• Geophysics and Geophysical Exploration
• /
• v.25 no.3
• /
• pp.140-161
• /
• 2022

Seismic data with missing traces are often obtained regularly or irregularly due to environmental and economic constraints in their acquisition. Accordingly, seismic data interpolation is an essential step in seismic data processing. Recently, research activity on machine learning-based seismic data interpolation has been flourishing. In particular, convolutional neural network (CNN) and generative adversarial network (GAN), which are widely used algorithms for super-resolution problem solving in the image processing field, are also used for seismic data interpolation. In this study, CNN-based algorithm, U-Net and GAN-based algorithm, and conditional Wasserstein GAN (cWGAN) were used as seismic data interpolation methods. The results and performances of the methods were evaluated thoroughly to find an optimal interpolation method, which reconstructs with high accuracy missing seismic data. The work process for model training and performance evaluation was divided into two cases (i.e., Cases I and II). In Case I, we trained the model using only the regularly sampled data with 50% missing traces. We evaluated the model performance by applying the trained model to a total of six different test datasets, which consisted of a combination of regular, irregular, and sampling ratios. In Case II, six different models were generated using the training datasets sampled in the same way as the six test datasets. The models were applied to the same test datasets used in Case I to compare the results. We found that cWGAN showed better prediction performance than U-Net with higher PSNR and SSIM. However, cWGAN generated additional noise to the prediction results; thus, an ensemble technique was performed to remove the noise and improve the accuracy. The cWGAN ensemble model removed successfully the noise and showed improved PSNR and SSIM compared with existing individual models.

• Journal of Science Education
• /
• v.47 no.2
• /
• pp.127-138
• /
• 2023

In this study, the implementation of science classes by pre-service elementary teachers were explored to foster the digital literacy of elementary students. To this end, interviews were conducted with four university students in their 4th year of study at 'G' Education University in Gwangju, also their teaching plans and class implementation videos were analyzed. Interview questions were divided into the science class planning stage, the science class implementation stage, and the post-science class implementation stage, and the interview contents were transcribed and categorized after recording. The results of the study are as follows. First, in the science class planning stage, pre-service elementary teachers considered the degree of elementary students' familiarity when choosing digital teaching and learning tools to foster digital literacy. While writing the teaching plan, it was difficult to set standards related to fostering digital literacy. Second, while pre-service elementary teachers had positive experiences such as improving students' participation in classes and increasing student-specific guidance through real-time data sharing, class time was exceeded due to poor use of digital teaching and learning tools. Third, after teaching their science classes, pre-service teachers presented reflections and requirements for the class. Based on these results, the necessity of developing digital literacy for pre-service elementary teachers and the need to improve the curriculum of education university were proposed.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2021.06a
• /
• pp.246-246
• /
• 2021

본 연구는 강우 발생시 유량을 추정하는 것에 목적이 있다. 이를 위해 본 연구는 선행연구의 모형 개발방법론에서 벗어나 딥러닝 알고리즘 중 하나인 합성곱 신경망 (convolution neural network)과 수문학적 이미지 (hydrological image)를 이용하여 강우 발생시 유량을 추정하였다. 합성곱 신경망은 일반적으로 분류 문제 (classification)을 해결하기 위한 목적으로 개발되었기 때문에 불특정 연속변수인 유량을 모의하기에는 적합하지 않다. 이를 위해 본 연구에서는 합성곱 신경망의 완전 연결층 (Fully connected layer)를 개선하여 연속변수를 모의할 수 있도록 개선하였다. 대부분 합성곱 신경망은 RGB (red, green, blue) 사진 (photograph)을 이용하여 해당 사진이 나타내는 것을 예측하는 목적으로 사용하지만, 본 연구의 경우 일반 RGB 사진을 이용하여 유출량을 예측하는 것은 경험적 모형의 전제(독립변수와 종속변수의 관계)를 무너뜨리는 결과를 초래할 수 있다. 이를 위해 본 연구에서는 임의의 유역에 대해 2차원 공간에서 무차원의 수문학적 속성을 갖는 grid의 집합으로 정의되는 수문학적 이미지는 입력자료로 활용했다. 합성곱 신경망의 구조는 Convolution Layer와 Pulling Layer가 5회 반복하는 구조로 설정하고, 이후 Flatten Layer, 2개의 Dense Layer, 1개의 Batch Normalization Layer를 배열하고, 다시 1개의 Dense Layer가 이어지는 구조로 설계하였다. 마지막 Dense Layer의 활성화 함수는 분류모형에 이용되는 softmax 또는 sigmoid 함수를 대신하여 회귀모형에서 자주 사용되는 Linear 함수로 설정하였다. 이와 함께 각 층의 활성화 함수는 정규화 선형함수 (ReLu)를 이용하였으며, 모형의 학습 평가 및 검정을 판단하기 위해 MSE 및 MAE를 사용했다. 또한, 모형평가는 NSE와 RMSE를 이용하였다. 그 결과, 모형의 학습 평가에 대한 MSE는 11.629.8 m3/s에서 118.6 m3/s로, MAE는 25.4 m3/s에서 4.7 m3/s로 감소하였으며, 모형의 검정에 대한 MSE는 1,997.9 m3/s에서 527.9 m3/s로, MAE는 21.5 m3/s에서 9.4 m3/s로 감소한 것으로 나타났다. 또한, 모형평가를 위한 NSE는 0.7, RMSE는 27.0 m3/s로 나타나, 본 연구의 모형은 양호(moderate)한 것으로 판단하였다. 이에, 본 연구를 통해 제시된 방법론에 기반을 두어 CNN 모형 구조의 확장과 수문학적 이미지의 개선 또는 새로운 이미지 개발 등을 추진할 경우 모형의 예측 성능이 향상될 수 있는 여지가 있으며, 원격탐사 분야나, 위성 영상을 이용한 전 지구적 또는 광역 단위의 실시간 유량 모의 분야 등으로의 응용이 가능할 것으로 기대된다.

• Industry Promotion Research
• /
• v.8 no.3
• /
• pp.181-199
• /
• 2023

E-commerce is gradually upgrading from traditional text and image formats to short video and livestreaming formats. Livestreaming e-commerce enriches the content and forms of information dissemination and product display, enhances the consumer's shopping experience, and gradually becomes the mainstream new consumer scene. However, there are many negative phenomena in the development of livestreaming e-commerce, such as false propaganda, counterfeit goods, and various negative events, which seriously affect the level of consumer trust in livestreaming e-commerce. Trust is the core competitive factor of livestreaming e-commerce. Based on previous research on trust theory and combined with the characteristic elements of "people, goods, and scenes" of livestreaming e-commerce, this article constructs a trust model for livestreaming e-commerce, proposes hypotheses, and proves through empirical research that factors such as store characteristics, livestream host characteristics, brand image, product information, platform reputation, livestreaming situation, and trust tendency have a significant positive impact on consumer trust. Based on the research conclusions, this article provides insights and management suggestions, such as emphasizing the construction of store characteristic indicators, creating desirable livestream host characteristics, focusing on product brand building and selection, maintaining the display of product information, selecting suitable livestreaming platforms, and creating rich content for livestreaming situations.

• Journal of Korea Water Resources Association
• /
• v.56 no.2
• /
• pp.139-150
• /
• 2023

Terretial LiDAR surveying is highly useful for maintenance of civil facilities as it can easily detect the temporal deformation of structures or topography. However, for river facilities such as levess, it is difficult to detect the deformation of the topography or structure under vegetations due to the influence of vegetation. Vegetation filters can be divided into color filters and morphological filters. In this study, combined filters with color and morphology are developed to improve the accuracy of vegetation filters. 8 color filters, 6 morphological filters, and 4 combined filters are applied to the vegetation removal on the embankment slope, and their accuracy and calculation time are compared. Color filters show a short calculation time, but the accuracy was low in the vegetation area. Morphological filters show high accuracy in the vegetation area, but low accuracy in places with severe local topographical changes such as heavy rocks. Combined filters also show a tendency similar to morphological filters, but in the case of ExGGM, the accuracy is excellent in both the vegetation and rock area. Considering the accuracy and calculation time, the combined filter ExGGM is suitable for general cases, and the shape filter GrMIn or the complex filter ExGISL is suitable for cases where the local topographical change is not severe.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.5
• /
• pp.217-228
• /
• 2023

Video instance segmentation is an intelligent visual task with high complexity because it not only requires object instance segmentation for each image frame constituting a video, but also requires accurate tracking of instances throughout the frame sequence of the video. In special, human instance segmentation in drama videos has an unique characteristic that requires accurate tracking of several main characters interacting in various places and times. Also, it is also characterized by a kind of the class imbalance problem because there is a significant difference between the frequency of main characters and that of supporting or auxiliary characters in drama videos. In this paper, we introduce a new human instance datatset called MHIS, which is built upon drama videos, Miseang, and then propose a novel video data augmentation method, CDVA, in order to overcome the data imbalance problem between character classes. Different from the previous video data augmentation methods, the proposed CDVA generates more realistic augmented videos by deciding the optimal location within the background clip for a target human instance to be inserted with taking rich spatio-temporal context embedded in videos into account. Therefore, the proposed augmentation method, CDVA, can improve the performance of a deep neural network model for video instance segmentation. Conducting both quantitative and qualitative experiments using the MHIS dataset, we prove the usefulness and effectiveness of the proposed video data augmentation method.

• Journal of Korean Society of Archives and Records Management
• /
• v.23 no.2
• /
• pp.49-67
• /
• 2023

In recent years, the film industry in South Korea has experienced a period of prosperity, evidenced by the numerous awards won at major international film festivals. Furthermore, growing global interest in K-content and the expansion of the OTT industry following the COVID-19 pandemic are providing favorable conditions for the development of the domestic film industry. Sound effects play a crucial role in conveying the atmosphere and emotions of a film, making them an essential element of film production. In response, the Jeonju IT & CT Industry Promotion Agency has been promoting the development of Korean-style sound effects since 2013. Furthermore, the agency launched an online service called the "K-Sound Library," a sound effect archive, in 2021. However, the service has not been widely utilized because of issues with the database's construction and the system's problems. Therefore, this study aims to identify the K-Sound Library's problems through interviews with sound effects specialists about the online service of the first sound effect archive in South Korea. Based on the interviews and analyses of foreign cases, the study suggests ways to improve the search services' usability and the sound effects classification system.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.42-42
• /
• 2022

지금까지도 유역에서의 녹조 모니터링은 현장채수를 통한 점 단위 모니터링에 크게 의존하고 있어 기후, 유속, 수온조건 등에 따라 수체에 광범위하게 발생하는 녹조를 효율적으로 모니터링하고 대응하기에는 어려운 점들이 있어왔다. 또한, 그동안 제한된 관측 데이터로 인해 현장 측정된 실측 데이터 보다는 녹조와 관련이 높은 NDVI, FGAI, SEI 등의 파생적인 지수를 산정하여 원격탐사자료와 매핑하는 방식의 분석연구 등이 선행되었다. 본 연구는 녹조의 모니터링시 정확도와 효율성을 향상을 목표로 하여, 우선은 녹조 측정장비를 활용, 7000개 이상의 녹조 관측 데이터를 확보하였으며, 이를 바탕으로 동기간의 고해상도 위성 자료와 실측자료를 매핑하기 위해 다양한Machine Learning기법을 적용함으로써 그 효과성을 검토하고자 하였다. 연구대상지는 낙동강 내성천 상류에 위치한 영주댐 유역으로서 데이터 수집단계에서는 면단위 현장(in-situ) 관측을 위해 2020년 2~9월까지 4회에 걸쳐 7291개의 녹조를 측정하고, 동일 시간 및 공간의 Sentinel-2자료 중 Band 1~12까지 총 13개(Band 8은 8과 8A로 2개)의 분광특성자료를 추출하였다. 다음으로 Machine Learning 분석기법의 적용을 위해 algae_monitoring Python library를 구축하였다. 개발된 library는 1) Training Set과 Test Set의 구분을 위한 Data 준비단계, 2) Random Forest, Gradient Boosting Regression, XGBoosting 알고리즘 중 선택하여 적용할 수 있는 모델적용단계, 3) 모델적용결과를 확인하는 Performance test단계(R², MSE, MAE, RMSE, NSE, KGE 등), 4) 모델결과의 Visualization단계, 5) 선정된 모델을 활용 위성자료를 녹조값으로 변환하는 적용단계로 구분하여 영주댐뿐만 아니라 다양한 유역에 범용적으로 적용할 수 있도록 구성하였다. 본 연구의 사례에서는 Sentinel-2위성의 12개 밴드, 기상자료(대기온도, 구름비율) 총 14개자료를 활용하여 Machine Learning기법 중 Random Forest를 적용하였을 경우에, 전반적으로 가장 높은 적합도를 나타내었으며, 적용결과 Test Set을 기준으로 NSE(Nash Sutcliffe Efficiency)가 0.96(Training Set의 경우에는 0.99) 수준의 성능을 나타내어, 광역적인 위성자료와 충분히 확보된 현장실측 자료간의 데이터 학습을 통해서 조류 모니터링 분석의 효율성이 획기적으로 증대될 수 있음을 확인하였다.

• Journal of Digestive Cancer Research
• /
• v.2 no.2
• /
• pp.75-77
• /
• 2014

Ampulla of Vater (AOV) cancer is rare malignant tumor which arises within the vicinity of the AOV. Metastatic AOV adenocarcinoma has poor prognosis, with an overall survival rate at 2 years ranging from 5 to 10%. The Surveillance, Epidemiology and End Results Program of the National Cancer Institute indicated that lymph node metastasis was present in as many as half of patients which were associated with poor prognosis and liver was the second most common site of distant metastasis in AOV cancer. In this case report, we describe a case of complete resolution of AOV cancer, which was already spread to retroperitoneal lymph node and liver. The patient underwent gemcitabine plus cisplatin chemotherapy for palliative aim. After 12 month of chemotherapy, image study showed partial remission, so intraoperative radiofrequency ablation therapy and pylorus preserving pancreaticoduodenectomy was done. AOV cancer was completely resected and the patient was followed up without recurrence for 7 months.

• Journal of the Korean Society of Radiology
• /
• v.17 no.7
• /
• pp.1025-1031
• /
• 2023

We aimed to evaluate the radiation dose and image quality by changing the Scout view voltage in low-dose chest CT (LDCT) and applying scan parameters such as AEC (auto exposure control) and ASIR (adaptive statistical iterative reconstruction) to find the optimal protocol. Scout view voltage was varied at 80, 100, 120, 140 kV and after measuring the dose 5 times using the existing low-dose chest CT protocol, the appropriate kV was selected for the study using the Dose report provided by the equipment. After taking a basic LDCT shot at 120 kV, 30 mAs, ASIR 50% was applied to this condition. Signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were assessed by measuring Background noise (B/N). For dose comparison, CTDI_vol and DLP provided by the equipment were compared and analyzed using the formulas. The results indicated that the protocol of scout 140 + LDCT + ASIR 50 + AEC reduced radiation exposure and improved image quality compared to traditional LDCT, providing an optimal protocol. As demonstrated in the experiment, LDCT screenings for asymptomatic normal individuals are crucial, as they involve concerns over excessive radiation exposure per examination. Therefore, applying appropriate parameters is important, and it is expected to contribute positively to the public health in future LDCT based health screenings.