• Korean Journal of Remote Sensing
• /
• v.39 no.6_3
• /
• pp.1679-1692
• /
• 2023

South Korea is a country that emits a large amount of pollutants as a result of population growth and industrial development and is also severely affected by transboundary air pollution due to its geographical location. As pollutants from both domestic and foreign sources contribute to air pollution in Korea, the location of air pollutant emission sources is crucial for understanding the movement and distribution of pollutants in the atmosphere and establishing national-level air pollution management and response strategies. Based on this background, this study aims to effectively acquire spatial information on domestic and international air pollutant emission sources, which is essential for analyzing air pollution status, by utilizing high-resolution optical satellite images and deep learning-based image segmentation models. In particular, industrial parks and quarries, which have been evaluated as contributing significantly to transboundary air pollution, were selected as the main research subjects, and images of these areas from multi-purpose satellites 3 and 3A were collected, preprocessed, and converted into input and label data for model training. As a result of training the U-Net model using this data, the overall accuracy of 0.8484 and mean Intersection over Union (mIoU) of 0.6490 were achieved, and the predicted maps showed significant results in extracting object boundaries more accurately than the label data created by course annotations.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.6
• /
• pp.255-266
• /
• 2022

The quality control of coarse aggregate among aggregates, which are the main ingredients of concrete, is currently carried out by SPC(Statistical Process Control) method through sampling. We construct a smart factory for manufacturing innovation by changing the quality control of coarse aggregates to inspect the coarse aggregates based on this image by acquired images through the camera instead of the current sieve analysis. First, obtained images were preprocessed, and HED(Hollistically-nested Edge Detection) which is the filter learned by deep learning segment each object. After analyzing each aggregate by image processing the segmentation result, fineness modulus and the aggregate shape rate are determined by analyzing result. The quality of aggregate obtained through the video was examined by calculate fineness modulus and aggregate shape rate and the accuracy of the algorithm was more than 90% accurate compared to that of aggregates through the sieve analysis. Furthermore, the aggregate shape rate could not be examined by conventional methods, but the content of this paper also allowed the measurement of the aggregate shape rate. For the aggregate shape rate, it was verified with the length of models, which showed a difference of ±4.5%. In the case of measuring the length of the aggregate, the algorithm result and actual length of the aggregate showed a ±6% difference. Analyzing the actual three-dimensional data in a two-dimensional video made a difference from the actual data, which requires further research.

• Journal of the Institute of Convergence Signal Processing
• /
• v.24 no.1
• /
• pp.42-50
• /
• 2023

The development of technology has achieved digital innovation throughout the media industry, including production techniques and editing technologies, and has brought diversity in the form of consumer viewing through the OTT service and streaming era. The convergence of big data and deep learning networks automatically generated text in format such as news articles, novels, and scripts, but there were insufficient studies that reflected the author's intention and generated story with contextually smooth. In this paper, we describe the flow of pictures in the storyboard with image caption generation techniques, and the automatic generation of story-tailored scenarios through language models. Image caption using CNN and Attention Mechanism, we generate sentences describing pictures on the storyboard, and input the generated sentences into the artificial intelligence natural language processing model KoGPT-2 in order to automatically generate scenarios that meet the planning intention. Through this paper, the author's intention and story customized scenarios are created in large quantities to alleviate the pain of content creation, and artificial intelligence participates in the overall process of digital content production to activate media intelligence.

• Korean Journal of Radiology
• /
• v.22 no.11
• /
• pp.1764-1776
• /
• 2021

Objective: This study aimed to validate a deep learning-based fully automatic calcium scoring (coronary artery calcium [CAC]_auto) system using previously published cardiac computed tomography (CT) cohort data with the manually segmented coronary calcium scoring (CAC_hand) system as the reference standard. Materials and Methods: We developed the CAC_auto system using 100 co-registered, non-enhanced and contrast-enhanced CT scans. For the validation of the CAC_auto system, three previously published CT cohorts (n = 2985) were chosen to represent different clinical scenarios (i.e., 2647 asymptomatic, 220 symptomatic, 118 valve disease) and four CT models. The performance of the CAC_auto system in detecting coronary calcium was determined. The reliability of the system in measuring the Agatston score as compared with CAC_hand was also evaluated per vessel and per patient using intraclass correlation coefficients (ICCs) and Bland-Altman analysis. The agreement between CAC_auto and CAC_hand based on the cardiovascular risk stratification categories (Agatston score: 0, 1-10, 11-100, 101-400, > 400) was evaluated. Results: In 2985 patients, 6218 coronary calcium lesions were identified using CAC_hand. The per-lesion sensitivity and false-positive rate of the CAC_auto system in detecting coronary calcium were 93.3% (5800 of 6218) and 0.11 false-positive lesions per patient, respectively. The CAC_auto system, in measuring the Agatston score, yielded ICCs of 0.99 for all the vessels (left main 0.91, left anterior descending 0.99, left circumflex 0.96, right coronary 0.99). The limits of agreement between CAC_auto and CAC_hand were 1.6 ± 52.2. The linearly weighted kappa value for the Agatston score categorization was 0.94. The main causes of false-positive results were image noise (29.1%, 97/333 lesions), aortic wall calcification (25.5%, 85/333 lesions), and pericardial calcification (24.3%, 81/333 lesions). Conclusion: The atlas-based CAC_auto empowered by deep learning provided accurate calcium score measurement as compared with manual method and risk category classification, which could potentially streamline CAC imaging workflows.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.5
• /
• pp.21-29
• /
• 2024

Reducing the radiation dose during CT scanning can lower the risk of radiation exposure, but not only does the image resolution significantly deteriorate, but the effectiveness of diagnosis is reduced due to the generation of noise. Therefore, noise removal from CT images is a very important and essential processing process in the image restoration. Until now, there are limitations in removing only the noise by separating the noise and the original signal in the image area. In this paper, we aim to effectively remove noise from CT images using the wavelet transform-based GAN model, that is, the WT-GAN model in the frequency domain. The GAN model used here generates images with noise removed through a U-Net structured generator and a PatchGAN structured discriminator. To evaluate the performance of the WT-GAN model proposed in this paper, experiments were conducted on CT images damaged by various noises, namely Gaussian noise, Poisson noise, and speckle noise. As a result of the performance experiment, the WT-GAN model is better than the traditional filter, that is, the BM3D filter, as well as the existing deep learning models, such as DnCNN, CDAE model, and U-Net GAN model, in qualitative and quantitative measures, that is, PSNR (Peak Signal-to-Noise Ratio) and SSIM (Structural Similarity Index Measure) showed excellent results.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.3
• /
• pp.67-74
• /
• 2024

In this paper, we present a study aimed at analyzing how different rainfall measurement methods affect the performance of reservoir water level predictions. This work is particularly timely given the increasing emphasis on climate change and the sustainable management of water resources. To this end, we have employed rainfall data from ASOS, AWS, and Thiessen Network-based measures provided by the KMA Weather Data Service to train our neural network models for reservoir yield predictions. Our analysis, which encompasses 34 reservoirs in Jeollabuk-do Province, examines how each method contributes to enhancing prediction accuracy. The results reveal that models using rainfall data based on the Thiessen Network's area rainfall ratio yield the highest accuracy. This can be attributed to the method's accounting for precise distances between observation stations, offering a more accurate reflection of the actual rainfall across different regions. These findings underscore the importance of precise regional rainfall data in predicting reservoir yields. Additionally, the paper underscores the significance of meticulous rainfall measurement and data analysis, and discusses the prediction model's potential applications in agriculture, urban planning, and flood management.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.63 no.6
• /
• pp.77-88
• /
• 2021

Duck industry had a rapid growth in recent years. Nevertheless, researches to improve duck house environment are still not sufficient enough. Moisture generation of duck house litter is an important factor because it may cause severe illness and low productivity. However, the measuring process is difficult because it could be disturbed with animal excrements and other factors. Therefore, it has to be calculated according to the environmental data around the duck house litter. To cut through all these procedures, we built several machine learning regression model forecasting moisture generation of litter by measured environment data (air temperature, relative humidity, wind velocity and water contents). 5 models (Multi Linear Regression, k-Nearest Neighbors, Support Vector Regression, Random Forest and Deep Neural Network). have been selected for regression. By using R-Square, RMSE and MAE as evaluation metrics, the best accurate model was estimated according to the variables for each machine learning model. In addition, to address the small amount of data acquired through lab experiments, bootstrapping method, a technique utilized in statistics, was used. As a result, the most accurate model selected was Random Forest, with parameters of n-estimator 200 by bootstrapping the original data nine times.

• KIPS Transactions on Computer and Communication Systems
• /
• v.10 no.3
• /
• pp.71-80
• /
• 2021

The selection of an appropriate neural network algorithm is an important step for accurate data prediction in machine learning. Many algorithms based on basic artificial neural networks have been devised to efficiently predict future data. These networks include deep neural networks (DNNs), recurrent neural networks (RNNs), long short-term memory (LSTM) networks, and gated recurrent unit (GRU) neural networks. Developers face difficulties when choosing among these networks because sufficient information on their performance is unavailable. To alleviate this difficulty, we evaluated the performance of each algorithm by comparing their errors and processing times. Each neural network model was trained using a tax dataset, and the trained model was used for data prediction to compare accuracies among the various algorithms. Furthermore, the effects of activation functions and various optimizers on the performance of the models were analyzed The experimental results show that the GRU and LSTM algorithms yields the lowest prediction error with an average RMSE of 0.12 and an average R2 score of 0.78 and 0.75 respectively, and the basic DNN model achieves the lowest processing time but highest average RMSE of 0.163. Furthermore, the Adam optimizer yields the best performance (with DNN, GRU, and LSTM) in terms of error and the worst performance in terms of processing time. The findings of this study are thus expected to be useful for scientists and developers.

• Smart Media Journal
• /
• v.12 no.1
• /
• pp.9-16
• /
• 2023

Recently, a convolutional neural network (CNN) based system is being developed to overcome the limitations of human resources in the apple quality classification of farmhouse. However, since convolutional neural networks receive only images of the same size, preprocessing such as sampling may be required, and in the case of oversampling, information loss of the original image such as image quality degradation and blurring occurs. In this paper, in order to minimize the above problem, to generate a image patch based graph of an original image and propose a random walk-based positional encoding method to apply the graph transformer model. The above method continuously learns the position embedding information of patches which don't have a positional information based on the random walk algorithm, and finds the optimal graph structure by aggregating useful node information through the self-attention technique of graph transformer model. Therefore, it is robust and shows good performance even in a new graph structure of random node order and an arbitrary graph structure according to the location of an object in an image. As a result, when experimented with 5 apple quality datasets, the learning accuracy was higher than other GNN models by a minimum of 1.3% to a maximum of 4.7%, and the number of parameters was 3.59M, which was about 15% less than the 23.52M of the ResNet18 model. Therefore, it shows fast reasoning speed according to the reduction of the amount of computation and proves the effect.

• Journal of the Institute of Convergence Signal Processing
• /
• v.24 no.3
• /
• pp.134-139
• /
• 2023

Rehabilitators perform outpatient treatment and daily rehabilitation exercises to recover physical function with the aim of quickly returning to society after surgical treatment. Unlike performing exercises in a hospital with the help of a professional therapist, there are many difficulties in performing rehabilitation exercises by the patient on a daily basis. In this paper, we propose a CNN-LSTM-based upper limb rehabilitation real-time monitoring system so that patients can perform rehabilitation efficiently and with correct posture on a daily basis. The proposed system measures biological signals through shoulder-mounted hardware equipped with EMG and IMU, performs preprocessing and normalization for learning, and uses them as a learning dataset. The implemented model consists of three polling layers of three synthetic stacks for feature detection and two LSTM layers for classification, and we were able to confirm a learning result of 97.44% on the validation data. After that, we conducted a comparative evaluation with the Teachable machine, and as a result of the comparative evaluation, we confirmed that the model was implemented at 93.6% and the Teachable machine at 94.4%, and both models showed similar classification performance.