• Journal of Korea Game Society
• /
• v.19 no.5
• /
• pp.103-114
• /
• 2019

We have been studying a deep-learning program that can communicate with other users in online communities since 2017. But there were problems with processing a Korean data set because of Korean characteristics. Also, low usage of GPUs of RNN models was a problem too. In this study, as Natural Language Processing models are improved, we aim to make better results using these improved models. To archive this, we use a Transformer model which includes Self-Attention mechanism. Also we use MeCab, korean morphological analyzer, to address a problem with processing korean words.

• Journal of Korea Multimedia Society
• /
• v.22 no.3
• /
• pp.349-356
• /
• 2019

When a prescription change occurs in the hospital depending on a patient's improvement status, pharmacists directly classify manually returned pills which are not taken by a patient. There are hundreds of kinds of pills to classify. Because it is manual, mistakes can occur and which can lead to medical accidents. In this study, we have compared YOLO, Faster R-CNN and RetinaNet to classify and detect pills. The data consisted of 10 classes and used 100 images per class. To evaluate the performance of each model, we used cross-validation. As a result, the YOLO Model had sensitivity of 91.05%, FPs/image of 0.0507. The Faster R-CNN's sensitivity was 99.6% and FPs/image was 0.0089. The RetinaNet showed sensitivity of 98.31% and FPs/image of 0.0119. Faster RCNN showed the best performance among these three models tested. Thus, the most appropriate model for classifying pills among the three models is the Faster R-CNN with the most accurate detection and classification results and a low FP/image.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.101-101
• /
• 2020

Traditional stochastic simulation of hydroclimatological variables often underestimates the variability and correlation structure of larger timescale due to the difficulty in preserving long-term memory. However, the Long Short-Term Memory (LSTM) model illustrates a remarkable long-term memory from the recursive hidden and cell states. The current study, therefore, employed the LSTM model in stochastic generation of hydrologic and climate variables to examine how much the LSTM model can preserve the long-term memory and overcome the drawbacks of conventional time series models such as autoregressive (AR). A trigonometric function and the Rössler system as well as real case studies for hydrological and climatological variables were tested. Results presented that the LSTM model reproduced the variability and correlation structure of the larger timescale as well as the key statistics of the original time domain better than the AR and other traditional models. The hidden and cell states of the LSTM containing the long-memory and oscillation structure following the observations allows better performance compared to the other tested conventional models. This good representation of the long-term variability can be important in water manager since future water resources planning and management is highly related with this long-term variability.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.9
• /
• pp.2334-2347
• /
• 2023

Metaverse services generate text data, data of ubiquitous computing, in real-time to analyze user emotions. Analysis of user emotions is an important task in metaverse services. This study aims to classify user sentiments using deep learning and pre-trained language models based on the transformer structure. Previous studies collected data from a single platform, whereas the current study incorporated the review data as "Metaverse" keyword from the YouTube and Google Play Store platforms for general utilization. As a result, the Bidirectional Encoder Representations from Transformers (BERT) and Robustly optimized BERT approach (RoBERTa) models using the soft voting mechanism achieved a highest accuracy of 88.57%. In addition, the area under the curve (AUC) score of the ensemble model comprising RoBERTa, BERT, and A Lite BERT (ALBERT) was 0.9458. The results demonstrate that the ensemble combined with the RoBERTa model exhibits good performance. Therefore, the RoBERTa model can be applied on platforms that provide metaverse services. The findings contribute to the advancement of natural language processing techniques in metaverse services, which are increasingly important in digital platforms and virtual environments. Overall, this study provides empirical evidence that sentiment analysis using deep learning and pre-trained language models is a promising approach to improving user experiences in metaverse services.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.5
• /
• pp.737-745
• /
• 2023

This paper discusses a study on developing an artificial intelligence model to detect incidents of fires in distribution switchboard using thermal images. The objective of the research is to preprocess collected thermal images into suitable data for object detection models and design a model capable of determining the occurrence of fires within distribution panels. The study utilizes thermal image data from AI-HUB's industrial complex for training. Two CNN-based deep learning object detection algorithms, namely Faster R-CNN and RetinaNet, are employed to construct models. The paper compares and analyzes these two models, ultimately proposing the optimal model for the task.

• Proceedings of the Korean Society of Crop Science Conference
• /
• 2022.10a
• /
• pp.80-80
• /
• 2022

Estimation of heading date of paddy rice is laborious and time consuming. Therefore, automatic estimation of heading date of paddy rice is highly essential. In this experiment, deep learning classification models were used to classify two difference categories of rice (vegetative and reproductive stage) based on the panicle initiation of paddy field. Specifically, the dataset includes 444 slanted view images belonging to two categories and was then expanded to include 1,497 images via IMGAUG data augmentation technique. We adopt two transfer learning strategies: (First, used transferring model weights already trained on ImageNet to six classification network models: VGGNet, ResNet, DenseNet, InceptionV3, Xception and MobileNet, Second, fine-tuned some layers of the network according to our dataset). After training the CNN model, we used several evaluation metrics commonly used for classification tasks, including Accuracy, Precision, Recall, and F1-score. In addition, GradCAM was used to generate visual explanations for each image patch. Experimental results showed that the InceptionV3 is the best performing model in terms of the accuracy, average recall, precision, and F1-score. The fine-tuned InceptionV3 model achieved an overall classification accuracy of 0.95 with a high F1-score of 0.95. Our CNN model also represented the change of rice heading date under different date of transplanting. This study demonstrated that image based deep learning model can reliably be used as an automatic monitoring system to detect the heading date of rice crops using CCTV camera.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.20 no.11
• /
• pp.305-313
• /
• 2019

Recently, spatial information is being constructed actively based on the images obtained by drones. Because occlusion areas occur due to buildings as well as many obstacles, such as trees, pedestrians, and banners in the urban areas, an efficient way to resolve the problem is necessary. Instead of the traditional way, which replaces the occlusion area with other images obtained at different positions, various models based on deep learning were examined and compared. A comparison of a type of feature descriptor, HOG, to the machine learning-based SVM, deep learning-based DNN, CNN, and RNN showed that the CNN is used broadly to detect and classify objects. Until now, many studies have focused on the development and application of models so that it is impossible to select an optimal model. On the other hand, the upgrade of a deep learning-based detection and classification technique is expected because many researchers have attempted to upgrade the accuracy of the model as well as reduce the computation time. In that case, the procedures for generating spatial information will be changed to detect the occlusion area and replace it with simulated images automatically, and the efficiency of time, cost, and workforce will also be improved.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.40 no.4
• /
• pp.251-260
• /
• 2022

Although the number of trees affected by pine wilt disease is decreasing, the affected area is expanding across the country. Recently, with the development of deep learning technology, it is being rapidly applied to the detection study of pine wilt nematodes and dead trees. The purpose of this study is to efficiently acquire deep learning training data and acquire accurate true values to further improve the detection ability of U-Net models through learning. To achieve this purpose, by using a filtering method applying a step-by-step deep learning algorithm the ambiguous analysis basis of the deep learning model is minimized, enabling efficient analysis and judgment. As a result of the analysis the U-Net model using the true values analyzed by period in the detection and performance improvement of dead pine trees of wilt nematode using the U-Net algorithm had a recall rate of -0.5%p than the U-Net model using the previously provided true values, precision was 7.6%p and F-1 score was 4.1%p. In the future, it is judged that there is a possibility to increase the precision of wilt detection by applying various filtering techniques, and it is judged that the drone surveillance method using drone orthographic images and artificial intelligence can be used in the pine wilt nematode disaster prevention project.

• Computers and Concrete
• /
• v.32 no.6
• /
• pp.607-613
• /
• 2023

The objective of this work is to determine the compressive strength of geopolymer concrete utilizing four distinct machine learning approaches. These techniques are known as gradient boosting machine (GBM), generalized linear model (GLM), extremely randomized trees (XRT), and deep learning (DL). Experimentation is performed to collect the data that is then utilized for training the models. Compressive strength is the response variable, whereas curing days, curing temperature, silica fume, and nanosilica concentration are the different input parameters that are taken into consideration. Several kinds of errors, including root mean square error (RMSE), coefficient of correlation (CC), variance account for (VAF), RMSE to observation's standard deviation ratio (RSR), and Nash-Sutcliffe effectiveness (NSE), were computed to determine the effectiveness of each algorithm. It was observed that, among all the models that were investigated, the GBM is the surrogate model that can predict the compressive strength of the geopolymer concrete with the highest degree of precision.

• Journal of the Korean Society of Visualization
• /
• v.22 no.1
• /
• pp.49-60
• /
• 2024

The purpose of this study establishes the deep neural network (DNN) and Decoder models to predict the flow and thermal fields of three-dimensional wavy wings as a passive flow control. The wide ranges of the wavy geometric parameters of wave amplitude and wave number are considered for the various the angles of attack and the aspect ratios of a wing. The huge dataset for training and test of the deep learning models are generated using computational fluid dynamics (CFD). The DNN and Decoder models exhibit quantitatively accurate predictions for aerodynamic coefficients and Nusselt numbers, also qualitative pressure, limiting streamlines, and Nusselt number distributions on the surface. Particularly, Decoder model regenerates the important flow features of tiny vortices in the valleys, which makes a delay of the stall. Also, the spiral vortical formation is realized by the Decoder model, which enhances the lift.