• Journal of the Korea Society of Computer and Information
• /
• v.26 no.4
• /
• pp.21-28
• /
• 2021

This study proposes a deep learning-based flat foot classification methodology using transfer learning. We used a transfer learning with VGG16 pre-trained model and a data augmentation technique to generate a model with high predictive accuracy from a total of 176 image data consisting of 88 flat feet and 88 normal feet. To evaluate the performance of the proposed model, we performed an experiment comparing the prediction accuracy of the basic CNN-based model and the prediction model derived through the proposed methodology. In the case of the basic CNN model, the training accuracy was 77.27%, the validation accuracy was 61.36%, and the test accuracy was 59.09%. Meanwhile, in the case of our proposed model, the training accuracy was 94.32%, the validation accuracy was 86.36%, and the test accuracy was 84.09%, indicating that the accuracy of our model was significantly higher than that of the basic CNN model.

• International Journal of Computer Science & Network Security
• /
• v.22 no.5
• /
• pp.175-181
• /
• 2022

COVID-19 is an acute respiratory syndrome that affects the host's breathing and respiratory system. The novel disease's first case was reported in 2019 and has created a state of emergency in the whole world and declared a global pandemic within months after the first case. The disease created elements of socioeconomic crisis globally. The emergency has made it imperative for professionals to take the necessary measures to make early diagnoses of the disease. The conventional diagnosis for COVID-19 is through Polymerase Chain Reaction (PCR) testing. However, in a lot of rural societies, these tests are not available or take a lot of time to provide results. Hence, we propose a COVID-19 classification system by means of machine learning and transfer learning models. The proposed approach identifies individuals with COVID-19 and distinguishes them from those who are healthy with the help of Deep Visual Embeddings (DVE). Five state-of-the-art models: VGG-19, ResNet50, Inceptionv3, MobileNetv3, and EfficientNetB7, were used in this study along with five different pooling schemes to perform deep feature extraction. In addition, the features are normalized using standard scaling, and 4-fold cross-validation is used to validate the performance over multiple versions of the validation data. The best results of 88.86% UAR, 88.27% Specificity, 89.44% Sensitivity, 88.62% Accuracy, 89.06% Precision, and 87.52% F1-score were obtained using ResNet-50 with Average Pooling and Logistic regression with class weight as the classifier.

• The korean journal of orthodontics
• /
• v.52 no.2
• /
• pp.112-122
• /
• 2022

Objective: This study aimed to present and evaluate a new deep learning model for determining cervical vertebral maturation (CVM) degree and growth spurts by analyzing lateral cephalometric radiographs. Methods: The study sample included 890 cephalograms. The images were classified into six cervical stages independently by two orthodontists. The images were also categorized into three degrees on the basis of the growth spurt: pre-pubertal, growth spurt, and post-pubertal. Subsequently, the samples were fed to a transfer learning model implemented using the Python programming language and PyTorch library. In the last step, the test set of cephalograms was randomly coded and provided to two new orthodontists in order to compare their diagnosis to the artificial intelligence (AI) model's performance using weighted kappa and Cohen's kappa statistical analyses. Results: The model's validation and test accuracy for the six-class CVM diagnosis were 62.63% and 61.62%, respectively. Moreover, the model's validation and test accuracy for the three-class classification were 75.76% and 82.83%, respectively. Furthermore, substantial agreements were observed between the two orthodontists as well as one of them and the AI model. Conclusions: The newly developed AI model had reasonable accuracy in detecting the CVM stage and high reliability in detecting the pubertal stage. However, its accuracy was still less than that of human observers. With further improvements in data quality, this model should be able to provide practical assistance to practicing dentists in the future.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.22 no.4
• /
• pp.239-249
• /
• 2020

Rice flour varieties have been developed to replace wheat, and consumption of rice flour has been encouraged. damage related to pre-harvest sprouting was occurring due to a weather disaster during the ripening period. Thus, it is necessary to develop pre-harvest sprouting rate prediction system to minimize damage for pre-harvest sprouting. Rice cultivation experiments from 20 17 to 20 19 were conducted with three rice flour varieties at six regions in Gangwon-do, Chungcheongbuk-do, and Gyeongsangbuk-do. Survey components were the heading date and pre-harvest sprouting at the harvest date. The weather data were collected daily mean temperature, relative humidity, and rainfall using Automated Synoptic Observing System (ASOS) with the same region name. Gradient Boosting Machine (GBM) which is a machine learning model, was used to predict the pre-harvest sprouting rate, and the training input variables were mean temperature, relative humidity, and total rainfall. Also, the experiment for the period from days after the heading date (DAH) to the subsequent period (DA2H) was conducted to establish the period related to pre-harvest sprouting. The data were divided into training-set and vali-set for calibration of period related to pre-harvest sprouting, and test-set for validation. The result for training-set and vali-set showed the highest score for a period of 22 DAH and 24 DA2H. The result for test-set tended to overpredict pre-harvest sprouting rate on a section smaller than 3.0 %. However, the result showed a high prediction performance (R2=0.76). Therefore, it is expected that the pre-harvest sprouting rate could be able to easily predict with weather components for a specific period using machine learning.

• Korean Journal of Radiology
• /
• v.24 no.6
• /
• pp.541-552
• /
• 2023

Objective: Detection of pneumoperitoneum using abdominal radiography, particularly in the supine position, is often challenging. This study aimed to develop and externally validate a deep learning model for the detection of pneumoperitoneum using supine and erect abdominal radiography. Materials and Methods: A model that can utilize "pneumoperitoneum" and "non-pneumoperitoneum" classes was developed through knowledge distillation. To train the proposed model with limited training data and weak labels, it was trained using a recently proposed semi-supervised learning method called distillation for self-supervised and self-train learning (DISTL), which leverages the Vision Transformer. The proposed model was first pre-trained with chest radiographs to utilize common knowledge between modalities, fine-tuned, and self-trained on labeled and unlabeled abdominal radiographs. The proposed model was trained using data from supine and erect abdominal radiographs. In total, 191212 chest radiographs (CheXpert data) were used for pre-training, and 5518 labeled and 16671 unlabeled abdominal radiographs were used for fine-tuning and self-supervised learning, respectively. The proposed model was internally validated on 389 abdominal radiographs and externally validated on 475 and 798 abdominal radiographs from the two institutions. We evaluated the performance in diagnosing pneumoperitoneum using the area under the receiver operating characteristic curve (AUC) and compared it with that of radiologists. Results: In the internal validation, the proposed model had an AUC, sensitivity, and specificity of 0.881, 85.4%, and 73.3% and 0.968, 91.1, and 95.0 for supine and erect positions, respectively. In the external validation at the two institutions, the AUCs were 0.835 and 0.852 for the supine position and 0.909 and 0.944 for the erect position. In the reader study, the readers' performances improved with the assistance of the proposed model. Conclusion: The proposed model trained with the DISTL method can accurately detect pneumoperitoneum on abdominal radiography in both the supine and erect positions.

• Journal of Korean Academy of Oral and Maxillofacial Radiology
• /
• v.28 no.2
• /
• pp.363-385
• /
• 1998

In orthodontics and orthognathic surgery. cephalogram has been routine practice in diagnosis and treatment evaluation of craniofacial deformity. But its inherent distortion of actual length and angles during projecting three dimensional object to two dimensional plane might cause errors in quantitative analysis of shape and size. Therefore, it is desirable that three dimensional object is diagnosed and evaluated three dimensionally and three dimensional CT image is best for three dimensional analysis. Development of clinic necessitates evaluation of result of treatment and comparison before and after surgery. It is desirable that patient that was diagnosed and planned by three dimensional computed tomography before surgery is evaluated by three dimensional computed tomography after surgery. too. But Because there is no standardized normal values in three dimension now and three dimensional Computed Tomography needs expensive equipments and because of its expenses and amount of exposure to radiation. limitations still remain to be solved in its application to routine practice. If postoperative three dimensional image is constructed by pre and postoperative lateral and postero-anterior cephalograms and preoperative three dimensional computed tomogram. pre and postoperative image will be compared and evaluated three dimensionally without three dimensional computed tomography after surgery and that will contribute to standardize normal values in three dimension. This study introduced new method that computer-simulated three dimensional image was constructed by preoperative three dimensional computed tomogram and pre and postoperative lateral and postero-anterior cephalograms. and for validation of new method. in four cases of dry skull that position of mandible was displaced and four patients of orthognathic surgery. computer-simulated three dimensional image and actual postoperative three dimensional image were compared. The results were as follows. 1. In four cases of dry skull that position of mandible was displaced. range of displacement between computer-simulated three dimensional images and actual postoperative three dimensional images in co-ordinates values was from -1.8 mm to 1.8 mm and 94％ in displacement of all co-ordinates values was from -1.0 mm to 1.0 mm and no significant difference between computer-simulated three dimensional images and actual postoperative three dimensional images was noticed(p>0.05). 2. In four cases of orthognathic surgery patients, range of displacement between computer­simulated three dimensional images and actual postoperative three dimensional images in coordinates values was from -6.7 mm to 7.7 mm and 90％ in displacement of all co-ordinates values was from -4.0 to 4.0 mm and no significant difference between computer-simulated three dimensional images and actual postoperative three dimensional images was noticed(p>0.05). Conclusively. computer-simulated three dimensional image was constructed by preoperative three dimensional computed tomogram and pre and postoperative lateral and postero-anterior cephalograms. Therefore. potentiality that can construct postoperative three dimensional image without three dimensional computed tomography after surgery was presented.

• Journal of the Korea Concrete Institute
• /
• v.28 no.6
• /
• pp.685-693
• /
• 2016

The outer tank of a liquefied natural gas (LNG) storage tank is a longitudinally and meridionally pre-stressed concrete (PSC) wall structure. Because of the current trend of constructing larger LNG storage tanks, the pre-stressing forces required to increase wall strength must be significantly increased. Because of the increase in tank sizes and pre-stressing forces, an extreme loading scenario such as a bomb blast or an airplane crash needs to be investigated. Therefore, in this study, the blast resistance performance of LNG storage tanks was analyzed by conducting a blast simulation to investigate the safety of larger LNG storage tanks. Test data validation for a blast simulation of reinforced concrete panels was performed using a specific FEM code, LS-DYNA, prior to a full-scale blast simulation of the outer tank of a 270,000-kL LNG storage tank. Another objective of this study was to evaluate the safety and serviceability of an LNG storage tank with respect to varying amounts of explosive charge. The results of this study can be used as basic data for the design and safety evaluation of PSC LNG storage tanks.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1512-1512
• /
• 2001

The University of Cordoba conducts since 1991 a breeding program to obtain new olive cultivars from intraspecific crosses. The objective is to obtain new early bearing and high-quality cultivars. In plant breeding, many seedlings must be tested to increased the chance of getting desirable genotypes. Therefore, fast, cheap and accurate methods of analysis are necessary. The conventional laboratory techniques are costly and time-consuming. Near Infrared Spectroscopy (NIRS) can satisfy the characteristics requested by plant breeders and offers many advantages such as the simultaneous analysis of many traits and cheap cost. The objective of this work was to asses the performance of NIRS to estimate oil fruit components (fruit weight, flesh moisture, flesh/stone ratio and oil flesh content in dry weight basis) and fatty acid composition in olive fruit. Genotypes from reciprocal crosses between ‘Arbequina’, ‘Frantoio’ and ‘Picual’ cultivars have been used in this study. A total of 287 samples, each from a single plant, were scanned using a DA-7000 Diode Array VIS/NIR Analysis System (Perten Instruments), which covers the visible and NIR range from 400-1700 nm. All samples were analysed for fatty acid composition (gas chromatography) and 220 for oil fruit components (oil content by nuclear magnetic resonance), 70% and 30% of samples were randomly assign for the calibration and validation sets respectively. The preliminary results shows that calibration for palmitic, oleic and linoleic acids were highly accurate with calibration and validation values of $r^2$ from 0.85 to 0.95 and 0.76 to 0.91 respectively. Calibration for palmitoleic and estearic acids were less accurate, probably because of the narrow range of variability available for these fatty acids. For the oil fruit components, calibration were high accurate for flesh moisture and oil flesh content in dry weight basis ($r^2$ higher than 0.90 in both calibration and validation sets) and less accurate for the other characteristics evaluated. The first results obtained indicate that NIRS analysis could be an ideal technique to reduce the cost, time and chemical wasted necessary to evaluate a large number of genotypes and it is accurate enough to use for pre-selecting genotypes in a breeding program.

• International Journal of Highway Engineering
• /
• v.20 no.3
• /
• pp.85-92
• /
• 2018

PURPOSES : In this study, analyze the characteristics of IOC indicator 'threshold' which is needed when evaluating the traffic signal operation status with ESPRESSO in various grade road traffic environment of Seoul metropolitan city and derive suggested value to use in field practice. METHODS : Using the computerized database program (Postgresql), we extracted data with regional characteristics (Arterial, Collector road) and temporal characteristics (peak hour, non-peak hour). Analysis of variance and Duncan's validation were performed using statistical analysis program (SPSS) to confirm whether the extracted data contains statistical significance. RESULTS : The analysis period of the main and secondary arterial roads was confirmed to be suitable from 14 days to 60 days. For the arterial, it is suggested to use 20 km/h as the critical speed for PM peak hour and weekly non peak hour. It is suggested to use 25 km/h as the critical speed for AM peak hour and night non peak hour. As for the collector road, it is suggested to use 20 km/h as the critical speed for PM peak hour and weekly non peak hour. It is suggested to use 30 km/h as the critical speed for AM peak hour and night non peak hour. CONCLUSIONS : It is meaningful from a methodological point of view that it is possible to make a reasonable comparative analysis on the signal intersection pre-post analysis when the signal operation DB is renewed by breaking the existing traffic signal operation evaluation method.

• Journal of the Korean Regional Science Association
• /
• v.31 no.4
• /
• pp.91-105
• /
• 2015

Mortgage-Backed Securities (MBS) was introduced in 1999 in order to stabilize housing market and prevent potential speculation. However, research on MBS is limited, so this paper try to narrow the gap by focusing on the factors relating the pre-payment risk of MBS. We used Granger Causality Validation, Vector Auto Regressive, and HP-filtering with time-series data from 2004 to 2014. This paper shows that the prepayment rate of MBS increases as Mortgage rate decreases because borrowers tend to refinance existing MBS with new lower-rate MBS. In addition, it reveals that the rate increases as housing price increases. This outcome support the hypothesis that introduction of low-rate MBS invites more investment or speculation, and hence the housing price rises. The relationship between the MBS pre-payment rate and housing price is yet a peculiar characteristic of the MBS in Korea.