3 skin detection questions need answers

Hello guys! This is an $10 usd project for answering me some questions. The deliverable is text. Maybe I will hire the freelancer who helps me later to implement some matlab code. I planning to create a skin mole/nevus detection algorithm for skin diseases. I have 3 datasets for face,skin tone and nevus detection so i need to feed the algorithms with. Firstly I attach the skin dataset ([login to view URL]) below together with 3 pdfs so we create the algorithm for skin detection. I think a good approach would be to check first in this dataset for the extreme values of the pixels YCbCr first and then to combine rgb, hsv to make the skin detection algorithm. Is that a good idea? Im open to your idea on that. I have these 3 questions i need answered for $10 USD from a expert on this field. I may need his help later on the implementation of some part of the code for more money later.I just need some help at the start 1. How do you suggest me doing the training? In the dataset that have sent there are pictures of colorful people that made me difficult - for the time I suppose we have to fall into the model a caucassius human. Finding the extreme values of colour unfortunately did not help me if you have some solution on this? 2. How do you have in mind that it is better to make the detection? I thought no image would be discarded any picture would still be considered to be simply not to control all the non-skin pixels just by the regions with a different color from black. If you have a better idea, I will be happy to hear it (all for simple skin detection and on the way we see) 3. I wrote some matlab code ([login to view URL] & [login to view URL]) about the bordering and i think it is a good start. The point is that the process is done for all the pictures of a folder together. If the folder has just one image its OK, but if it has more images the code presents them all in just one figure. Please if you can check the code and say your suggestion on how to move forward. I send some of the Links I also found: [login to view URL] [login to view URL] [login to view URL] I will be happy hearing from you and working on a long term maybe Thank you a lot in advance