Recently I played with simple steganography techniques and made a program that can encode an image into the lowest 4 bits of a carrier image. Although 50% of the information/detail from both pictures was lost, it wasn't that noticeable. Couldn't this be used in computer graphics?
To be more specific, is there some process in computer graphics, where it would make sense to lose 50% detail and save 50% of space (memory/storage) but still keep, lets say, ~80% of the visible detail?
What I thought so far about was storing mipmaps in the original picture or maybe a normal map on top of the original. Sadly I don't know that much about computer graphics, and it may be that I am missing something or that it doesn't work like that.
Any ideas or feedback would be greatly appreciated!