Thursday, 1 May 2014

Summary of Weekly Meeting - (8): Co-meeting with Matina

I. Memo for the 1st component:
   
    - It is fine but not robust enough under rotation and viewpoint change: 
      try to use some rotation invariant descriptors to substitute the SURF descriptor (...SIFT is one of the common one to try)
    - Don't care too much about the number of matches, the number of 'correct' matches is more important
    - Don't worried about the speed of time, this is the big constraint for now, i.e., offline process (...So this is one of the reason that SURF descriptor can/should be substituted)
    - Fine tune for this 1st component while I am available (... solve the issue in rotation invariant)

    *Suggestion:
      Keep going to the 2nd component, then back to 1st component to do some improvement. Time is limited, and current algorithm somehow is fine with a main issue of rotation invariant. Moreover, this step may take ages to try and test in order to find the best/acceptable solution.

II. Memo for the 2nd component:

     - This is for 2 purposes:
        1. Do 2D to 3D integration
        2. As the most simple way to do matching under 'RST invariant'... which can be a ground truth method to improve/confirm the result and performance of my proposed method (i.e., the 1st component)
        3. The Strength in using markers:
            - The most common way in MIS
            - To locate an object without considering its properties such as texture
            - A group of markers which can circle out an object
            - The group of markers can be moved to simulate Affine transformation

     - Tools (...use one of them):
        1. Color marks: red, yellow, blue, green (...patterns is not necessary)
        2. A white paper with painted color markers: red, yellow, blue, green
        3. Directly assign value into 2 matrix 
        4. Directly choose the information of location of points in 2 matrix

     - Main steps:
        1. Color segmentation
        2. Estimate Homography (i.e., Affine Transformation): H = [R T]
        3. Kindly note from Matina and George:



       *Suggestion:
         - Check the idea and method of homography in MATLAB / Source Code
         - Check the book - The Bible of OpenGL
         - Asking people
       

No comments:

Post a Comment