• Imaging Science in Dentistry
• /
• 제52권3호
• /
• pp.275-281
• /
• 2022

Purpose: The aim of this study was to assess the performance of a deep learning system for permanent tooth germ detection on pediatric panoramic radiographs. Materials and Methods: In total, 4518 anonymized panoramic radiographs of children between 5 and 13 years of age were collected. YOLOv4, a convolutional neural network (CNN)-based object detection model, was used to automatically detect permanent tooth germs. Panoramic images of children processed in LabelImg were trained and tested in the YOLOv4 algorithm. True-positive, false-positive, and false-negative rates were calculated. A confusion matrix was used to evaluate the performance of the model. Results: The YOLOv4 model, which detected permanent tooth germs on pediatric panoramic radiographs, provided an average precision value of 94.16% and an F1 value of 0.90, indicating a high level of significance. The average YOLOv4 inference time was 90 ms. Conclusion: The detection of permanent tooth germs on pediatric panoramic X-rays using a deep learning-based approach may facilitate the early diagnosis of tooth deficiency or supernumerary teeth and help dental practitioners find more accurate treatment options while saving time and effort

• Asian Pacific Journal of Cancer Prevention
• /
• 제14권12호
• /
• pp.7257-7260
• /
• 2013

To determine the pattern of disease in patients presenting with breast lumps and to determine the sensitivity and specificity of fine needle aspiration cytology of benign and malignant diseases as a diagnostic tool by correlating with histopathology findings. This retrospective study was carried out in the Department of Pathology, Maharaja Agrasen Medical College, Agroha, from Jan 2008 to April 2012. Fine needle aspiration cytology was performed on 370 cases and out of these 52 cases were received in the Department for histopathological examination. Fibroadenoma was the most common disease encountered, in 88 (24%), with a peak incidence in second and third decade of life. Fibrocystic disease was second, being common in the third and fourth decades of life. Peak incidences of duct ectasia, granulomatous and tubercular mastitis were seen in the third decade. Gynaecomastia showed two peak incidences in second and sixth decades of life. Out of total 370 cases undergoing fine needle aspiration, benign cases were 316 (85.4%), malignant and suspicious were 54 (14.6%) and 10 (2.70%) respectively. Out of total 22 histological confirmed malignant lesions 19 were interpreted as malignant cytologically while two as suspicious and one as benign. All thirty histologically confirmed benign cases were diagnosed as benign cytologically. The sensitivity, specificity, positive and negative predictive values were 98%, 100%, 100% and 96.4% respectively. FNA cytology is highly accurate for diagnosis of breast masses. However, the clinician should correlate FNA cytological results with physical examination and imaging findings to prevent false negative and false positive events and to obtain optimal management of their patients.

• 한국시뮬레이션학회논문지
• /
• 제17권3호
• /
• pp.9-18
• /
• 2008

단순히 퍼지만을 사용하여 시스템을 연동하는 경우와 퍼지로직을 같이 사용하여 침입 탐지 에이전트의 시스템 성능을 향상시키는 경우에 관한 연구로서, 블랙보드 기반의 비퍼지로직을 사용하는 경우와 블랙보드 기반의 퍼지 로직을 사용하는 경우을 비교한다. 또한 BBA를 통해 정적으로 대응하던 시스템을 향상시켜 동적 대응이 가능하게 구성하여 현실적인 시스템이 되도록 구성하였다. 대상 시스템의 성능을 평가하기 위하여 시뮬레이션을 수행하였다. 퍼지 시스템을 사용함으로써 false negative를 줄일 수 있었다. 분산 침입탐지를 위해 포함된 퍼지로직은 다양한 요소를 고려하기 때문에 침입의 성능을 높일 수 있다. 퍼지시스템을 사용하는 경우와 비 퍼지 시스템의 성능을 비교함으로써 퍼지 시스템의 성능 향상을 보이며, 이러한 비교를 통해 전체 시스템의 성능 향상을 보인다.

• Clinical and Experimental Reproductive Medicine
• /
• 제32권3호
• /
• pp.223-230
• /
• 2005

Objective: This study was performed to evaluate the accuracy of hysterosalpingography (HSG) for evaluating female infertility patients by comparison with hysteroscopic and laparoscopic examination. Methods and Material: Total 219 infertile patients were retrospectively analyzed between January 1, 2002 and December 31, 2003. Ninety seven patients (44.3%) were primary infertility, 122 patients (55.7%) were secondary infertility. We performed hysteroscopic and laparoscopic examination on next cycle when HSG revealed any abnormal finding, and 3~6 cycles later if HSG was normal. Results: The accuracy of HSG was 65.2% compared with hysteroscopic examination (sensitivity 88.4%, specificity 46.4%, false positive rate 53.6%, false negative rate 11.6%). The most common abnormal finding of hysteroscopy was uterine synechia (67.4%) followed by endometrial polyp, uterine anomaly (e.g. uterine septum), endometrial hyperplasia. Compared with laparoscopic examination, the accuracy of HSG was 76.9% (sensitivity 98.9%, specificity 70.6%, +LR 3.36, -LR 0.02). The positive predictive value of normal patent tube was excellent (99.6%) but that of proximal tubal blockage was only 46.7%. The unilateral tubal obstruction of HSG was poor accuracy (+LR 3.85 -LR 0.68) and 70% of those was patent by laparoscopic examination. Laparoscopic examination also revealed that 53% of patients had peritubal adhesion and 37% of patients has additional pelvic findings, especially endometriosis. Among the patients had normal HSG, 53.5% patients with normal ultrasonography was diagnosed endometriosis (25.6% of them had endometriosis stage I-II). Conclusion: Normal HSG shows a high negative predictive value. Nevertheless, the incidence of associated pelvic disease in the normal HSG group is high enough to warrant diagnostic laparoscopy if nonsurgical treatment is unsuccessful. Because HSG has poor accuracy in predicting distal tubal blockage and peritubal adhesion, and poor positive predictive value of proximal tubal blockage, laparoscopic examination could be considered in abnormal HSG group.

• Journal of Trauma and Injury
• /
• 제31권3호
• /
• pp.151-158
• /
• 2018

Purpose: The regional emergency medical centers manage the patients with major blunt trauma according to the process appropriate to each hospital rather than standardized protocol of the major trauma centers. The primary purpose of this study is to evaluate the effectiveness and influence on prognosis of additional cervical-thoracic-lumbar-spine computed tomography (CTL-spine CT) scan in diagnosis of spinal injury from the victim of major blunt trauma with impaired consciousness. Methods: The study included patients visited the urban emergency medical center with major blunt trauma who were over 18 years of age from January 2013 to December 2016. Data were collected from retrospective review of medical records. Sensitivity, specificity, positive predictive value, and negative predictive value were measured for evaluation of the performance of diagnostic methods. Results: One hundred patients with Glasgow coma scale ${\leq}13$ underwent additional CTL-spine CT scan. Mechanism of injury was in the following order: driver, pedestrian traffic accident, fall and passenger accident. Thirty-one patients were diagnosed of spinal injury, six of them underwent surgical management. The sensitivity of chest, abdomen and pelvis CT (CAP CT) was 72%, specificity 97%, false positive rate 3%, false negative rate 28% and diagnostic accuracy 87%. Eleven patients were not diagnosed of spinal injury with CAP CT and C-spine lateral view, but all of them were diagnosed of stable fractures. Conclusions: C-spine CT scan be actively considered in the initial examination process. When CAP CT scan is performed in major blunt trauma patients with impaired consciousness, CTL-spine CT scan or simple spinal radiography has no significant effect on the prognosis of the patient and can be performed if necessary.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제9권12호
• /
• pp.411-418
• /
• 2020

정보보안을 위한 IDS(Intrusion Detection Systems)는 통상적으로 서명기반(signature based) 침입탐지시스템과 이상기반(anomaly-based) 침입 탐지시스템으로 분류한다. 이 중에서도 네트워크에서 발생하는 트래픽 데이터를 기계학습으로 분석하는 이상기반 IDS 연구가 활발하게 진행됐다. 본 논문에서는 공격 유형 학습에 사용되는 데이터에 존재하는 희소 클래스 문제로 인한 성능 저하를 해결하기 위한 전처리 방안에 대해 연구했다. 희소 클래스(Rare Class)와 준 희소 클래스(Semi Rare Class)를 기준으로 데이터를 재구성하여 기계학습의 분류 성능의 개선에 대하여 실험했다. 재구성된 3종의 데이터 세트에 대하여 Wrapper와 Filter 방식을 연이어 적용하는 하이브리드 특징 선택을 수행한 이후에 Quantile Scaler로 정규화를 처리하여 전처리를 완료한다. 준비된 데이터는 DNN(Deep Neural Network) 모델로 학습한 후 TP(True Positive)와 FN(False Negative)를 기준으로 분류 성능을 평가했다. 이 연구를 통해 3종류의 데이터 세트에서 분류 성능이 모두 개선되는 결과를 얻었다.

• 한국프랜차이즈경영연구
• /
• 제12권2호
• /
• pp.35-49
• /
• 2021

Purpose: The major objective of this study is to develop a model for the impact of franchisors' Gapjil (verbal·nonverbal Gapjil, abusing bargaining position, refusing transaction, false or exaggerated information, restrictive practices, unfair damage compensation) on franchisee's recontract intention. We also examine the mediating role of economic satisfaction and social satisfaction in the relationship between franchisors' Gapjil and franchisee's contract intention. Research design, data, and methodology: Data were collected from franchisee owners located nationwide in Korea. Out of 256 questionaires distributed, a total of 256 questionnaires were returned. After excluding 10 invalid respondent questionnaires, we coded and analyzed 246 valid questionnaires (effective response rate of 96.09%) using frequency, confirmatory factor analysis, correlations analysis, and structural equation modeling with SPSS 22.O and SmartPLS 3.0. Results: The findings of this study are summarized as follows: First, among the Gapjil of the franchisors, restrictive practices and unfair damage compensation had negative effects on economic and social satisfaction, but verbal and nonverbal Gapjil for economic and social satisfaction was not significant. Second, abusing bargaining positions and false or exaggerated information had negative effects on social satisfaction, but for economic satisfaction, found to be insignificant. Third, economic and social satisfaction had positive effects on the franchisee's recontract intention to the franchisor. Conclusion: The following implications of this study are as follows. First, the construct of Gapjil that occurs between the franchisors and the franchisees was first presented, and the franchisors' Gapjil is divided into interpersonal Gapjil and structural Gapjil. Second, the Gapjil of the franchisors can be an important predictor variable in maintaining and developing a long-term relationship between the franchisors and the franchisees. Third, solving conflict due to the Gapjil problem between franchisors and franchisees can be an important factor for franchisors and franchisees to co-survive and thrive in Korean franchise system. Fourth, this study suggest that managing the Gapjil of the franchisors was a important antecedent factor in maintaining long-term relationship between the franchisors and the franchisees. Therefore, this study will help franchisors formulate effective symbiotic marketing strategies to satisfy relationships with franchisees and consequently enhance long-term orientation.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권2호
• /
• pp.742-762
• /
• 2015

In this study, we propose a distance metric learning approach called discriminant metric learning (DML) for face verification, which addresses a binary-class problem for classifying whether or not two input images are of the same subject. The critical issue for solving this problem is determining the method to be used for measuring the distance between two images. Among various methods, the large margin nearest neighbor (LMNN) method is a state-of-the-art algorithm. However, to compensate the LMNN's entangled data distribution due to high levels of appearance variations in unconstrained environments, DML's goal is to penalize violations of the negative pair distance relationship, i.e., the images with different labels, while being integrated with LMNN to model the distance relation between positive pairs, i.e., the images with the same label. The likelihoods of the input images, estimated using DML and LMNN metrics, are then weighted and combined for further analysis. Additionally, rather than using the k-nearest neighbor (k-NN) classification mechanism, we propose a verification mechanism that measures the correlation of the class label distribution of neighbors to reduce the false negative rate of positive pairs. From the experimental results, we see that DML can modify the relation of negative pairs in the original LMNN space and compensate for LMNN's performance on faces with large variances, such as pose and expression.

• Asian Pacific Journal of Cancer Prevention
• /
• 제17권8호
• /
• pp.4071-4075
• /
• 2016

Background: FDG PET/CT is at an equivocal stage to recommend for staging of colorectal cancer as compared to contrast-enhanced CT (ceCT). This study was intended to evaluate the value of FDG PET/ceCT in colorectal cancer staging as compared to ceCT alone. Materials and Methods: PET/ceCT was performed for 61 colorectal cancer patients who were prospectively enrolled in the study. Three patients were excluded due to loss to follow-up. PET/ceCT findings and ceCT results alone were read separately. The treatment planning was then determined by tumor board consensus. The criteria for T staging were determined by the findings of ceCT. Nodal positive by PET/ceCT imaging was determined by visual analysis of FDG uptake greater than regional background blood pool activity. The diagnostic accuracy of T and N staging was determined only in patients who received surgery without any neoadjuvant treatment. Results: Of 58 patients, there were 40 with colon cancers including sigmoid cancers and 18 with rectal cancers. PET/ceCT in pre-operative staging detected bone metastasis and metastatic inguinal lymph nodes (M1a) that were undepicted on CT in 2 patients (3%), clearly defined 19 equivocal lesions on ceCT in 18 patients (31%) and excluded 6 metastatic lesions diagnosed by ceCT in 6 patients (10%). These resulted in alteration of management plan in 15 out of the 58 cases (26%) i.e. changing from chemotherapy to surgery (4), changing extent of surgery (9) and avoidance of futile surgery (2). Forty four patients underwent surgery within 45 days after PET/CT. The diagnostic accuracy for N staging with PET/ceCT and ceCT alone was 66% and 48% with false positive rates of 24% (6/25) and 76% (19/25) and false negative rates of 47% (9/19) and 21% (4/19), respectively. All of the false negative lymph nodes from PET/ceCT were less than a centimeter in size and located in peri-lesional regions. The diagnostic accuracy for T staging was 82%. The sensitivity of the peri-lesional fat stranding sign in determining T3 stage was 94% and the specificity was 54%. Conclusions: Our study suggested promising roles of PET/ceCT in initial staging of colorectal cancer with better diagnostic accuracy facilitating management planning.

• Journal of the Korean Association of Oral and Maxillofacial Surgeons
• /
• 제32권2호
• /
• pp.151-156
• /
• 2006

술전 경부 임파절 평가의 정확성은 신체검사와 방사선학적 소견 모두에서 정확성의 한계를 보였으며, 정확성을 보다 높이기 위해서는 임상의사와 방사선의사간에 의견 교환이 중요하다고 사료된다. 그리고 필요에 따라서는 초음파 촬영, 세침 흡인세포검사(FNAC) 등의 보조적인 임파절 평가 방법을 이용하는 것도 전체 정진율을 높일 수 있을 것이다. 또한 구강암의 원발 부위가 고위험군인 경우에는 가음성도 비율이 높고, 방사선학적 검사에서 전체 정진율이 낮은 결과를 보이는 것으로 볼때, 지금까지 알려진대로 예방적 경부 곽청술을 보다 적극적으로 고려하는 것이 필요하다고 사료된다.