When dropping an image into the preview window, the map for the current window type is computed, then the view switches to the "input image" tab, and when the user goes back to the map that was already computed, it is computed again.