• 한국가시화정보학회지
• /
• 제19권1호
• /
• pp.68-73
• /
• 2021

In this study, the detection of rubber o-ring defects was carried out using k-fold cross validation and Support Vector Machine (SVM) algorithm. The data process was carried out in 3 steps. First, we proceeded with a frame alignment to eliminate unnecessary regions in the learning and secondly, we applied gray-scale changes for computational reduction. Finally, data processing was carried out using image augmentation to prevent data overfitting. After processing data, SVM algorithm was used to obtain normal and defect detection accuracy. In addition, we applied the SVM algorithm through the k-fold cross validation method to compare the classification accuracy. As a result, we obtain results that show better performance by applying the k-fold cross validation method.

• Journal of Korean Neurosurgical Society
• /
• 제44권4호
• /
• pp.211-216
• /
• 2008

Objective : The cranioplasty and ventriculoperitoneal (VP) shunt operation have been used to treat a large cranial defect with posttraumatic hydrocephalus (PTH). The aim of this study was to evlauate the difference of outcomes between in the shunting after the cranioplasty (group 1) and the cranioplasty after the shunting (group 2) in a large flaccid cranial defect with PTH. Methods : In this study, a retrospective review was done on 23 patients undergoing the cranioplasty and VP shunt operation after the decompressive craniectomy for a refractory intracranial hypertension from 2002 to 2005. All of 23 cases had a large flaccid concave cranial defect and PTH. Ten cases belong to group 1 and 13 cases to group 2. The outcomes after operations were compared in two groups 6 months later. Results : The improvement of Glasgow outcome scale (GOS) was seen in 8 cases (80.0%) of total 10 cases in group 1, and 6 cases (46.2%) of 13 cases in group 2. Three (75.0%) of 4 cases with hemiparesis in group 1 and 3 of 6 cases (50.0%) in group 2 were improved. All cases (2 cases) with decrease of visual acuity were improved in each group. Dysphasia was improved in 3 of 5 cases (60%) in group 1 and 4 of 6 cases (66.6%) in group 2. Conclusion : These results suggest that outcomes in group 1 may be better than in group 2 for a large flaccid concave cranial defect with PTH.

• 한국압력기기공학회 논문집
• /
• 제16권2호
• /
• pp.1-9
• /
• 2020

Large-scale industrial facility structures continue to deteriorate due to the effects of operating and environmental conditions. The problems of these industrial facilities are potentially causing economic losses, environmental pollution, casualties, and national losses. Accordingly, in order to prevent disaster accidents of large structures in advance, the necessity of diagnosing structures using non-destructive inspection techniques is being highlighted. The defect occurrence, location and defect type of the structure are important parameters for predicting the remaining life of the structure, so continuous defect observation is very important. Recently, many researchers have been actively researching real-time monitoring technology to solve these problems. Structure Health Monitoring Inspection is a technology that can identify and respond to the occurrence of defects in real time, but there is a limit to check the degree of defects and the direction of growth of defects. In order to compensate for the shortcomings of these technologies, the importance of defect imaging techniques is emerging, and in order to find defects in large structures, a method of inspecting a wide range using guided ultrasonic is effective. The work presented here introduces a calculation for the shape factor for evaluation of the damaged area, as well as a variable β parameter technique to correct a damaged shape. Also, we perform research in modeling simulation and an experiment for comparison with a suggested inspection method and verify its validity. The curved structure image obtained by the advanced RAPID algorithm showed a good match between the defect area and the shape.

• 대한기계학회논문집A
• /
• 제36권10호
• /
• pp.1227-1234
• /
• 2012

본 연구에서는 감육결함이 가공된 실배관 곡관 시편을 대상으로 10 MPa의 내압과 반복 굽힘하중 조건에서 피로 시험을 수행하였다. 시험에는 감육결함이 곡관의 외호부와 내호부에 존재하는 경우를 고려하였으며, 반복 하중으로는 완전 역방향의 변위제어 형태로 작용되는 In-plane 상의 굽힘하중이 고려되었다. 실험 결과, 감육결함이 곡관의 외호부에 존재하는 경우에는 결함부가 아닌 곡관의 측면에서 축방향 균열이 생성되었다. 또한, ASME Sec.III의 설계 피로곡선에서 예측된 건전 곡관의 피로수명보다 긴 피로수명을 보였으며, 피로수명이 결함 길이에 크게 영향을 받지 않았다. 반면, 결함이 내호부에 존재하는 곡관은 내호부에서 원주방향 균열이 발생하였으며, 피로수명은 ASME Sec.III의 설계 피로곡선에서 예측한 건전 곡관의 피로수명보다 짧은 값을 보였다. 또한, 내호부 감육 곡관의 피로수명은 결함 길이가 감소함에 따라 뚜렷이 감소하였다.

• Industrial Engineering and Management Systems
• /
• 제9권3호
• /
• pp.251-261
• /
• 2010

This article demonstrated the application of the Fuzzy Data Envelopment Analytical Hierarchy Process (FDEAHP) to evaluate the root causes of critical defect problems occurring in the production of liquid medicine. The methodology of the research began by collecting the defect data by using Check Sheets, and ranking the significant problems by using a Pareto Diagram. Two types of major problems were found to occur, including glass fragments in the medicine and damaged lid threads. The causes of each problem were then analyzed by using Cause and Effect Diagrams. The significant causes were ranked by FDEAHP under three criteria, Severity (S), Occurrence (O) and Detection (D), followed by the framework of the FMEA Technique. Two causes with the highest Final Weight (FW) of each problem were selected to be improved, such as installing auxiliary equipment, using the Poka-Yoke system, setting the scale of the shaft and lathing the bushes of each bottle size. The results demonstrated a reduction in defects from 3.209% to 1.669% and showed that improving a few significant root causes, identified by an experienced decision maker, was sufficient to reduce the defect rate.

• 한국임상약학회지
• /
• 제29권2호
• /
• pp.133-137
• /
• 2019

Atorvastatin is one of the most widely prescribed medications for dyslipidemia treatment. In Korea, post combined therapy with ezetimibe, a 73-year-old woman was reported by a community pharmacy to have experienced visual field defect, which recovered after drug discontinuation. She had never experienced this symptom before, and several studies have reported an association between use of statins and visual disorders such as blurred vision, diplopia, and cataract. Blockage of cholesterol accumulation, oxidative stress, or myopathy is expected to be a cause of this symptom. Naranjo scale, Korean causality assessment algorithm (Ver.2), and World Health Organization-Uppsala Monitoring Center (WHO-UMC) criteria were the three tools used to determine causality between the visual disorder and atorvastatin. The results represent 'probable', 'certain', and 'probable/likely' causality, respectively. Our results, in combination with a review of literature, indicate that ocular adverse effects are highly likely related to atorvastatin.

• 대한두개안면성형외과학회지
• /
• 제21권6호
• /
• pp.345-350
• /
• 2020

Background: This study evaluated the efficacy of the endoscopic medial orbital wall repair by comparing it with the conventional transcaruncular method. This surgical approach differs from the established endoscopic technique in that we push the mesh inside the orbit rather than placing it over the defect. Methods: We retrospectively reviewed 40 patients with isolated medial orbital blowout fractures who underwent medial orbital wall reconstruction. Twenty-six patients underwent endoscopic repair, and 14 patients underwent external repair. All patients had preoperative computed tomography scans taken to determine the defect size. Pre- and postoperative exophthalmometry, operation time, the existence of diplopia, and pain were evaluated and compared between the two methods. We present a case showing our procedure. Results: The operation time was significantly shorter in the endoscopic group (44.7 minutes vs. 73.9 minutes, p= 0.035). The preoperative defect size, enophthalmos correction rate, and pain did not significantly differ between the two groups. All patients with preoperative diplopia, eyeball movement limitation, or enophthalmos had their symptoms resolved, except for one patient who had preexisting strabismus. Conclusion: This study demonstrates that endoscopic medial orbital wall repair is not inferior to the transcaruncular method. The endoscopic approach seems to reduce the operation time, probably because the dissection process is shorter, and no wound repair is needed. Compared to the previous endoscopic method, our method is not complicated, and is more physiological. Larger scale studies should be performed for validation.

• 문화기술의 융합
• /
• 제9권2호
• /
• pp.569-574
• /
• 2023

콘크리트 구조물의 적절한 유지관리를 위해서 다양한 결함에 대해 사전에 파악하고 유지관리하는 것이 필요하다. 현재 방법으로는 규모가 큰 사회기반시설물의 점검 시 효율성, 안전성, 신뢰성에 문제가 있어 새로운 점검 방식의 도입이 필요하다. 최근에는 영상에 대한 딥러닝 기술이 발달함에 따라 콘크리트 결함 분류 연구가 활발히 진행되고 있다. 하지만, 균열 외에 오염과 박락 등에 대한 연구는 제한적이다. 본 연구에서는 사전에 학습된 딥러닝 모델에 대한 전이학습을 통한 다양한 콘크리트 결함 유형 분류 모델을 개발하고, 정확도를 저하시키는 요인을 도출 및 향후 발전 방향을 제시하였다. 이는 향후 콘크리트 유지관리 분야에서 활용도가 높을 것으로 예상된다.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2001년도 춘계학술대회논문집A
• /
• pp.339-344
• /
• 2001

The transmission pipeline industry spends many millions of dollars annually performing inline inspections, excavating sites of possible corrosion, and repairing or replacing damaged sections of pipe. New criteria for evaluation the integrity of corroded pipe have been developed in recent years to help in controlling these costs. These new criteria vary widely in their estimates of integrity and the most appropriate criterion for a given pipeline is not always clear. This paper presents an overview, comparison and evaluation of acceptability criteria for corrosion defects in pipelines. By full scale burst tests, this paper have assessed the relative accuracy of each of theses criteria in predicting failure and remaining strength. Many of the criteria appear to be excessively conservative and indicate that defects must be repaired when none is needed, based upon burst test data.

• 스마트미디어저널
• /
• 제12권9호
• /
• pp.45-59
• /
• 2023

Weakly supervised object localization (WSOL) is a task of localizing an object in an image using only image-level labels. Previous studies have followed the conventional class activation mapping (CAM) pipeline. However, we reveal the current CAM approach suffers from problems which cause original CAM could not capture the complete defects features. This work utilizes a convolutional neural network (CNN) pretrained on image-level labels to generate class activation maps in a multi-scale manner to highlight discriminative regions. Additionally, a vision transformer (ViT) pretrained was treated to produce multi-head attention maps as an auxiliary detector. By integrating the CNN-based CAMs and attention maps, our approach localizes defective regions without requiring bounding box or pixel-level supervision during training. We evaluate our approach on a dataset of apple images with only image-level labels of defect categories. Experiments demonstrate our proposed method aligns with several Object Detection models performance, hold a promise for improving localization.