Image Details

Choose export citation format:

Data-driven Deep SwinTransUnet Enables the Prediction of 4K Future Solar Extreme Ultraviolet Images

  • Authors: Hao Li, Tao Shu, Zhou Chen

Hao Li et al 2026 The Astrophysical Journal Supplement Series 287 .

  • Provider: AAS Journals

Caption: Figure 6.

Results of image stitching and increased prediction frames, evaluated on the original DN scale. (a) Overall visualization of the stitched full-disk images, showing two consecutive predicted frames at 04:12 and 04:24 UT on 2010 December 31. For each frame, the ground truth (GT), Swin-4KEUV prediction, enlarged local regions, and residual maps are presented. The residual is computed as prediction minus GT, and the reported MSE, PSNR, and SSIM values are all calculated on the original DN scale. (b) Comparison of the first-frame prediction results under different prediction lengths (Ty = 3 to Ty = 8), including GT, Swin-4KEUV predictions, and residual maps, all shown on the original DN scale. (c) Boxplots of the MSE, PSNR, and SSIM for the first two predicted frames over the complete test set (approximately 22 data pairs from 2020 December 25 to December 31), with all metrics evaluated on the original DN scale.

Other Images in This Article
Copyright and Terms & Conditions

Additional terms of reuse