• IEMEK Journal of Embedded Systems and Applications
• /
• 제17권5호
• /
• pp.297-307
• /
• 2022

The simple procedural segment selection algorithm commonly used in Dynamic Adaptive Streaming over HTTP (DASH) reveals severe weakness to provide high-quality streaming services in the integrated mobile networks of various wired and wireless links. A major issue could be how to properly cope with dynamically changing underlying network conditions. The key to meet it should be to make the segment selection algorithm much more adaptive to fluctuation of network traffics. This paper presents a system architecture that replaces the existing procedural segment selection algorithm with a deep reinforcement learning algorithm based on the Asynchronous Advantage Actor-Critic (A3C). The distributed A3C-based deep learning server is designed and implemented to allow multiple clients in different network conditions to stream videos simultaneously, collect learning data quickly, and learn asynchronously, resulting in greatly improved learning speed as the number of video clients increases. The performance analysis shows that the proposed algorithm outperforms both the conventional DASH algorithm and the Deep Q-Network algorithm in terms of the user's quality of experience and the speed of deep learning.

• Nuclear Engineering and Technology
• /
• 제55권2호
• /
• pp.506-515
• /
• 2023

This study is to find a method to learn engine sound after the start-up of a diesel generator installed in nuclear power plant with an unsupervised deep learning algorithm (CNN autoencoder) and a new method to predict the failure of a diesel generator using it. In order to learn the sound of a diesel generator with a deep learning algorithm, sound data recorded before and after the start-up of two diesel generators was used. The sound data of 20 min and 2 h were cut into 7 s, and the split sound was converted into a spectrogram image. 1200 and 7200 spectrogram images were created from sound data of 20 min and 2 h, respectively. Using two different deep learning algorithms (CNN autoencoder and binary classification), it was investigated whether the diesel generator post-start sounds were learned as normal. It was possible to accurately determine the post-start sounds as normal and the pre-start sounds as abnormal. It was also confirmed that the deep learning algorithm could detect the virtual abnormal sounds created by mixing the unusual sounds with the post-start sounds. This study showed that the unsupervised anomaly detection algorithm has a good accuracy increased about 3% with comparing to the binary classification algorithm.

• Journal of Software Assessment and Valuation
• /
• 제15권1호
• /
• pp.63-70
• /
• 2019

These days, as Interest in Image recognition with deep learning is increasing, there has been a lot of research in image recognition using deep learning. In this study, we propose a system for classifying rocks through rock images of 18 types of rock(6 types of igneous, 6 types of metamorphic, 6 types of sedimentary rock) which are addressed in the high school curriculum, using CNN model based on Tensorflow, deep learning open source framework. As a result, we developed a classifier to distinguish rocks by learning the images of rocks and confirmed the classification performance of rock classifier. Finally, through the mobile application implemented, students can use the application as a learning tool in classroom or on-site experience.

• Journal of the Korea Convergence Society
• /
• 제10권5호
• /
• pp.9-16
• /
• 2019

In this paper, it will analyze security problems, so technology's potential can apply to business security area. First, in order to deep learning do security tasks sufficiently in the business area, deep learning requires repetitive learning with large amounts of data. In this paper, to acquire learning ability to do stable business tasks, it must detect abnormal IP packets and attack such as normal software with malicious code. Therefore, this paper will analyze whether deep learning has the cognitive ability to detect various attack. In this paper, to deep learning to reach the system and reliably execute the business model which has problem, this paper will develop deep learning technology which is equipped with security engine to analyze new IP about Session and do log analysis and solve the problem of mathematical role which can extract abnormal data and distinguish infringement of system data. Then it will apply to business model to drop the vulnerability and improve the business performance.

• Smart Media Journal
• /
• 제12권11호
• /
• pp.134-144
• /
• 2023

Nowadays, Deep Learning (DL) technology is being used in several government departments. South Korea imports a lot of seafood. If the demand for fishery products is not accurately predicted, then there will be a shortage of fishery products and the price of the fishery product may rise sharply. So, South Korea's Ministry of Ocean and Fisheries is attempting to accurately predict seafood imports using deep learning. This paper introduces the solution for the fish import prediction in South Korea using the Long Short-Term Memory (LSTM) method. It was found that there was a huge gap between the sum of consumption and export against the sum of production especially in the case of two species that are Hairtail and Pollock. An import prediction is suggested in this research to fill the gap with some advanced Deep Learning methods. This research focuses on import prediction using Machine Learning (ML) and Deep Learning methods to predict the import amount more precisely. For the prediction, two Deep Learning methods were chosen which are Artificial Neural Network (ANN) and Long Short-Term Memory (LSTM). Moreover, the Machine Learning method was also selected for the comparison between the DL and ML. Root Mean Square Error (RMSE) was selected for the error measurement which shows the difference between the predicted and actual values. The results obtained were compared with the average RMSE scores and in terms of percentage. It was found that the LSTM has the lowest RMSE score which showed the prediction with higher accuracy. Meanwhile, ML's RMSE score was higher which shows lower accuracy in prediction. Moreover, Google Trend Search data was used as a new feature to find its impact on prediction outcomes. It was found that it had a positive impact on results as the RMSE values were lowered, increasing the accuracy of the prediction.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• 제19권3호
• /
• pp.28-37
• /
• 2020

This study developed a deep learning model that predicts rental demand for public bicycles. For this, public bicycle rental data, weather data, and subway usage data were collected. After building an exponential smoothing model, ARIMA model and LSTM-based deep learning model, forecasting errors were compared and evaluated using MSE and MAE evaluation indicators. Based on the analysis results, MSE 348.74 and MAE 14.15 were calculated using the exponential smoothing model. The ARIMA model produced MSE 170.10 and MAE 9.30 values. In addition, MSE 120.22 and MAE 6.76 values were calculated using the deep learning model. Compared to the value of the exponential smoothing model, the MSE of the ARIMA model decreased by 51% and the MAE by 34%. In addition, the MSE of the deep learning model decreased by 66% and the MAE by 52%, which was found to have the least error in the deep learning model. These results show that the prediction error in public bicycle rental demand forecasting can be greatly reduced by applying the deep learning model.

• Journal of the Korea Society of Computer and Information
• /
• 제28권10호
• /
• pp.17-25
• /
• 2023

In this paper, we propose a lightweight deep learning architecture tailored for efficient text detection in fashion design sketch images. Given the increasing prominence of Digital Transformation in the fashion industry, there is a growing emphasis on harnessing digital tools for creating fashion design sketches. As digitization becomes more pervasive in the fashion design process, the initial stages of text detection and recognition take on pivotal roles. In this study, a lightweight network was designed by building upon existing text detection deep learning models, taking into consideration the unique characteristics of apparel design drawings. Additionally, a separately collected dataset of apparel design drawings was added to train the deep learning model. Experimental results underscore the superior performance of our proposed deep learning model, outperforming existing text detection models by approximately 20% when applied to fashion design sketch images. As a result, this paper is expected to contribute to the Digital Transformation in the field of clothing design by means of research on optimizing deep learning models and detecting specialized text information.

• Journal of the Korean Society of Environmental Restoration Technology
• /
• 제26권6호
• /
• pp.85-96
• /
• 2023

This study proposes a method for forest vegetation monitoring using high-resolution aerial imagery captured by unmanned aerial vehicles(UAV) and deep learning technology. The research site was selected in the forested area of Mountain Dogo, Asan City, Chungcheongnam-do, and the target species for monitoring included Pinus densiflora, Quercus mongolica, and Quercus acutissima. To classify vegetation species at the pixel level in UAV imagery based on characteristics such as leaf shape, size, and color, the study employed the semantic segmentation method using the prominent U-net deep learning model. The research results indicated that it was possible to visually distinguish Pinus densiflora Siebold & Zucc, Quercus mongolica Fisch. ex Ledeb, and Quercus acutissima Carruth in 135 aerial images captured by UAV. Out of these, 104 images were used as training data for the deep learning model, while 31 images were used for inference. The optimization of the deep learning model resulted in an overall average pixel accuracy of 92.60, with mIoU at 0.80 and FIoU at 0.82, demonstrating the successful construction of a reliable deep learning model. This study is significant as a pilot case for the application of UAV and deep learning to monitor and manage representative species among climate-vulnerable vegetation, including Pinus densiflora, Quercus mongolica, and Quercus acutissima. It is expected that in the future, UAV and deep learning models can be applied to a variety of vegetation species to better address forest management.

• International Journal of Internet, Broadcasting and Communication
• /
• 제16권2호
• /
• pp.127-135
• /
• 2024

Recent studies have shown that inverse design using deep learning has the potential to rapidly generate the optimal design that satisfies the target performance without the need for iterative optimization processes. Unlike traditional methods, deep learning allows the network to rapidly generate a large number of solution candidates for the same objective after a single training, and enables the generation of diverse designs tailored to the objectives of inverse design. These inverse design techniques are expected to significantly enhance the efficiency and innovation of design processes in various fields such as aerospace, biology, medical, and engineering. We analyzes inverse design models that are mainly utilized in the nano and chemical fields, and proposes inverse design models based on supervised and unsupervised learning that can be applied to the engineering system. It is expected to present the possibility of effectively applying inverse design methodologies to the design optimization problem in the field of engineering according to each specific objective.

• Journal of Korea Multimedia Society
• /
• 제21권2호
• /
• pp.173-180
• /
• 2018

This paper proposes a method to improve the performance of face recognition via super-resolution method using sparse representation and deep learning from low-resolution facial images. Recently, there have been many researches on ultra-high-resolution images using deep learning techniques, but studies are still under way in real-time face recognition. In this paper, we combine the sparse representation and deep learning to generate super-resolution images to improve the performance of face recognition. We have also improved the processing speed by designing in parallel structure when applying sparse representation. Finally, experimental results show that the proposed method is superior to conventional methods on various images.