이 레포지터리는 제가 출판한(published) 논문에서 사용한 코드들을 정리해 둔 것입니다. 코드는 제가 작성했으며, AI의 검수를 통해 정리·정돈했습니다.
This repository organizes the code used in my published papers. The code was written by me and curated with the help of AI review.
| Directory | Publication | Topic | Status |
|---|---|---|---|
euv-pixel-translation/ |
ApJS 264, 33 (2023) | EUV pixel-level translation using FCN | Implemented |
farside-magnetogram/ |
Nature Astronomy 3, 397 (2019) | Far-side magnetogram generation | Implemented |
flare-forecasting/ |
ApJ 869, 91 (2018) | Solar flare prediction | Implemented |
magnetogram-denoising/ |
ApJL 891, L4 (2020) | Magnetogram noise reduction | Implemented |
magnetogram-to-euv/ |
ApJL 884, L23 (2019) | Magnetogram to EUV image translation | Implemented |
Paper: Park et al. (2023), "Pixel-to-pixel Translation of Solar Extreme-ultraviolet Images for DEMs by Fully Connected Networks", ApJS, 264, 33
Summary: Pixel-to-pixel image translation among solar EUV images using Fully Connected Networks. Translates SDO/AIA 3 channels (17.1, 19.3, 21.1 nm) to other 3 channels (9.4, 13.1, 33.5 nm) for DEM analysis.
Network: FCN-based encoder-decoder with U-Net style skip connections
Paper: Park et al. (2019), "Generation of Solar UV and EUV Images from SDO/HMI Magnetograms by Deep Learning", ApJL, 884, L23
Summary: Deep learning-based image-to-image translation from SDO/HMI magnetograms to SDO/AIA UV/EUV images (9 passbands). Compares L1-only model with L1+cGAN model.
Network: Pix2Pix (U-Net Generator + PatchGAN Discriminator)
Paper: Kim, Park et al. (2019), "Solar farside magnetograms from deep learning analysis of STEREO/EUVI data", Nature Astronomy, 3, 397
DOI: 10.1038/s41550-019-0711-5
Summary: Generation of solar far-side magnetograms from STEREO/EUVI 304nm images using cGAN. Trained on near-side SDO/AIA 304nm and SDO/HMI pairs, then applied to far-side STEREO/EUVI data.
Network: Pix2Pix (U-Net Generator + PatchGAN Discriminator)
Paper: Park et al. (2018), "Application of the Deep Convolutional Neural Network to the Forecast of Solar Flare Occurrence Using Full-disk Solar Magnetograms", ApJ, 869, 91
Summary: CNN-based binary classification for solar flare prediction (≥C1.0 class) using full-disk magnetograms from SOHO/MDI and SDO/HMI without preprocessing.
Network: GoogLeNet + DenseNet combination (proposed model with inception-style blocks and dense connectivity)
Paper: Park et al. (2020), "De-noising SDO/HMI Solar Magnetograms by Image Translation Method Based on Deep Learning", ApJL, 891, L4
Summary: DCGAN-based noise reduction for SDO/HMI magnetograms. Translates single noisy magnetograms to de-noised equivalents (21-frame stacked quality). Reduces noise level from 8.66 G to 3.21 G.
Network: Pix2Pix / DCGAN (U-Net Generator + PatchGAN Discriminator)
- Python 3.8+
- PyTorch 1.9+
- NumPy
- SunPy
- AIAPy (for EUV data processing)
| Package | Description | Installation |
|---|---|---|
| egghouse | Personal utility library for solar physics research | pip install git+https://github.com/eunsu-park/egghouse.git |
egghouse provides shared utilities used across all projects:
| Module | Usage |
|---|---|
egghouse.config |
Configuration management (BaseConfig class for YAML/CLI support) |
egghouse.sdo |
SDO/AIA, SDO/HMI data processing |
egghouse.image |
Image processing utilities |
# Install from GitHub
pip install git+https://github.com/eunsu-park/egghouse.git
# Or for development
git clone https://github.com/eunsu-park/egghouse.git
cd egghouse
pip install -e .MIT License
이 저장소의 콘텐츠는 부분적으로 AI(Claude)의 도움을 받아 작성되었습니다. Some content in this repository was created with the assistance of AI (Claude).
- 일부 학습 자료는 AI가 생성하고 저자가 검토함 Some study materials were generated by AI and reviewed by the author
- 일부는 저자가 작성하고 AI가 검토함 Some were written by the author and reviewed by AI
Eunsu Park