If You're working with different modalities like infrared or X-Ray, you should know the original color space of the image. In image processing 8-bit images are the standard one, but in different modalities like infrared or X-Ray it could be 16-bit or even more. It's possible that the first thing you're going to do is squeezing it to the 8-bit form. It will make visualization and general processing much easier, but you can lose much information in that way and make your goal much harder to achieve.
If for example you'd like to detect some really hot objects on the image, it's possible it's going to be really easy on 16-bit images, but quite challenging in the 8-bit color space.