Recognition of skin based on B,G,R values stored in a text file?

Hey all,

For my current project I have to create a MatLab code that will check through a set of 245057 R,G,B values provided in a text file to me and detect which ones are skin samples and which are not while using fuzzy logic.

I am struggling to get started and as such have only managed to import the data stored in the text file into my workspace. If anyone can point me to an example of something the same or similar, or simply clue me in on what to do then that would be greatly appreciated. I am currently at odds with this software needless to say :(.

[Image Analyst]

What do you have as a training set that you declare for a fact that the colors in the training set are truly skin? You need to know what is defined as skin, otherwise all you have is 245057 random colors.

[Sam]

As far as I know and based on what my lecturer has said we don't have a readily accessible training set. Although, he did say I need to create a decision tree. This is where the dataset is from

http://archive.ics.uci.edu/ml/datasets/Skin+Segmentation

I'm really stumped!