High Resolution Solar Image Generation Using Generative Adversarial Networks

We applied Generative Adversarial Networks (GANs) to perform solar image-to-image translation — from Solar Dynamics Observatory (SDO)/Helioseismic and Magnetic Imager (HMI) line-of-sight magnetogram images to SDO/Atmospheric Imaging Assembly (AIA) 0304-Å images. UV/EUV observations like SDO/AIA 0304-Å were only available starting in the late 1990s, even though magnetic field observations like SDO/HMI have been available since the 1970s. By leveraging GANs, we can give scientists access to more complete datasets for analysis, using the Pix2PixHD and Pix2Pix algorithms trained and tested on 2012–2014 data. Our models generate high-resolution (1024×1024 pixel) AIA0304 images from HMI magnetograms, with a pixel-to-pixel Pearson Correlation Coefficient as high as 0.99 (Pix2PixHD) and 0.962 (Pix2Pix), outperforming prior work. To our knowledge, this is the first attempt to leverage Pix2PixHD for SDO/HMI-to-SDO/AIA0304 translation.

Recommended citation: Dash, A., Ye, J., Wang, G., & Jin, H. (2024). "High Resolution Solar Image Generation Using Generative Adversarial Networks." Annals of Data Science, 11(5), 1545–1561.
Download Paper