• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.10
• /
• pp.397-402
• /
• 2019

Accurately predicting cancer prognosis to provide appropriate treatment strategies for patients is one of the critical challenges in bioinformatics. Many researches have suggested machine learning models to predict patients' outcomes based on their gene expression data. Gene expression data is high-dimensional numerical data containing about 17,000 genes, so traditional researches used feature selection or dimensionality reduction approaches to elevate the performance of prognostic prediction models. These approaches, however, have an issue of making it difficult for the predictive models to grasp any biological interaction between the selected genes because feature selection and model training stages are performed independently. In this paper, we propose a novel two-dimensional image formatting approach for gene expression data to achieve feature selection and prognostic prediction effectively. Node2Vec is exploited to integrate biological interaction network and gene expression data and a convolutional neural network learns the integrated two-dimensional gene expression image data and predicts cancer prognosis. We evaluated our proposed model through double cross-validation and confirmed superior prognostic prediction accuracy to traditional machine learning models based on raw gene expression data. As our proposed approach is able to improve prediction models without loss of information caused by feature selection steps, we expect this will contribute to development of personalized medicine.

• Journal of Internet Computing and Services
• /
• v.20 no.5
• /
• pp.105-111
• /
• 2019

Core technologies that lead the Fourth Industrial Revolution era, such as artificial intelligence, big data, and autonomous driving, are implemented and serviced through the rapid development of computing power and hyper-connected networks based on the Internet of Things. In this paper, we implement two different models for drivable area segmentation in various environment, and propose a better model by comparing the results. The models for drivable area segmentation are using DeepLab V3+ and Mask R-CNN, which have great performances in the field of image segmentation and are used in many studies in autonomous driving technology. For driving information in various environment, we use BDD dataset which provides driving videos and images in various weather conditions and day&night time. The result of two different models shows that Mask R-CNN has higher performance with 68.33% IoU than DeepLab V3+ with 48.97% IoU. In addition, the result of visual inspection of drivable area segmentation on driving image, the accuracy of Mask R-CNN is 83% and DeepLab V3+ is 69%. It indicates Mask R-CNN is more efficient than DeepLab V3+ in drivable area segmentation.

• Journal of the Korean Society of Radiology
• /
• v.14 no.7
• /
• pp.991-1001
• /
• 2020

Supervised deep learning technologies for improving the image quality of computed tomography (CT) need a lot of training data. When input images have different characteristics with training images, the technologies cause structural distortion in output images. In this study, an imaging model based on the deep reinforcement learning (DRL) was developed for overcoming the drawbacks of the supervised deep learning technologies and reducing noise in CT images. The DRL model was consisted of shared, value and policy networks, and the networks included convolutional layers, rectified linear unit (ReLU), dilation factors and gate rotation unit (GRU) in order to extract noise features from CT images and improve the performance of the DRL model. Also, the quality of the CT images obtained by using the DRL model was compared to that obtained by using the supervised deep learning model. The results showed that the image accuracy for the DRL model was higher than that for the supervised deep learning model, and the image noise for the DRL model was smaller than that for the supervised deep learning model. Also, the DRL model reduced the noise of the CT images, which had different characteristics with training images. Therefore, the DRL model is able to reduce image noise as well as maintain the structural information of CT images.

• The Journal of Society for e-Business Studies
• /
• v.26 no.1
• /
• pp.93-106
• /
• 2021

Many recent studies on deep learning-based road crack detection have shown significantly more improved performances than previous works using algorithm-based conventional approaches. However, many deep learning-based studies are still focused on classifying the types of cracks. The classification of crack types is highly anticipated in that it can improve the crack detection process, which is currently relying on manual intervention. However, it is essential to calculate the severity of the cracks as well as identifying the type of cracks in actual pavement maintenance planning, but studies related to road crack detection have not progressed enough to automated calculation of the severity of cracks. In order to calculate the severity of the crack, the type of crack and the area of the crack in the image must be identified together. This study deals with a method of using Mobilenet-SSD that is deep learning-based object detection techniques to effectively automate the simultaneous detection of crack types and crack areas. To improve the accuracy of object-detection for road cracks, several experiments were conducted to combine the U-Net for automatic segmentation of input image and object-detection model, and the results were summarized. As a result, image masking with U-Net is able to maximize object-detection performance with 0.9315 mAP value. While referring the results of this study, it is expected that the automation of the crack detection functionality on pave management system can be further enhanced.

• Therapeutic Science for Rehabilitation
• /
• v.10 no.1
• /
• pp.105-114
• /
• 2021

Objective : The purpose of this study was to identify and compare the sex differences in episodic memory and spatial cognition in healthy young adults. Methods : Forty-eight undergraduates (male=24, female=24) were assessed for sex differences using the visual stimuli episodic memory task and the virtual reality-based spatial cognition task. The accuracy rates (%) for the What, Where, and When conditions of the episodic memory task and the average distance error (cm) for 10 trials of the spatial cognition task were analyzed. Results : There were no significant sex differences between the three conditions. The male participants showed a significantly higher performance on the spatial cognition task than the female participants Conclusion : The results of this study indicated that the sex differences in episodic memory could be altered by the test methods. Although episodic memory and spatial cognition mainly depend on the hippocampus, the sex-related differences between the two functions were inconsistent, suggesting that these two functions are independent.

• Journal of Convergence for Information Technology
• /
• v.11 no.2
• /
• pp.98-106
• /
• 2021

The paper deals with comparative study on sensitivity analysis using various methods regarding to design of experiments for structure design of an active type DSF (Deck support frame) that was developed for float-over installation of offshore plant. The thickness sizing variables of structure member of the active type DSF were considered the design factors. The output responses were defined from the weight and the strength performances. Various methods such as orthogonal array design, Box-Behnken design, and Latin hypercube design were applied to the comparative study. In order to evaluate the approximation performance of the design space exploration according to the design of experiments, response surface method was generated for each design of experiment, and the accuracy characteristics of the approximation were reviewed. The design enhancement results such as numerical costs, weight minimization, etc. via the design of experiment methods were compared to the results of the best design. The orthogonal array design method represented the most improved results for the structure design of the active type DSF.

• Korean Journal of Food Science and Technology
• /
• v.52 no.6
• /
• pp.580-586
• /
• 2020

Fishery resources have been widely consumed as protein- and vitamin-rich food sources in the Korean diet. However, information regarding their vitamin levels is extremely limited. In this study, trienzyme-Lactobacillus casei method was validated and used to determine the folate contents in fishery foods. The trienzyme-L. casei method for folate analysis showed excellent accuracy (85.2 to 95.3% recovery) and precision (repeatability 1.4% RSD and reproducibility 2.4% RSD). Folate contents of 20 fish foods (4 fish, 3 crustaceans, 3 sea algae, 3 cephalopods, 4 shellfish, and 3 others) ranged from 1.75 to 97.98 ㎍/100 g. Furthermore, we found that the folate content in seaweed fusiforme was the highest, followed by gulfweed (69.73 ㎍/100 g). Folate analysis using the trienzyme-L. casei method was determined excellent based on the z-score of -0.3 in the Food Analysis Performance Assessment Scheme test. Analytical and method validation data generated in this study could be used to update the national food composition table on vitamin B9 in Korean fishery resources.

• Korean Journal of Remote Sensing
• /
• v.37 no.3
• /
• pp.503-514
• /
• 2021

Rice production with adequate level of area is important for decision making of rice supply and demand policy. It is essential to grasp rice cultivation areas in advance for estimating rice production of the year. This study was carried out to classify paddy rice cultivation in Gimje-si using sentinel-1 SAR (synthetic aperture radar) and UAV imagery in early July. Time-series Sentinel-1A and 1B images acquired from early May to early July were processed to convert into sigma naught (dB) images using SNAP (SeNtinel application platform, Version 8.0) toolbox provided by European Space Agency. Farm map and parcel map, which are spatial data of vector polygon, were used to stratify paddy field population for classifying rice paddy cultivation. To distinguish paddy rice from other crops grown in the paddy fields, we used the decision tree method using threshold levels and random forest model. Random forest model, trained by mainly rice cultivation area and rice and soybean cultivation area in UAV image area, showed the best performance as overall accuracy 89.9%, Kappa coefficient 0.774. Through this, we were able to confirm the possibility of early estimation of rice cultivation area in Gimje-si using UAV image.

• Journal of the Korean Society of Radiology
• /
• v.15 no.3
• /
• pp.307-312
• /
• 2021

In nuclear medicine, radioactive isotope tracers are administered to the human body to obtain and evaluate disease morphological information and biological function information. Dose calibrator is a device used to measure the radioactivity of a single nuclide in medical institutions. Administration of the correct dose to the human body acts as an important factor in diagnosis and treatment, and measurement through a dose calibrator before administration is the most important factor. Dose calibrator performs daily quality control after installation in each medical institution. Quality control is a means of guaranteeing quality control after installation, and is essential for improving the quality of treatment and promoting patient safety. Therefore, accurate and standardized performance evaluation methods should be established. In this study, 3D printing was used for quantitative evaluation of quality control by increasing the accuracy and standardization of quality control. When the 3D printer was installed and reproducibility was tested, the error range of the expected value and reading value decreased by 0.302% in the F-18 nuclide and 0.09% in the ^99mTc-pertechnate nuclide than when the 3D printer was installed. The error rate for other nuclides was also found to have a low error rate for reproducibility tests when 3D printing was installed.

• The Journal of the Korea Contents Association
• /
• v.21 no.6
• /
• pp.35-43
• /
• 2021

Citrus quality is generally determined by its sugar content and acidity. In particular, sugar content is a very important factor because it determines the taste of citrus. Currently, the most commonly used method of measuring citrus sugar content in farms is a portable juiced sugar meter and a non-destructive sugar meter. This method can be easily measured by individuals, but the accuracy of the sugar content is inferior to that of the citrus NongHyup official machine. In particular, there is an error difference of 0.5 Brix or more, which is still insufficient for use in the field. Therefore, in this paper, we propose an AI model that predicts the citrus sugar content of unmeasured days within the error range of 0.5 Brix or less based on the previously collected citrus sugar content and meteorological data (average temperature, humidity, rainfall, solar radiation, and average wind speed). In addition, it was confirmed that the prediction model proposed through performance evaluation had an mean absolute error of 0.1154 for Seongsan area and 0.1983 for the Hawon area in Jeju Island. Lastly, the proposed model supports an error difference of less than 0.5 Brix and is a technology that supports predictive measurement, so it is expected that its usability will be highly progressive.