Why use Camera Projection instead of Image Texture

Matte painters (usually using Nuke) seem to love using camera projection to texture models, often just a card or simple geo if a camera move is involved. I can’t work out why they don’t just use a regular image texture for the geo. What am I missing?

Doing some camera mapping is handy when you’ve got a real photo to match or a filmed shot.

I can see how this can be useful in matte painting if for instance, you export several layers from photoshop and map them using the same camera. No need to do some guesswork, every layer will be aligned the same way. And if you look from the camera you’ve got the perfect spot to watch the matte , matching 100% the photoshop file.

All this doesn’t mean you should always use it, especially if you work in 3D and go beyond planes and simple geometry. I guess it’s kind of a fast way to have something that match the photoshop file.