• Journal of Biomedical Engineering Research
• /
• v.40 no.2
• /
• pp.48-54
• /
• 2019

This paper proposed a method of validation data augmentation for improving the grading accuracy of diabetic macular edema (DME) using deep learning. The data augmentation technique is basically applied in order to secure diversity of data by transforming one image to several images through random translation, rotation, scaling and reflection in preparation of input data of the deep neural network (DNN). In this paper, we apply this technique in the validation process of the trained DNN, and improve the grading accuracy by combining the classification results of the augmented images. To verify the effectiveness, 1,200 retinal images of Messidor dataset was divided into training and validation data at the ratio 7:3. By applying random augmentation to 359 validation data, $1.61{\pm}0.55%$ accuracy improvement was achieved in the case of six times augmentation (N=6). This simple method has shown that the accuracy can be improved in the N range from 2 to 6 with the correlation coefficient of 0.5667. Therefore, it is expected to help improve the diagnostic accuracy of DME with the grading information provided by the proposed DNN.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.189-189
• /
• 2022

물 공급 시설의 효율적이고 안정적인 운영을 위한 운영 계획의 수립 및 검증을 위해서는 장기간의 유입량 자료가 필요하다. 하지만, 현실적으로 얻을 수 있는 실측 자료는 제한적이며, 유입량이 부족하여 댐 운영에 영향을 미치는 자료는 더욱 적을 수밖에 없다. 이를 개선하고자 장기간의 모의 유입량을 생성해 운영 계획을 수립하는 방법이 종종 사용되지만, 실측 자료를 기반으로 모의하기 때문에 이 역시 가뭄의 빈도가 낮아, 장기 가뭄이나 짧은 간격으로 가뭄이 발생할 시 안정적인 운영이 어렵다. 본 연구에서는 장기 가뭄 발생 시에도 안정적인 물 공급이 가능한 운영 계획 수립을 위해 가뭄 빈도를 증가시킨 유입량 모의 기법을 제안하고자 한다. 제안하는 모의 기법은 최근 머신러닝에서 사용되는 SMOTE 알고리즘을 기반으로 한다. SMOTE 알고리즘은 데이터의 불균형을 처리하기 위한 오버 샘플링 기법으로, 소수 그룹을 단순 복제하지 않고 새로운 복제본을 생성해 과적합의 위험이 적으며, 원자료의 정보가 손실되지 않는 장점이 있다. 본 연구에서는 미국 캘리포니아주에 위치한 Folsom 댐을 대상으로 고빈도 가뭄 유입량을 모의했으며, 고빈도 가뭄 유입량을 사용한 운영 계획을 수립하였다. Folsom 댐의 과거 관측 유입량 자료를 기반으로 고빈도 가뭄 유입량을 사용한 운영 계획과 일반적인 가뭄 빈도의 유입량을 사용한 운영 계획을 적용했을 때 발생하는 공급 부족량과 과잉 방류량의 차이를 비교해 고빈도 가뭄 유입량의 사용이 물 공급 시설의 안정적인 운영에 끼치는 영향을 확인하고자 한다.

• Annual Conference of KIPS
• /
• 2022.05a
• /
• pp.705-707
• /
• 2022

인체의 흉부 X-ray 영상으로부터 척추질환과 관련된 의료 진단지표를 자동으로 추출하는 과정을 위하여 흉추조직의 정확한 분할이 필요하다. 본 연구에서는 HRNet 기반의 학습을 통하여 흉추조직을 분할하는 방법을 고찰한다. 분할 과정에서 영상 내의 상대적인 위치 정보가 효과적으로 반영될 수 있도록, 계층별로 영상의 고해상도의 표현이 그대로 유지되는 구조와 저해상도의 특징 지도로 변환되는 구조가 병렬적으로 연결되는 형태의 심층 신경망 모델을 채택하였다. 흉부 X-ray 영상에서 콥각도(Cobb's angle)를 산출하는 문제를 대상으로 흉추 분할을 위한 학습 방법, 진단지표 추출 방법 등을 소개하며, 부수적으로 피사체의 위치 변화 및 크기 변화 등에 강인한 성능을 제공하기 위하여 학습 데이터를 증강하는 방법론을 제시하였다. 총 145개의 영상을 사용한 실험을 통하여 제안된 이론의 타당성을 평가하였다.

• Annual Conference of KIPS
• /
• 2023.11a
• /
• pp.894-895
• /
• 2023

증강현실(AR)이란, 현실의 이미지나 배경에 3차원 가상 이미지를 입혀 실시간으로 출력되는 컴퓨터 그래픽 기법이다. 본 논문에서는 AR 서비스를 이용한 여행 기록 애플리케이션을 제안한다. 현재 제공되는 서비스에서는 장소에 대해 제공하는 정보가 매우 한정적이다. 따라서 공유 커뮤니티를 형성할 수 있도록 하여 다양한 정보를 전달하고 직접적인 시각 자료를 제공받을 수 있는 애플리케이션이 필요하다고 생각하여 기획하게 되었다. ReactNative와 SpringBoot를 사용하며, MySQL을 이용하여 데이터를 저장한다. Google Map API를 활용하여 사용자가 기록하고자 하는 장소에 대한 위도와 경도 정보를 받을 수 있도록 하며, AR 기능은 Unity로 구현하였다. AR을 포함한 혼합 현실 서비스를 제공하여 사용자들이 본인의 경험에만 의존하는 것이 아닌 풍부한 몰입력과 현장감을 통해 보다 가시적이고 생생한 다이어리 기록을 만들 수 있다.

• The Journal of Korea Robotics Society
• /
• v.19 no.2
• /
• pp.203-212
• /
• 2024

Simultaneous motion is essential in the activities of daily living (ADL). For motion intention recognition, surface electromyogram (sEMG) and corresponding motion label is necessary. However, this process is time-consuming and it may increase the burden of the user. Therefore, we propose a simultaneous motion recognition framework using data augmentation based on muscle activation model. The model consists of multiple point sources to be optimized while the number of point sources and their initial parameters are automatically determined. From the experimental results, it is shown that the framework has generated the data which are similar to the real one. This aspect is quantified with the following two metrics: structural similarity index measure (SSIM) and mean squared error (MSE). Furthermore, with k-nearest neighbor (k-NN) or support vector machine (SVM), the classification accuracy is also enhanced with the proposed framework. From these results, it can be concluded that the generalization property of the training data is enhanced and the classification accuracy is increased accordingly. We expect that this framework reduces the burden of the user from the excessive and time-consuming data acquisition.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.27 no.3
• /
• pp.319-328
• /
• 2024

Semi-supervised learning is a good way to train a classification model using a small number of labeled and large number of unlabeled data. We applied semi-supervised learning to a synthetic aperture radar(SAR) image classification model with a limited number of datasets that are difficult to create. To address the previous difficulties, semi-supervised learning uses a model trained with a small amount of labeled data to generate and learn pseudo labels. Besides, a lot of number of papers use a single fixed threshold to create pseudo labels. In this paper, we present a semi-supervised synthetic aperture radar(SAR) image classification method that applies different thresholds for each class instead of all classes sharing a fixed threshold to improve SAR classification performance with a small number of labeled datasets.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.5
• /
• pp.243-250
• /
• 2016

Recently, public interest in virtual reality (VR) and augmented reality (AR) has increased. Therefore, computer graphics-related research has been actively conducted. This has included research on virtual space related to human posture implementation. However, such research has focused on general posture in humans. This paper presents a system with reference to the basic posture in sports climbing and the inverse kinematics method for generating the positions and behavior of virtual characteristics in a three-dimensional virtual space. The simulation based on the inverse kinematics method, produced with an inverse kinematics solver and initial pose animation from motion capture, provides realistic and natural movement. We designed a simulation system to generate correct posture and motions similar to those in sports climbing by applying the basic procedure of sports climbing. The simulation system provides help for producing content about sports climbing, such as learning programs for novice climbers and sports climbing games.

• Journal of Digital Convergence
• /
• v.16 no.8
• /
• pp.129-136
• /
• 2018

This study analyzed technology application field and technology transfer type related to the 4th industrial revolution using frequency, visualization, and association analysis of text mining of Big Data. The analysis was conducted between the last three years (2015 - 2017) registered with the NTB of KIAT transfer technology database was utilized. As a result of analysis, First, First, transfer technologies called core technologies of the Fourth Industrial Revolution are a lot of about robots, 3D, autonomous driving, and wearables. Second, as the year go by, transfer technolgy registration such as IoT, Cloud, VR is increasing. Third, the results of the association analysis of technology transfer type are as follows. IoT and VR showed preference for technology trading and licensing, autonomous driving technology trading, wearable licensing, robots preferring technology cooperation, licensing, and technology trading.

• Journal of Broadcast Engineering
• /
• v.23 no.3
• /
• pp.345-350
• /
• 2018

CNN (convolution neural network) has become the most popular artificial intelligence technique and shows remarkable performance in image classification task. In this paper, we propose a CNN-based classification method for various street images as well as a method of evaluating the safety score for the street. The proposed method consists of learning four types of street images using CNN and classifying input street images using the learned CNN model followed by evaluating the safety score. During the learning process, four types of street images are collected and augmented, and then CNN learning is performed. It is shown that learned CNN model classifies input images correctly and the safety scores are evaluated quantitatively by combining the probabilities of different street types.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.5
• /
• pp.55-64
• /
• 2018

In this paper, we propose fully convolutional neural networks based automatic detection of a pothole that directly causes driver's safety accidents and the vehicle damage. First, the training DB is collected through the camera installed in the vehicle while driving on the road, and the model is trained in the form of a semantic segmentation using the fully convolutional neural networks. In order to generate robust performance in a dark environment, we augmented the training DB according to brightness, and finally generated a total of 30,000 training images. In addition, a total of 450 evaluation DB was created to verify the performance of the proposed automatic pothole detection, and a total of four experts evaluated each image. As a result, the proposed pothole detection showed robust performance for missing.