Help, should i be using compare with PSNR?

[petalpetra]

Hi fmw42 and snibgo!

Thank guys! You are great and definitely deserve acknowledgement if this ever gets published haha!

I will try what you guys have suggested... I may be back with more Q's!

fmw42: If we have the amount of pixels for white bone calculated from a blank diagram, and then i calculate the amoutn of white pixels once a bone has been blacked out, then can't i just divide the two and get a percentage of bone? Why would you suggest transparent background (not that i know how to even do that)? This is all based on the premise that i am not currently using colour, but actually would like to, to represent different things.

[fmw42]

Transparent pixels are not included in the statistics. I assumed that your image had, bone, missing bone (black), and background black. I was concerned that you were not distinguishing between background and missing bone. If this diagram is just bone and background and no missing bone, then your concept should be fine. Get the total pixels from the black and white full skeleton image. Use the same size and matching image but with some bone missing. Get the total white pixels from that. If the two image are not the same size and scale skeleton, then this concept does not work. The image should be in registration and be the same size for most accurate results.

percent missing bone = 100 * ( (total white pixels from full bone image) - (total white pixels from missing bone image) ) / (total white pixels from full bone image)