Help me with my master thesis. It takes you only 10-15 minutes!

[NewsBot]

I need some help for my master thesis.*I'm creating an algorithm to detect skin and I want to be able to compare my algorithm with the results on how humans classify skin.

 

Therefore, I have taken images and cut them into pieces. I have done that because humans usually classify skin via the context; and I needed to remove that exact context.

 

What does "classify by context" mean? It means that we see a person and know it's a person and therefore know that this person has skin. We are then able to find the skin easier.

 

But*detecting people is*already a very hard task for a computer. For now there's no algorithm that always finds a person in an image.

 

So far so good. I have cut the images into pieces and created a website that gives you exactly these pieces (in three different sizes). On the website you will find also four buttons saying: No Skin, 1/3 skin, 2/3 skin and All skin. Click on the button that gives you the best approximation. 1/3 skin, for example, means that you think that 1/3 of the pixels in the image on the page are skin.

 

Also: you don't need to try to "see" things!*The image you see might make no sense at all.*Let yourself guide by the colors and the first impression.

 

To help me click on this link: www.liensberger.it/master/ and categorize between 200-500 images. You can also categorize less. Each vote counts and I'm happy for all votes that I get...

 

Thank you very much!

 

More...

 

View All Our Microsoft Related Feeds