hey ... I'm doing a prj to identify skin color from image. So for tht u'll get a RGB value.. But for my expert system i need to catogorize whther the skin color is fair or dark etc..
So how can u get a range for a particular color, as a color is a combination of Red, Green and Blue.. Ex, Fair skin can be of varing RGB values.. so how to give a range that if RGB from say 10 to 100, take it as Fair..
Don't knw whether my explanation is clear.. Any help is appreciated.. Tnx
I suppose color could be part of a biometric identification scheme. There are plenty of statistical tools that get applied to numeric classification problems. The difficulties I see: 1. locating "skin" pixels in a larger image 2. compensating for illumination - daylight vrs artificial etc etc 3. normalizing absolute RGB values to fit the statistical tool
A google search for "numeric classification algorithms" finds lots of interesting sites. Bill
First, a bit of business: you may not have read our naming policy on the way in. It requires that you use a full, real (sounding) first and last name for your display name. Single names, joke names and "handles" are not acceptable at the Ranch. You can change your display name here. Thanks!
This reminds me of a talk I went to a while back about an application used by police departments to track major crimes, especially inbetween departments. The had long lists of codes for identifying physical aspects of the perps, including code for describing their hair: bald, receding, slightly receding (never understood the difference between those two, apart from ego!) etc... I immediately noticed that there was no code for "mullet"! I'd like to think my suggestion has lead to faster arrest rates in rural Ontario