Testing cases of our proposed models: the top row displays results from Model A for ARs, and the bottom row shows results from Model B for the QS. The columns from left to right represent (1) the observed intensity used as the training label (Obs); (2) the predicted 10830 Å photospheric signal generated by the model (Gen); (3) the ratio map indicating relative deviations (Obs/Gen); (4) the distribution of the Obs/Gen values representing the model’s prediction accuracy; and (5) the intensity values of pixels sampled along the diagonal from the bottom left to the top right of the images.