It’s easy to get a shaded image from a decal and a normal map. But doing the inverse work would be very useful for creating decals and normal maps from photos for example. Do such tools already exist? What method should be used to make this tool?
Something like that would involve lots of photos with different lighting, and some analysis. You would obviously need some equipment and a software for this.
I’m reminded of PTM here. http://www.hpl.hp.com/ptm/
As featured in the ATI biased book - ShaderX (http://www.shaderx.com)
The chap generates a normal map as part of his processes. Limited but good enough for his app.