1. Use Chehra Tracker on stereo pair facial image.
Note: you can use toys
2. Apply stereo depth finding algorithms and disparity algorithms ( state of art algorithms )
3. Get depth of noise by using chehra tracker points and then but artificial noise on it and see what is the difference in depth.
more explanation:
1. Set up your two camera with fixed baseline and known camera parameters
2. Take right and left images of the objective ( it should be human face or toys face
Or
Use any of stereo data set available online one of them called (Middlebury dataset)
[login to view URL]
3. Establish correspondence point between two images use stereo matching algorithms table below might be helpful for you
4. Get disparity map and depth map for step 3
Section 2
Repeat all steps above by using chehra algorithms
sir i can help u with this project i have done masters in computer engg with specilization in image processing and pattern recognition this is a dummy bid i can give u an exact estimate after we discuss the project hopping for your kind consideration