YCYeji Choi

Yeji Choi

*This is an algorithmically generated author record based on publicly available information.

Yeji Choi has a total of 54 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Generation of Rainfall Maps from GK2A Satellite Images Using Deep Learning

    by Yerim Lim, Yeji Choi, Eunbin Kim, Yong-Jae Moon, Hyun-Jin Jeong

    Accurate rainfall monitoring is essential for mitigating hydrometeorological disasters and understanding hydrological changes under climate change. This study presents a deep learning-based rainfall estimation framework using multispectral GEO-KOMPSAT-2A (GK2A) satellite imagery. The analysis primarily focuses on daytime observations to take advantage of visible channel information, which provides richer representations of cloud characteristics during daylight conditions. The core model, Model-HSP, is built on the Pix2PixCC architecture and trained with Hybrid Surface Precipitation (HSP) data from weather radar. To further enhance accuracy, an ensemble model (Model-ENS) integrates the outputs of Model-HSP and a radar based Model-CMX, leveraging their complementary strengths for improved generalization, robustness, and stability across rainfall regimes. Performance was evaluated over two periods—a one year period from May 2023 to April 2024 and the August 2023 monsoon season—at 2 km and 4 km spatial resolutions, using RMSE and CC as quantitative metrics. Case analyses confirmed the superior capability of Model-ENS in capturing rainfall distribution, intensity, and temporal evolution across diverse weather conditions. These findings show that deep learning greatly enhances GEO satellite rainfall estimation, enabling real-time, high-resolution monitoring even in radar sparse or limited coverage regions, and offering strong potential for global and regional hydrometeorological and climate research applications.


  • Open Access

    Seq2Seq LSTM Model Optimized for Urban Short‐Term Multivariate Predictions

    by Yeji Choi, Sang-Hyun Lee, Yong-Ki Ma

    Weather prediction has traditionally relied on numerical weather prediction (NWP) models. However, these models have encountered significant challenges, such as high computational demands and inherent prediction biases, particularly in high‐resolution urban simulations. Recent advancements in deep learning (DL) techniques have provided potential solutions to overcome these limitations. This study presents a Seq2Seq long short‐term memory (LSTM) model optimized for short‐term (up to 6 h) operational predictions of near‐ground temperature, relative humidity, wind speed, and wind direction in a highly urbanized area. The optimized LSTM model was constructed by testing various structures and hyperparameters (“ model-centric approach ”) and considering the effectiveness of different training datasets (“ data-centric approach ”) with a focus on jointly handling multivariate forecasting within a Seq2Seq LSTM framework. A 1‐month forecast evaluation showed that the Seq2Seq LSTM model improved upon the local data assimilation and prediction system (LDAPS), the Korea operational NWP model, with the most pronounced improvements observed in temperature (RMSE: ~1.4°C) and wind speed (RMSE: ~0.6 m s −1 ) forecasts, while improvements in relative humidity and wind direction were more limited. The model could predict more precisely time‐averaged trends of the forecast variables than their deviations, with deteriorated performance in abrupt meteorological changes. These results imply the essential role of NWP forecasts as input sequences in improving the model performance and reducing training data length.

  • Guided Super Resolution of Land Surface Temperature Using Multisatellite Imageries

    by Sunju Lee, Yeji Choi, Beomkyu Choi, Junghoon Seo, Minki Song, Eunha Sohn, Sewoong Ahn

    As understanding and monitoring global warming and heatwaves have become increasingly important, the demand for higher spatial and temporal resolution in satellite-observed Land Surface Temperature (LST) data has risen. LST derived from geostationary satellite observations plays a crucial role in temperature monitoring, offering high temporal resolution across wide areas. However, a primary limitation of geostationary satellite-derived LST products is their low spatial resolution. In this study, we aim to overcome this limitation by employing a deep learning-based super-resolution (SR) approach and propose a model called References and Residual in Residual Blocks to Perform Super Resolution Network (3R-Net). This model incorporates terrain-guided reference images and residual blocks to enable more accurate super-resolution. By using LST products from GEO-KOMPSAT-2A (GK2A) as low-resolution input, visible channel imagery from GEO-KOMPSAT-2B (GK2B) as terrain-reflective high-resolution reference, and Landsat-8 LST products as high-resolution target images, our model effectively enhances the 2 km resolution of GK2A LST products to the 500 m resolution of Landsat-8 with improved spatial detail and accuracy. Unlike many previous studies relying on single-image super-resolution with synthetically downsampled inputs, our approach uses real-world LR inputs, HR targets, and reference images, making the learning process more realistic and practical. Experimental results show that, when clear guidance is provided through reference images, 3R-Net surpasses existing SR methods, achieving higher PSNR, SSIM, and lower RMSE while capturing critical spatial and temporal features, including surface characteristics and daily heating patterns. By integrating residual-in-residual blocks our model achieves a simple yet powerful enhancement in capturing fine-grained spatial and temporal patterns. These advancements suggest that 3R-Net can provide enhanced LST data crucial for climate research, environmental monitoring, and early warning systems.


  • Probabilistic Weather Forecasting with Deterministic Guidance-Based Diffusion Model

    by Donggeun Yoon, Minseok Seo, Doyi Kim, Yeji Choi, Donghyeon Cho


  • Open Access

    Cloud Cover Prediction Model Using Multichannel Geostationary Satellite Images

    by Eunbin Cho, Eunbin Kim, Yeji Choi

    Cloud cover influences solar radiation reaching the Earth’s surface, impacting industries. Recently, advancements in weather prediction have been made through the use of satellite images and deep learning methods for enhancing the accuracy of cloud variability forecasts. Despite these advancements, computational limitations arise due to the large size of the satellite images. Although conventional practices involving the cropping or downscaling of images into smaller sizes have been employed, these processes have been observed to compromise the accuracy of the predicted images. In this study, we introduce Cloudstream, a novel approach that combines a CNN-based encoder, decoder, and PredRNN-V2 as a backbone model. This approach prioritizes computational efficiency while also maintaining prediction accuracy. Cloudstream predicted future cloud detection image data, training with the dataset of the sequential cloud detection and infrared channel images from the Korean Geostationary Meteorological Satellite GEO-KOMPSAT-2A. Additionally, we explored the utilization of non-patch images in the development of Cloudstream. A quantitative evaluation of the model was performed using two different input sizes for the same geographic area: 128×128 pixels and 512×512 pixels. There are no significant differences in F1 scores between Cloudstream and PredRNN-V2 when processing 128×128 inputs, however, Cloudstream required three times fewer floating-point operations than PredRNN-V2. Additionally, we found that 512×512 high-resolution input images exhibit superior prediction performance compared to 128×128 low-resolution input images. This study contributes to the refinement of deep learning-based video frame prediction models by focusing on optimizing satellite image prediction, addressing computational challenges.


  • Open Access

    Deterministic Guidance Diffusion Model for Probabilistic Weather Forecasting

    by Donggeun Yoon, Minseok Seo, Doyi Kim, Yeji Choi, Donghyeon Cho

    Weather forecasting requires not only accuracy but also the ability to perform probabilistic prediction. However, deterministic weather forecasting methods do not support probabilistic predictions, and conversely, probabilistic models tend to be less accurate. To address these challenges, in this paper, we introduce the \textbf{\textit{D}}eterministic \textbf{\textit{G}}uidance \textbf{\textit{D}}iffusion \textbf{\textit{M}}odel (DGDM) for probabilistic weather forecasting, integrating benefits of both deterministic and probabilistic approaches. During the forward process, both the deterministic and probabilistic models are trained end-to-end. In the reverse process, weather forecasting leverages the predicted result from the deterministic model, using as an intermediate starting point for the probabilistic model. By fusing deterministic models with probabilistic models in this manner, DGDM is capable of providing accurate forecasts while also offering probabilistic predictions. To evaluate DGDM, we assess it on the global weather forecasting dataset (WeatherBench) and the common video frame prediction benchmark (Moving MNIST). We also introduce and evaluate the Pacific Northwest Windstorm (PNW)-Typhoon weather satellite dataset to verify the effectiveness of DGDM in high-resolution regional forecasting. As a result of our experiments, DGDM achieves state-of-the-art results not only in global forecasting but also in regional forecasting. The code is available at: \url{https://github.com/DongGeun-Yoon/DGDM}.

  • Explainable AI-Based Interface System for Weather Forecasting Model

    by Soyeon Kim, Junho Choi, Yeji Choi, Subeen Lee, Artyom Stitsyuk, Minkyoung Park, Seongyeop Jeong, You-Hyun Baek, Jaesik Choi


  • Open Access

    Short-term forecasting of typhoon rainfall with a deep-learning-based disaster monitoring model

    by Doyi Kim, Yeji Choi, Minseok Seo, Seungheon Shin, Hyun-Jin Jeong

    Accurate and reliable disaster forecasting is vital for saving lives and property. Hence, effective disaster management is necessary to reduce the impact of natural disasters and to accelerate recovery and reconstruction. Typhoons are one of the major disasters related to heavy rainfall in Korea. As a typhoon develops in the far ocean, satellite observations are the only means to monitor them. Our study uses satellite observations to propose a deep-learning-based disaster monitoring model for short-term typhoon rainfall forecasting. For this, we consider two deep learning models: a video frame prediction model, Warp and Refine Network (WR-Net), to predict future satellite observations and an image-to-image translation model, geostationary rainfall product (GeorAIn) (based on the Pix2PixCC model), to generate rainfall maps from predicted satellite images. Typhoon Hinnamnor, the worst typhoon case in 2022 in Korea, is selected as a target case for model verification. The results show that the predicted satellite images can capture the structures and patterns of the typhoon. The rainfall maps generated from the GeorAIn model using predicted satellite images show a correlation coefficient of 0.81 for 3-hr and 0.56 for 7-hr predictions. The proposed disaster monitoring model can provide us with practical implications for disaster alerting systems and can be extended to flood-monitoring systems.


  • Open Access

    Improved Flood Insights: Diffusion-Based SAR to EO Image Translation

    by Minseok Seo, Youngtack Oh, Doyi Kim, Dongmin Kang, Yeji Choi

    Driven by rapid climate change, the frequency and intensity of flood events are increasing. Electro-Optical (EO) satellite imagery is commonly utilized for rapid response. However, its utilities in flood situations are hampered by issues such as cloud cover and limitations during nighttime, making accurate assessment of damage challenging. Several alternative flood detection techniques utilizing Synthetic Aperture Radar (SAR) data have been proposed. Despite the advantages of SAR over EO in the aforementioned situations, SAR presents a distinct drawback: human analysts often struggle with data interpretation. To tackle this issue, this paper introduces a novel framework, Diffusion-Based SAR to EO Image Translation (DSE). The DSE framework converts SAR images into EO images, thereby enhancing the interpretability of flood insights for humans. Experimental results on the Sen1Floods11 and SEN12-FLOOD datasets confirm that the DSE framework not only delivers enhanced visual information but also improves performance across all tested flood segmentation baselines.