hachreak
commented
2 weeks ago Hi @rose-jinyang sorry for the long delay but github did not send me an email when a new issue is opened.. T___T Thanks for your interest about our project.
The model has been trained on 2 datasets (sen2venus and oli2msi), where 4 bands are used for 2x, 3 bands are used for 3x and 4 bands are used for 4x. More on detail:
- For 2x upscaling, we used Sen2Venμs dataset: B2, B3, B4, B8 sentinel2 and B3, B4, B7, B11 Venμs bands, 10m and 5m resolution, respectively.
- For 3x upscaling, we used OLI2MSI dataset: blue, green, red from Landsat-8 and B2, B3, B4 from Sentinel2 bands, 30m and 10m resolution, respectively.
- For 4x upscaling, we used Sen2Venμs dataset: B5, B6, B7, B8A sentinel2 and B8, B9, B10, B11 Venμs bands, 20m and 5m resolution.
Let me know if you have more questions.
Hello How are you? What is the format of input image to this model? Is it a general RGB image?