대면 수업 상황에서 치위생과 학생들이 인지하는 치위생 전공 교육에 적응하고 학업에 집중할 수 있도록 바람직한 비전을 제시하고자 한다. 자료 수집은 2021년 6월 20일부터 8월 19일까지 서울·충청·강원 지역 대학 치위생과 학생들을 편의표본추출 하였으며 자기기입식 설문법으로 작성하였다. 연구대상자의 일반적 특성에 따른 변수의 차이는 t-검정, 일원분산분석을 하였고, 사후검정은 Tukey로 검증하였다. 변수 간의 상관관계는 Pearson's correlation을 하였고, 치위생과 학생의 전공만족도에 영향을 미치는 요인은 Multiple regression으로 하였다. 치위생과 학생의 자기 주도적 학습능력은 3.57점이었고, 학습몰입은 3.02점이었으며, 전공만족도는 3.54점이었다. 치위생과 학생들은 학년이 높고, 대학생활 만족도가 높을수록, 학습몰입과 자기 주도적 학습능력이 높을수록 전공만족도가 높은 것으로 나타났다. 따라서 코로나 19로 어려운 여건 속에서 전공만족도 향상을 위해 체계적인 프로그램 개발 및 적용이 필요한 것으로 나타났다.

• Asia pacific journal of information systems
• /
• 제29권4호
• /
• pp.752-770
• /
• 2019

The purpose of this study is to incorporate telemarketing processes to improve telemarketing performance. For this application, we have attempted to mix the model of machine learning to extract potential customers with personalisation techniques to derive recommended products from actual contact. Most of traditional recommendation systems were mainly in ways such as collaborative filtering, which predicts items with a high likelihood of future purchase, based on existing purchase transactions or preferences for products. But, under these systems, new users or items added to the system do not have sufficient information, and generally cause problems such as a cold start that can not obtain satisfactory recommendation items. Also, indiscriminate telemarketing attempts can backfire as they increase the dissatisfaction and fatigue of customers who do not want to be contacted. To this purpose, this study presented a multi-purpose hybrid recommendation algorithm to achieve two goals: to select customers with high possibility of contact, and to recommend products to selected customers. In addition, we used subscription data from telemarketing agency that handles insurance products to derive realistic applicability of the proposed recommendation system. Our proposed recommendation system would certainly solve the cold start and scarcity problem of existing recommendation algorithm by using contents information such as customer master information and telemarketing history. Also. the model could show excellent performance not only in terms of overall performance but also in terms of the recommendation success rate of the unpopular product.

• 한국조명전기설비학회:학술대회논문집
• /
• 한국조명전기설비학회 1999년도 학술대회논문집-국제 전기방전 및 플라즈마 심포지엄 Proceedings of 1999 KIIEE Annual Conference-International Symposium of Electrical Discharge and Plasma
• /
• pp.121-126
• /
• 1999

In this paper, to bring under robust and accurate control of auto-equipment systems which disturbance, parameter alteration of system, uncertainty and so forth exist, neural network controller called dynamic neural processor(DNP) is designed. In order to perform a elaborate task like as assembly, manufacturing and so forth of components, tracking control on the trajectory of power coming in contact with a target as well as tracking control on the movement course trajectory of end-effector is indispensable. Also, the learning architecture to compute inverse kinematic coordinates transformations in the manipulator of auto-equipment systems is developed and the example that DNP can be used is explained. The architecture and learning algorithm of the proposed dynamic neural network, the DNP, are described and computer simulations are provided to demonstrate the effectiveness of the proposed learning method using the DNP.

• 한국조명전기설비학회지:조명전기설비
• /
• 제13권2호
• /
• pp.187-187
• /
• 1999

in order to perform a elaborate task like as assembly, manufacturing and so forth of components, tracking control on the trajectory of power coming in contact with a target as well as tracking control on the movement course trajectory of end-effector is indispensable. In this paper, to bring under robust and accurate control of auto-equipment systems which disturbance, parameter alteration of system, uncertainty and so forth exist, neural network controller called dynamic neural processor(DNP) is designed. Also, the learning architecture to compute inverse kinematic coordinates transformations in the manipulator of auto-equipment system is developed and the example that DNP can be used is explained. The architecture and learning algorithm of the proposed dynamic neural network, the DNP, are described and computer simulation are provided to demonstrate the effectiveness of the proposed learning method using the DNP.

• International Journal of Internet, Broadcasting and Communication
• /
• 제16권2호
• /
• pp.267-277
• /
• 2024

This paper explores the use of machine learning in game production to create goal-oriented, realistic animations for skeleton monsters. The purpose of this research is to enhance realism by implementing intelligent movements in monsters within game development. To achieve this, we designed and implemented a learning model for skeleton monsters using reinforcement learning algorithms. During the machine learning process, various reward conditions were established, including the monster's speed, direction, leg movements, and goal contact. The use of configurable joints introduced physical constraints. The experimental method validated performance through seven statistical graphs generated using machine learning methods. The results demonstrated that the developed model allows skeleton monsters to move to their target points efficiently and with natural animation. This paper has implemented a method for creating game monster animations using machine learning, which can be applied in various gaming environments in the future. The year 2024 is expected to bring expanded innovation in the gaming industry. Currently, advancements in technology such as virtual reality, AI, and cloud computing are redefining the sector, providing new experiences and various opportunities. Innovative content optimized for this period is needed to offer new gaming experiences. A high level of interaction and realism, along with the immersion and fun it induces, must be established as the foundation for the environment in which these can be implemented. Recent advancements in AI technology are significantly impacting the gaming industry. By applying many elements necessary for game development, AI can efficiently optimize the game production environment. Through this research, We demonstrate that the application of machine learning to Unity and game engines in game development can contribute to creating more dynamic and realistic game environments. To ensure that VR gaming does not end as a mere craze, we propose new methods in this study to enhance realism and immersion, thereby increasing enjoyment for continuous user engagement.

• 스마트미디어저널
• /
• 제9권3호
• /
• pp.122-129
• /
• 2020

전 세계적으로 무증상의 코로나바이러스 감염증-19 감염자가 자신이 감염된 것을 모르고 주변인들에게 전파할 수 있다는 가능성은 국민이 전염병 확산에 대한 불안과 두려움에서 벗어나지 못하고 있다는 점에서 여전히 매우 중요한 이슈이다. 본 논문에서는 K-평균 군집화 알고리즘 및 딥러닝 기반 군중 집계를 이용한 전염병 확진자 접촉 가능성 여부 판단 모니터링 시스템을 제안하였다. 모든 입력 학습 영상에 대해 300회 반복 학습한 결과, PSNR값은 21.51, 전체 데이터 셋에 대한 최종 MAE값은 67.984였다. 이는 확진자와 주변인과의 거리와 감염률 산출, 잠재적 환자 동선 주변 인원의 위험도 순 그룹 및 감염률 예측에 대한 영상 속 화질 정보, 관측치 간의 평균 절대 오차를 의미하며 각 CCTV 장면에서 군중의 수가 4,000명 이하일 때에는 평균 절대 오차 값이 0에 가까움을 증명하였다.

• Smart Structures and Systems
• /
• 제24권6호
• /
• pp.723-732
• /
• 2019

Globally road transport networks are subjected to continuous levels of stress from increasing loading and environmental effects. As the most popular mean of transport in the UK the condition of this civil infrastructure is a key indicator of economic growth and productivity. Structural Health Monitoring (SHM) systems can provide a valuable insight to the true condition of our aging infrastructure. In particular, monitoring of the displacement of a bridge structure under live loading can provide an accurate descriptor of bridge condition. In the past B-WIM systems have been used to collect traffic data and hence provide an indicator of bridge condition, however the use of such systems can be restricted by bridge type, assess issues and cost limitations. This research provides a non-contact low cost AI based solution for vehicle classification and associated bridge displacement using computer vision methods. Convolutional neural networks (CNNs) have been adapted to develop the QUBYOLO vehicle classification method from recorded traffic images. This vehicle classification was then accurately related to the corresponding bridge response obtained under live loading using non-contact methods. The successful identification of multiple vehicle types during field testing has shown that QUBYOLO is suitable for the fine-grained vehicle classification required to identify applied load to a bridge structure. The process of displacement analysis and vehicle classification for the purposes of load identification which was used in this research adds to the body of knowledge on the monitoring of existing bridge structures, particularly long span bridges, and establishes the significant potential of computer vision and Deep Learning to provide dependable results on the real response of our infrastructure to existing and potential increased loading.