Sea-Thru uses photogrammetry: it makes decisions about how far objects are away from the camera and other environmental factors before it corrects the scene, and it corrects each part of the scene differently as a result. It’s tested against reality (i.e. it’s trained on underwater scenes before and after turbid water is drained)
All I did was adjust the whole image using the color correction sliders. No science at all, just whatever made it look “right”.
Science will always have to contend with aesthetic preference, especially when the aesthetic is contrived to make the science look bad. 
