• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1997년도 추계학술대회 논문집
• /
• pp.253-256
• /
• 1997

In this study, we developed a portable device which monitors and analyzes a vibration signal happened to machines. This device is based on PDA which is smaller thant a palm of the hand, but it has powerful computing ability as much s a computer with 100MHz CPU and an operating system called Windows CE. As a preprocess for a diagnosis of a rotating machine, training artificial neural network based on PC is performed, and the device will diagnose the condition of a rotating machine using weight values as a result of the training ANN.

• 한국음향학회지
• /
• 제38권5호
• /
• pp.607-613
• /
• 2019

본 논문은 external knowledge를 사용한 lattice 없는 상호 정보 최대화(Lattice Free Maximum Mutual Information, LF-MMI) 기반 음향 모델링 방법을 제안한다. External knowledge란 음향 모델에서 사용하는 학습 데이터 이외의 문자열 데이터를 말한다. LF-MMI란 심층 신경망(Deep Neural Network, DNN) 학습의 최적화를 위한 목적 함수의 일종으로, 구별 학습에서 높은 성능을 보인다. LF-MMI에는 DNN의 사후 확률을 계산하기 위해 음소의 열을 사전 확률로 갖는다. 본 논문에서는 LF-MMI의 목적식의 사전 확률을 담당하는 음소 모델링에 external knowlege를 사용함으로써 과적합의 가능성을 낮추고, 음향 모델의 성능을 높이는 방법을 제안한다. External memory를 사용하여 사전 확률을 생성한 LF-MMI 모델을 사용했을 때 기존 LF-MMI와 비교하여 14 %의 상대적 성능 개선을 보였다.

• 컴퓨터교육학회논문지
• /
• 제20권3호
• /
• pp.1-11
• /
• 2017

본 연구는 문제 해결식 수업(PBL)을 대학교 컴퓨터 시스템프로그래밍 교육에 적용하고 이 수업이 학생들의 학업성취도와 직업기초능력에 어떤 영향을 미치는지 살펴보았다. 교과목 분석을 통해 과목을 크게 세 부분으로 나누고, 각 부분의 교육을 위해 한 문제씩을 출제하였다. 학업성취도에 미치는 영향을 측정하기 위해 2014년 학생들과의 성적을 비교하였고, 직업기초능력에 미치는 영향을 측정하기 위해서 한국직업능력개발원에서 주관중인 대학생 핵심역량진단시스템(K-CESA)를 활용하여 PBL 교육전과 후에 진단을 수행하였다. 분석 결과, 학업성취도와 직업기초능력 모두 유의미한 향상 효과가 있었다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제5권3호
• /
• pp.560-574
• /
• 2011

In this paper, we propose a hybrid method of Machine Learning (ML) algorithm and a rule-based algorithm to implement a lightweight Named Entity (NE) extraction system for Korean SMS text. NE extraction from Korean SMS text is a challenging theme due to the resource limitation on a mobile phone, corruptions in input text, need for extension to include personal information stored in a mobile phone, and sparsity of training data. The proposed hybrid method retaining the advantages of statistical ML and rule-based algorithms provides fully-automated procedures for the combination of ML approaches and their correction rules using a threshold-based soft decision function. The proposed method is applied to Korean SMS texts to extract person's names as well as location names which are key information in personal appointment management system. Our proposed system achieved 80.53% in F-measure in this domain, superior to those of the conventional ML approaches.

• Journal of Multimedia Information System
• /
• 제8권2호
• /
• pp.101-110
• /
• 2021

In computer vision, single-image super resolution has been an area of research for a significant period. Traditional techniques involve interpolation-based methods such as Nearest-neighbor, Bilinear, and Bicubic for image restoration. Although implementations of convolutional neural networks have provided outstanding results in recent years, efficiency and single model multi-scalability have been its challenges. Furthermore, previous works haven't placed enough emphasis on real-number scalability. Interpolation-based techniques, however, have no limit in terms of scalability as they are able to upscale images to any desired size. In this paper, we propose a convolutional neural network possessing the advantages of the interpolation-based techniques, which is also efficient, deeming it suitable in practical implementations. It consists of convolutional layers applied on the low-resolution space, post-up-sampling along the end hidden layers, and additional layers on high-resolution space. Up-sampling is applied on a multiple channeled feature map via bicubic interpolation using a single model. Experiments on architectural structure, layer reduction, and real-number scale training are executed with results proving efficient amongst multi-scale learning (including scale multi-path-learning) based models.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2021년도 춘계학술대회
• /
• pp.252-254
• /
• 2021

산업안전보건연구원에 따르면 2019년 산업재해자 수가 109,242명으로 2018년에 비해 6.8% 증가하였다. 이러한 산업 안전보건 분야는 질병보다 사고가 더 자주 발생하고 있다. 이러한 상황에서 정부와 기업은 건설 시공 분야에서 ICT 기반 현장 안전사고 예방 핵심 기술 개발이 논의되고 있는 실정이다. 이러한 분야에서 최근 컴퓨터 비전과 인공지능을 활용한 기술들이 많이 사용되고 있다. 본 논문에서는 공장 근로자들의 안전관리를 위한 학습 데이터를 구축하고 YOLOv4를 기반으로 모델을 학습시켰다. 이를 통해 공장에서 근로자들의 위험 상황을 예측하는 초기 연구로써 활용할 수 있을 것으로 사료된다.

• 대한간호학회지
• /
• 제25권4호
• /
• pp.617-632
• /
• 1995

Nowadays, most big hospitals have a computer system to manage their administration. For maxi mum effectiveness in managing the computer system, an analysis of the variables affecting its implementation is necessary from the beginning. This study was done to analyze the variables influencing the operation of a hospital information system (HIS). The theoretical base for this study considered the combined effects of user expectations of computerization, and computer-anxiety. The relationship between variables in the theoretical base were analyzed and the individual characteristics influencing each variable were also analyzed. This study was done in two steps. First, 344 nurses were given an initial questionnaire developed to evaluate the reliability of the items. Based on the results, a second revised questionnaire was administered to 88 nurses who had been working in the areas where HIS was applied. The results of the first and second steps of the study are as follows 1. The initial study was done with nurses who were trained on the computer system briefly before HIS was implemented. The individual characteristics influencing computer anxiety and expectation regarding computer system usage in that initial study included, length of career, type of degree or certification, previous experiences with a computer, training on a computer, desire for computer training, and level of acceptance of a computerized work environment. But in the second study with nurses working in areas of the hospital where HIS was introduced, the work site was the only influencing characteristics. There-fore, in applying a computer system, overcoming work-environment barriers will be more import-ant than any individual characteristics. 2. The computer anxiety of the nurses in both groups, before and after the computer system ap-plication, was below the average level but the expectation of the effects of computerization was above average. The nurses using the computer program showed an above average level of satis-faction with the computer system itself, and with its effect on their efficiency. Therefore, the ability of nurses operating HIS will be positively. predictive. 3. For the variables included in the theoretical framework of the study, all of the correlational coefficients were statistically significant in the analysis of variation correlation. Therefore, the theoretical base of the study, "expectation in con junction with computer anxiety" can be considered an model which can be evaluated. Accord-ing to our analysis, the higher the level of nurses' motivation to use the computer system and the lower the anxiety about computer usage, the higher the possibility of computer system acceptance by nurses. The results of this study showed that in applying a computer system in the hospital, the main characteristic influencing acceptance was where the individual worked rather than personal characteristics such as length of career, type of degree or certification, and previous experiences with a computer. Therefore, it is suggested that the first step in uncovering and eliminating hindrance factors in ap-plication of a computer system should be an analysis of working conditions in relation to the functional content of the computer system. The suitability of the theoretical model based on the hypothesis ap-plied in this study should be further tested.

• International Journal of Computer Science & Network Security
• /
• 제21권8호
• /
• pp.65-70
• /
• 2021

Waste accumulation is becoming a significant challenge in most urban areas and if it continues unchecked, is poised to have severe repercussions on our environment and health. The massive industrialisation in our cities has been followed by a commensurate waste creation that has become a bottleneck for even waste management systems. While recycling is a viable solution for waste management, it can be daunting to classify waste material for recycling accurately. In this study, transfer learning models were proposed to automatically classify wastes based on six materials (cardboard, glass, metal, paper, plastic, and trash). The tested pre-trained models were ResNet50, VGG16, InceptionV3, and Xception. Data augmentation was done using a Generative Adversarial Network (GAN) with various image generation percentages. It was found that models based on Xception and VGG16 were more robust. In contrast, models based on ResNet50 and InceptionV3 were sensitive to the added machine-generated images as the accuracy degrades significantly compared to training with no artificial data.

• 정보처리학회논문지:컴퓨터 및 통신 시스템
• /
• 제5권11호
• /
• pp.393-402
• /
• 2016

국방 L-V-C 시스템은 물리적 시간축에 따라 이벤트가 진행되는 Live 시스템, 컴퓨터 상에서 실제시간에 근접한 시간 사건에 의해 지배되는 Virtual 시스템 및 진행 시간에 관계없이 사건간 인과관계에만 의존적인 Constructive 시스템 등이 혼재된 분산형 복잡 시스템이다. 최근 이들 훈련 시스템들을 연동하여 최적의 훈련효과를 얻고자하는 LVC 연동 훈련체계에 대한 요구가 전세계적으로 증가하고 있다. 그러나, 기존에는 이론적이고 논리적인 접근 방식 혹은 부분적인 연동만이 제한적으로 제안되어 온 반면, 전 시스템적으로 LVC 훈련체계들을 연동시킬 수 있는 실제적인 기술은 국내외적으로 드문 상황이다. 이에 본 논문에서는 각 훈련시스템의 고유한 특성을 지원하는 분산시스템 연동 프로토콜들을 상위 개념에서 통합하고, 데이터와 이벤트에 대해 동일한 글로벌 시간과 상태를 유지하기 위한 데이터 중심 미들웨어 기반의 새로운 연동 아키텍쳐를 설계하고 구현한다. 또한, 구현된 연동 아키텍쳐를 기반으로 실제 L-V-C 시스템들을 모사한 시연 시스템들을 활용하여 그 성능을 검증하고 유효성을 증명한다.

• 한국컴퓨터정보학회논문지
• /
• 제16권6호
• /
• pp.71-77
• /
• 2011

본 논문에서는 번호판 인식 시스템에서 번호판 영상의 화질 개선을 위하여 국부 블록(Local block : LB) 학습기반의 초해상도 알고리즘을 제안한다. 본 논문에서 국부 블록은 영상 내에서 정보를 담고 있는 최소 단위로 정의하였으며, 학습의 기본 단위가 된다. 제안된 방법은 먼저 다양한 환경에 적합한 훈련 국부 블록 set을 생성하였다. 훈련 국부 블록 set은 고해상도 국부 블록과 저해상도 국부 블록의 순서쌍으로 구성되며 다양한 크기의 번호판과 열화 영상에 대응하기 위하여 다양한 크기와 열화를 갖는 저해상도 국부 블록 훈련 set을 구성하였다. 그 다음으로는 저해상도 입력 영상에서 복원해야할 정보를 훈련 국부 블록 set에서 추출/융합하는 과정을 제안하였다. 모의 실험결과, 열화된 저해상도 번호판 영상에 대해 제안한 방법이 효과적인 복원 성능을 나타내는 것을 확인할 수 있었다.