• 한국전자통신학회논문지
• /
• 제17권3호
• /
• pp.483-490
• /
• 2022

말라리아는 기생충에 의해 발생하는 질병으로 전 세계에 퍼져있다. 말라리아 셀을 인식하는데 일반적으로 두꺼운 혈흔과 얇은 혈흔 검사 방법이 사용되지만 이러한 방법은 많은 수작업 계산이 필요하여 효율성과 정확성이 매우 낮을 뿐만 아니라 빈민국에는 병리학자가 부족하여 말라리아 치명율이 높다. 본 논문에서는 특징 추출기, 잔류 구조와 완전 연결층으로 구성되고, 전이 학습을 이용한 말라리아셀 이미지를 인식하는 모델을 제안한다. VGG-19 모델의 사전 학습된 파라미터가 사용될 때 일부 컨볼루션층의 파라미터는 고정되고, 모델의 데이터에 맞추기 위하여 미세조정이 사용된다. 그리고 제안된 모델과 비교하기 위하여 잔류 구조가 없는 말라리아셀 인식 모델을 구현한다. 실험 결과 잔류 구조를 사용한 모델이 잔류 구조가 없는 모델에 비하여 성능이 우수 하였으며, 최신 논문과 비교하여 가장 높은 97.33%의 정확도를 보여주었다.

• Smart Structures and Systems
• /
• 제29권1호
• /
• pp.93-103
• /
• 2022

With the wider availability of sensor technology through easily affordable sensor devices, several Structural Health Monitoring (SHM) systems are deployed to monitor vital civil infrastructure. The continuous monitoring provides valuable information about the health of the structure that can help provide a decision support system for retrofits and other structural modifications. However, when the sensors are exposed to harsh environmental conditions, the data measured by the SHM systems tend to be affected by multiple anomalies caused by faulty or broken sensors. Given a deluge of high-dimensional data collected continuously over time, research into using machine learning methods to detect anomalies are a topic of great interest to the SHM community. This paper contributes to this effort by proposing a relatively new time series representation named "Shapelet Transform" in combination with a Random Forest classifier to autonomously identify anomalies in SHM data. The shapelet transform is a unique time series representation based solely on the shape of the time series data. Considering the individual characteristics unique to every anomaly, the application of this transform yields a new shape-based feature representation that can be combined with any standard machine learning algorithm to detect anomalous data with no manual intervention. For the present study, the anomaly detection framework consists of three steps: identifying unique shapes from anomalous data, using these shapes to transform the SHM data into a local-shape space and training machine learning algorithms on this transformed data to identify anomalies. The efficacy of this method is demonstrated by the identification of anomalies in acceleration data from an SHM system installed on a long-span bridge in China. The results show that multiple data anomalies in SHM data can be automatically detected with high accuracy using the proposed method.

• Smart Structures and Systems
• /
• 제29권1호
• /
• pp.117-127
• /
• 2022

Data anomalies seriously threaten the reliability of the bridge structural health monitoring system and may trigger system misjudgment. To overcome the above problem, an efficient and accurate data anomaly detection method is desiderated. Traditional anomaly detection methods extract various abnormal features as the key indicators to identify data anomalies. Then set thresholds artificially for various features to identify specific anomalies, which is the artificial experience method. However, limited by the poor generalization ability among sensors, this method often leads to high labor costs. Another approach to anomaly detection is a data-driven approach based on machine learning methods. Among these, the bidirectional long-short memory neural network (BiLSTM), as an effective classification method, excels at finding complex relationships in multivariate time series data. However, training unprocessed original signals often leads to low computation efficiency and poor convergence, for lacking appropriate feature selection. Therefore, this article combines the advantages of the two methods by proposing a deep learning method with manual experience statistical features fed into it. Experimental comparative studies illustrate that the BiLSTM model with appropriate feature input has an accuracy rate of over 87-94%. Meanwhile, this paper provides basic principles of data cleaning and discusses the typical features of various anomalies. Furthermore, the optimization strategies of the feature space selection based on artificial experience are also highlighted.

• Korean Journal of Acupuncture
• /
• 제39권3호
• /
• pp.71-83
• /
• 2022

Objectives : Instructional design is the systematic approach to the Analysis, Design, Development, Implementation, and Evaluation of learning materials and activities. We aimed to apply the rapid prototyping to instructional systems design (RPISD) in meridianology laboratory, a subject in which students train acupuncture to develop lesson plan. Methods : The needs of the stakeholders including client, subject matter expert and students were analyzed using the performance needs analysis model. Task analysis was implemented by observation and interview. First prototype was drafted and implemented in meridianology laboratory class once. The second prototype was modified from the first, by usability evaluation of the stakeholders. Results : The client requested an electronically documented manual to improve the quality of acupuncture training. The learner requested an extension of practice time and detailed practice guidelines. The main problems of students' performance were some cases of violation of clean needle technique, the lack of communication between the operator and recipient in direct, and lack of confidence in their own performance. Stakeholders were generally satisfied with the proposed first prototype. Second prototype of lesson plan was produced by modifying some contents. Conclusions : A lesson plan was developed by applying the systematic RPISD model. It is expected that the developed instructional design may contribute to the quality improvement of meridianology laboratory education.

• Smart Structures and Systems
• /
• 제31권4호
• /
• pp.365-381
• /
• 2023

The existing vision-based techniques for inspection and condition assessment of civil infrastructure are mostly manual and consequently time-consuming, expensive, subjective, and risky. As a viable alternative, researchers in the past resorted to deep learning-based autonomous damage detection algorithms for expedited post-disaster reconnaissance of structures. Although a number of automatic damage detection algorithms have been proposed, the scarcity of labeled training data remains a major concern. To address this issue, this study proposed a semi-supervised learning (SSL) framework based on consistency regularization and cross-supervision. Image data from post-earthquake reconnaissance, that contains cracks, spalling, and exposed rebars are used to evaluate the proposed solution. Experiments are carried out under different data partition protocols, and it is shown that the proposed SSL method can make use of unlabeled images to enhance the segmentation performance when limited amount of ground truth labels are provided. This study also proposes DeepLab-AASPP and modified versions of U-Net++ based on channel-wise attention mechanism to better segment the components and damage areas from images of reinforced concrete buildings. The channel-wise attention mechanism can effectively improve the performance of the network by dynamically scaling the feature maps so that the networks can focus on more informative feature maps in the concatenation layer. The proposed DeepLab-AASPP achieves the best performance on component segmentation and damage state segmentation tasks with mIoU scores of 0.9850 and 0.7032, respectively. For crack, spalling, and rebar segmentation tasks, modified U-Net++ obtains the best performance with Igou scores (excluding the background pixels) of 0.5449, 0.9375, and 0.5018, respectively. The proposed architectures win the second place in IC-SHM2021 competition in all five tasks of Project 2.

• 현장농수산연구지
• /
• 제10권1호
• /
• pp.29-41
• /
• 2008

This research was conducted to develop a technology transfer and farmer's extension of newly released technology of Bokto seeding method for crop and vegetable production based on the theory of Asian and Pacific Center for Transfer of Technology(APCTT). This technology has recently transferred to not only Korea but also other countries like North Korea, China, Japan, Taiwan, Russia and Africa(Cameroon, Sudan and South Africa) since 2005. It has known as a highly reduction of production cost in terms of labors, chemical fertilizer and pesticides as well as environmental friendly due to a deep and side banded placement of chemical fertilizer at basal application. In addition this technology was proven to a precision farming on sowing depth and mechanism of chemical application method and also highly resistant against disasters like typhoon, flooding, low temperature, drought and lodging due to silicate application. It has improved a constraints such as a poor seedling establishment, weed occurrence, lodging, low yield and poor grain and eating quality in the previous direct seeding methods but still have a problem in occurrence of weedy rice and ununiformed operation of wet or flooded soil condition. Also this technology has a limit in marketing and A/S system. Based on a theory of APCTT evaluation and analysis this technology may be more concentrated on establishment of a special cooperation team among researcher and scientists, extension workers, industry sections and governmental sectors in order to rapidly transfer this technology to farmer's field. Also there will be needed to operate a web site for this newly released technology to inform and exchange an idea, experiences and newly improved information. A feed back system might be operated in this technology as well to improve a technology under way on users' operation. Also user's manual will be internationally released and provided for farmer's instruction and training at field site.

• 근관절건강학회지
• /
• 제16권1호
• /
• pp.55-65
• /
• 2009

Purpose: The purpose of this study was to determine the effects of water exercise program on physical fitness, pain and quality of life(QoL) in patients with osteoarthritis. Methods: Thirty-four old women were divided into the exercise(EG, n=18) and control groups(CG, n=16) after 6-week self-help education program. The EG carried out 6-week supervised water exercise program(60min/day, 2days/wk). Physical fitness, pain, and QoL were assessed by Senior Fitness Test Manual(Rikli & Jones, 2005), Pain rating scale(0-10) and World Health Organization QoL BREF(Min et al., 2000), respectively. Results: Both EG and CG increased upper and lower-body strength(all, p<.05), which were measured by arm curl and chair stand, respectively. For flexibility test, the EG increased upper and lower body(all, p<.05). Balance of the EG increased(p<.05), but not in the CG. Pain significantly decreased in the EG post training(p=.000). However, both EG and CG did not significantly improve for QoL. Conclusion: Six weeks of water exercise program did induce significant improvement in physical fitness and pain control in patients with osteoarthritis.

• 문화기술의 융합
• /
• 제9권1호
• /
• pp.347-353
• /
• 2023

현재 軍(육군) 입대 장병은 신병훈련소에서 신체에 대한 치수 측정(자동, 수동) 및 샘플 피복을 착용해 본 후, 희망하는 치수로 피복을 지급받고 있다. 하지만, 민간 평상복보다 상대적으로 매우 세분화된 치수 체계를 적용하고 있는 軍에서는 이와 같은 치수 측정 과정에서 발생하는 측정된 치수의 낮은 정확도로 인해 지급받은 피복이 제대로 맞지 않아 피복을 교체하는 빈도가 매우 빈번히 발생하고 있다. 뿐만 아니라 서구적으로 변화된 MZ 세대의 체형변화를 반영하지 않고, 10여 년 전(前)에 수집된 구세대 체형 데이터 기반의 치수 체계를 적용함으로써 재고량이 비효율적으로 관리되는 문제점이 있다. 즉, 필요한 규격의 피복은 부족하고 불필요한 규격의 피복재고는 다수 발생하고 있다. 따라서, 피복 교체빈도를 감소시키고 재고관리의 효율성을 향상하기 위해 딥러닝 기반의 신체 치수 자동측정과 빅데이터 분석 및 머신러닝 기반의 "입대 장병 개인 맞춤형 피복 자동 추천 시스템"을 제안한다.

• 문화기술의 융합
• /
• 제9권3호
• /
• pp.435-442
• /
• 2023

본 연구의 목적은 간호대학생의 호흡기 감염질환 지식과 임상수행에 대해 표준화 환자를 활용한 시뮬레이션 교육의 효과를 분석하는 것이다. 간호학과 4학년생 112명을 대상으로 2020년 3월 2일부터 6월 15일까지 표준화 환자를 활용하여 전염성 호흡기계 환자 간호를 위한 단일군 전후 설계이다. 호흡기 감염병 교육 프로그램 '폐 감염병 지식', '손 씻기', '마스크 착용', '환자 및 간병인에게 마스크 착용 유도', '정맥주사' '3way 주사', '외과적 무균술', '소독 의료기기', '오염된 린넨 관리', '감염자 관리 매뉴얼' 등 10개 교육 과제 수행이며 강의, 기술 훈련, 표준화된 환자를 이용한 시뮬레이션, 디브리핑으로 구성되었다. 표준화된 환자를 이용한 시뮬레이션 교육 후 호흡기 감염질환에 대한 학생의 지식과 임상수행능력이 유의미한 향상을 보였으며, 다양한 감염관리 실습에 활용될 것으로 기대한다.

• 한국컴퓨터정보학회논문지
• /
• 제29권2호
• /
• pp.101-108
• /
• 2024

AI-OCR은 광학 문자 인식(OCR) 기술과 Artificial intelligence(AI)의 결합으로 사람의 인식이 필요하던 OCR의 단점을 보완하는 기술 향상을 이뤄내고 있다. AI-OCR의 성능을 높이기 위해서는 다양한 학습데이터의 훈련이 필요하다. 하지만 이미지 색상이 비슷한 밝기를 가진 경우에는 인식률이 떨어지기 때문에, Homomorphic filtering(HF)을 이용한 전처리 과정으로 색상 차이를 분명하게 하여 텍스트 인식률을 높이게 된다. HF은 감마값을 이용해 이미지의 고주파와 저주파를 각각 조절한다는 점에서 텍스트 추출에 적합하지만 감마값의 조절이 수동적으로 이뤄지는 단점이 존재한다. 본 연구는 시험적 과정을 거쳐 이미지의 대비, 밝기 및 엔트로피를 근거하는 감마의 임계값 범위를 제안한다. 제안된 감마값 범위를 적용한 HF의 실험 결과는 효율적인 AI-OCR의 높은 등장 가능성을 시사한다.