• ETRI Journal
• /
• v.46 no.3
• /
• pp.513-525
• /
• 2024

Deep neural networks trained on labeled medical data face major challenges owing to the economic costs of data acquisition through expensive medical imaging devices, expert labor for data annotation, and large datasets to achieve optimal model performance. The heterogeneity of diseases, such as Alzheimer's disease, further complicates deep learning because the test cases may substantially differ from the training data, possibly increasing the rate of false positives. We propose a reconstruction-based self-supervised anomaly detection model to overcome these challenges. It has a dual-subnetwork encoder that enhances feature encoding augmented by skip connections to the decoder for improving the gradient flow. The novel encoder captures local and global features to improve image reconstruction. In addition, we introduce an entropy-based image conversion method. Extensive evaluations show that the proposed model outperforms benchmark models in anomaly detection and classification using an encoder. The supervised and unsupervised models show improved performances when trained with data preprocessed using the proposed image conversion method.

• The Plant Pathology Journal
• /
• v.36 no.1
• /
• pp.54-66
• /
• 2020

This study was conducted to evaluate usefulness of numerical weather prediction data generated by the Unified Model (UM) for plant disease forecast. Using the UM06- and UM18-predicted weather data, which were released at 0600 and 1800 Universal Time Coordinated (UTC), respectively, by the Korea Meteorological Administration (KMA), disease forecast on bacterial grain rot (BGR) of rice was examined as compared with the model output based on the automated weather stations (AWS)-observed weather data. We analyzed performance of BGRcast based on the UM-predicted and the AWS-observed daily minimum temperature and average relative humidity in 2014 and 2015 from 29 locations representing major rice growing areas in Korea using regression analysis and two-way contingency table analysis. Temporal changes in weather conduciveness at two locations in 2014 were also analyzed with regard to daily weather conduciveness (C_i) and the 20-day and 7-day moving averages of C_i for the inoculum build-up phase (C_inc) prior to the panicle emergence of rice plants and the infection phase (C_inf) during the heading stage of rice plants, respectively. Based on C_inc and C_inf, we were able to obtain the same disease warnings at all locations regardless of the sources of weather data. In conclusion, the numerical weather prediction data from KMA could be reliable to apply as input data for plant disease forecast models. Weather prediction data would facilitate applications of weather-driven disease models for better disease management. Crop growers would have better options for disease control including both protective and curative measures when weather prediction data are used for disease warning.

• Herbal Formula Science
• /
• v.15 no.1
• /
• pp.145-161
• /
• 2007

Atopic dermatitis (AD) is a chronic inflammatory skin disease associated with cutaneous hyperreactivity to environmental triggers. The clinical phenotype that characterizes AD is the product of interactions between susceptible genes, the environmental factors, defective skin barrier function, and immunologic responses. This review summarizes recent progress in our understanding of the immunopathophysiology of AD and the implications for mouse models of AD in drug discovery from medicinal plants.

• Korean Journal of Agricultural Science
• /
• v.41 no.1
• /
• pp.23-27
• /
• 2014

Potato late blight occurrence is caused by various environmental factors, and the progress can be regularly predicted so that several predictive models have been developed. The models predict the timing of the disease occurrence, but they do not include the methods of the disease control. Effective fungicide control, economic threshold, prediction models were investigated in the study to reflect on customized control modules for the model forecasting the occurrence of potato late blight.

• Korean Journal of Agricultural Science
• /
• v.41 no.1
• /
• pp.29-34
• /
• 2014

Phytophthora blight occurrence is caused by various environmental factors, and the progress can be regularly predicted so that several predictive models have been developed. The models predict the timing of the disease occurrence, but they do not include the methods of the disease control. Effective fungicide control, control threshold, prediction models were investigated in the study to reflect on customized control modules for the model forecasting the occurrence of Phytophthora blight on hot pepper.

• BMB Reports
• /
• v.44 no.7
• /
• pp.435-445
• /
• 2011

Serous ovarian cancer is one of the most lethal gynecological malignancies. Progress on effective diagnostics and therapeutics for this disease are hampered by ambiguity as to the cellular origins of this histotype of ovarian cancer, as well as limited suitable animal models to analyze early stages of disease. In this report, we will review current animal models with respect to the two proposed progenitor cells for serous ovarian cancer, the ovarian surface epithelium and the fallopian tube epithelium.

• Asian-Australasian Journal of Animal Sciences
• /
• v.23 no.5
• /
• pp.672-679
• /
• 2010

Diabetes mellitus is a worldwide epidemic with high mortality. As concern over this disease rises, the number and value of research grants awarded by the National Research Foundation of Korea (NRF) have increased. Diabetes mellitus is classified into two groups. Type 1 diabetes requires insulin treatment, whereas type 2 diabetes, which is characterized by insulin resistance, can be treated using a variety of therapeutic approaches. Hyperglycemia is thought to be a primary factor in the onset of diabetes, although hyperlipidemia also plays a role. The major organs active in the regulation of blood glucose are the pancreas, liver, skeletal muscle, adipose tissue, intestine, and kidney. Diabetic complications are generally classified as macrovascular (e.g., stroke and heart disease) or microvascular (i.e., diabetic neuropathy, nephropathy, and retinopathy). Several animal models of diabetes have been used to develop oral therapeutic agents, including sulfonylureas, biguanides, thiazolidinediones, acarbose, and miglitol, for both type 1 and type 2 diseases. This review provides an overview of diabetes mellitus, describes oral therapeutic agents for diabetes and their targets, and discusses new developments in diabetic drug research.

• The Korean Journal of Applied Statistics
• /
• v.29 no.1
• /
• pp.145-155
• /
• 2016

In this paper, we compare forecasting models for disease occurrences in elderly patients due to the weather. For the analysis, the medical data of aged patients released from Health Insurance Review and the weather data of the Korea Meteorological Administration are weekly and regionally merged. The ARMAX model, the VARMAX model and the TSCS regression model are considered to analyze the number of weekly occurrences of some diseases attributable to climate conditions. These models are compared with MSE, MAPE, and MAE criteria.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.5
• /
• pp.1225-1239
• /
• 2016

We investigate pediatric tumor incidence data collected by the Florida Association for Pediatric Tumor program using various models commonly used in disease mapping analysis. Particularly, we consider Poisson normal models with various conditional autoregressive structure for spatial dependence, a zero-in ated component to capture excess zero counts and a spatio-temporal model to capture spatial and temporal dependence, together. We found that intrinsic conditional autoregressive model provides the smallest Deviance Information Criterion (DIC) among the models when only spatial dependence is considered. On the other hand, adding an autoregressive structure over time decreases DIC over the model without time dependence component. We adopt weighted ranks squared error loss to identify high risk regions which provides similar results with other researchers who have worked on the same data set (e.g. Zhang et al., 2014; Wang and Rodriguez, 2014). Our results, thus, provide additional statistical support on those identied high risk regions discovered by the other researchers.

• The Plant Pathology Journal
• /
• v.38 no.4
• /
• pp.395-402
• /
• 2022

To predict rice blast, many machine learning methods have been proposed. As the quality and quantity of input data are essential for machine learning techniques, this study develops three artificial neural network (ANN)-based rice blast prediction models by combining two ANN models, the feed-forward neural network (FFNN) and long short-term memory, with diverse input datasets, and compares their performance. The Blast_Weathe long short-term memory r_FFNN model had the highest recall score (66.3%) for rice blast prediction. This model requires two types of input data: blast occurrence data for the last 3 years and weather data (daily maximum temperature, relative humidity, and precipitation) between January and July of the prediction year. This study showed that the performance of an ANN-based disease prediction model was improved by applying suitable machine learning techniques together with the optimization of hyperparameter tuning involving input data. Moreover, we highlight the importance of the systematic collection of long-term disease data.