• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.46 no.4
• /
• pp.312-320
• /
• 2023

Recently, there has been an increasing attempt to replace defect detection inspections in the manufacturing industry using deep learning techniques. However, obtaining substantial high-quality labeled data to enhance the performance of deep learning models entails economic and temporal constraints. As a solution for this problem, semi-supervised learning, using a limited amount of labeled data, has been gaining traction. This study assesses the effectiveness of semi-supervised learning in the defect detection process of manufacturing using the MixMatch algorithm. The MixMatch algorithm incorporates three dominant paradigms in the semi-supervised field: Consistency regularization, Entropy minimization, and Generic regularization. The performance of semi-supervised learning based on the MixMatch algorithm was compared with that of supervised learning using defect image data from the metal casting process. For the experiments, the ratio of labeled data was adjusted to 5%, 10%, 25%, and 50% of the total data. At a labeled data ratio of 5%, semi-supervised learning achieved a classification accuracy of 90.19%, outperforming supervised learning by approximately 22%p. At a 10% ratio, it surpassed supervised learning by around 8%p, achieving a 92.89% accuracy. These results demonstrate that semi-supervised learning can achieve significant outcomes even with a very limited amount of labeled data, suggesting its invaluable application in real-world research and industrial settings where labeled data is limited.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2007.08a
• /
• pp.286-289
• /
• 2007

We have developed an automated circuit defect inspection and repair method that can be used to improve the yield ratio of TFT-LCD. The method focuses on correcting resist patterns after the development process to ensure shape regularity. We built a prototype system and confirmed that the method is valid.

• Archives of Plastic Surgery
• /
• v.34 no.1
• /
• pp.44-51
• /
• 2007

Purpose: Blow-out fracture is one of the most common fractures in facial trauma. It is diagnosed by Computed Tomography(CT) scan, which is considered as the most effective diagnostic tool. Since, the Picture Archiving Communication System(PACS) has been provided recently to many hospitals, doctors are more familiar with imaging software of PACS. Because this software has many useful measuring tools, doctors can measure orbital structure easily and make a plan for treatment with its data. Therefore, authors intended to analyze the data of orbital structure measured with PACS imaging software and evaluate its usefulness. Methods: The charts and CT images of 100 patients, which were 50 patients with medial wall fracture and 50 patients with floor fracture, were reviewed. Patients were selected by pre-determined criteria and their CT images were measured with image software of PACS. 'Extraocular muscle thickness', 'Defect ratio'(ratio of defect area to normal area) and 'Globe position index' were measured and analyzed statistically. Results: The thickness of inferior rectus muscle and medial rectus muscle was simultaneously increased in acute-stage of blow-out fracture. The medial rectus muscle was more thickened in medial wall fracture and inferior rectus was more thickened in floor fracture, respectively. In acute blow-out fracture, globe position is exophthalmic rather than enophthalmic. Especially in floor fracture, numerical value summed up thickness of all extraocular muscle is correlated to the defect ratio and globe position index. Conclusion: Clinicians can decide globe position or presume defect ratio in inferior wall fracture by measurement of CT image in acute blow-out fracture using PACS.

• Proceedings of the Korean Ceranic Society Conference
• /
• 2002.10a
• /
• pp.98.2-98
• /
• 2002

• 전자공학회논문지 IE
• /
• v.48 no.4
• /
• pp.1-6
• /
• 2011

In this paper we show the LCD simulator for defect inspection using image processing algorithm and neural network. The defect inspection algorithm of the LCD consists of preprocessing, feature extraction and defect classification. Preprocess removes noise from LCD image, using morphology operator and neural network is used for the defect classification. Sample images with scratch, pinhole, and spot from real LCD color filter image are used. From some experiments results, the proposed algorithms show that defect detected and classified in the ratio of 92.3% and 94.5 respectively. Accordingly, in this paper, a possibility of practical implementation of the LCD defect inspection system is finally suggested.

• 전자공학회논문지 IE
• /
• v.48 no.2
• /
• pp.26-31
• /
• 2011

In this paper we show the LCD simulator for defect inspection using image processing algorithm and neural network. The defect inspection algorithm of the LCD consists of preprocessing, feature extraction and defect classification. Preprocess removes noise from LCD image, using morphology operator and neural network is used for the defect classification. Sample images with scratch, pinhole, and spot from real LCD color filter image are used. From some experiments results, the proposed algorithms show that defect detected and classified in the ratio of 92.3% and 94.5 respectively. Accordingly, in this paper, a possibility of practical implementation of the LCD defect inspection system is finally suggested.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2018.11a
• /
• pp.75-76
• /
• 2018

MiDF, Micro Defect Free concrete, is the concrete made for high strength development by minimizing the defects such as micro pore and ITZ. Since MiDF exhibits strength through hydro-thermal synthesis reaction, it is essential to select the optimum C/S mole ratio. In this study, the basic characteristics of MiDF were evaluated by controlling the C/S mole ratio by SF and QP.

• Journal of Chest Surgery
• /
• v.17 no.4
• /
• pp.607-613
• /
• 1984

Thirty seven patients with ostium secundum atrial septal defect, operated from January, 1976 to September, 1984 at the Department of Thoracic and Cardiovascular Surgery, Kyungpook National University Hospital, were given clinical assessment. The following results were obtained. 1.Ostium secundum atrial septal defect was comprised of 18% of congenital heart disease. Their mean age was 15.7\ulcorner.42. Sex ratio [male:female] was 1:1.1. 2.Most frequent clinical symptom was dyspnea on exertion occurred in 26 patients [76.5%]. Only one patient had no symptom [2.9%]. 3.Pre-operative EKG findings revealed RVH in 61.8%, ICRBBB in 29.4%, and RAD in 41.2%. 4.Mean value of systolic pulmonary arterial pressure in patients over 20 years old was 37.8\ulcorner4.4mmHg and it was 28.1\ulcorner10.2mmHg in patients under 20 years old, but the difference between two groups was not statistically significant. 5. In large defect group [>3cm in diameter], Qp/Qs was significantly increased than small defect group [<3cm in diameter], but systolic pulmonary arterial pressure and Rp/Rs were not different between two groups. 6. Overall operative mortality was 5.4%.

• Journal of Chest Surgery
• /
• v.20 no.2
• /
• pp.411-415
• /
• 1987

We have experienced a rare case of traumatic ventricular septal defect caused by fall down injury. The patient suffered from congestive heart failure after trauma and represented new developing typical murmur. Cardiac catheterization revealed a large left-to-right shunt at the ventricular level and a 4.4:1 ratio of pulmonary to systemic flow. Left ventricular angiogram also showed ventricular septal defect low in the muscular septum. At operation, the right ventricle and pulmonary artery were found to be moderately dilated. A marked systolic thrill was palpable over the right ventricle near the apex. Through a small left ventriculotomy an oval defect, 2x.3 cm with a firm scarred margin, was found in the lower muscular ventricular septum. This defect was repaired using Dacron patch and ventriculotomy was closed with long strip of Teflon felt. Postoperative course was uneventful and discharged in good condition.

• Journal of Mechanical Science and Technology
• /
• v.14 no.5
• /
• pp.530-536
• /
• 2000

In this paper, the influence of several factors such as hardness, internal defect and non-propagating crack on fatigue limits was investigated with three kinds of ductile iron specimens. From the experimental results the fatigue limits were examined in relation with hardness and tensile strength in case of high strength specimens under austempering treatment; in consequence the marked improvement of fatigue limits were not showed. The maximum defect size was an important factor to predict and to evaluate the fatigue limits of ductile irons. And, the quantitative relationship between the fatigue limits$({\sigma}_w)$ and the maximum defect sizes $(\sqrt{area}_{max})$ was expressed as ${\sigma}_w^n{\cdot}{\sqrt{area}}_{max}=C_2$. Also, it was possible to explain the difference for the fatigue limits in three ductile irons by introduction of the non-propagating crack rates.