How can i apply RANSAC on the matched points of the SURF meathod

Hello everyone,

I have a video segemented into frames where i am comparing each frame of the video with a one constant picture.

I am using the method detectSURFFeatures()

How by ploting the matched features on the pictures,i noticed that i have arround 10% of the data that are wrong matches

I have read that the RANSAC method can help removing the outliers but so far, I had not lack to find a good example

here is the code snipped i am using (it might give you a better idea on what i did ) :

fixationImagePoints = detectSURFFeatures(fixationImage);

originalImagePoints = detectSURFFeatures(originalImage);

[fFixationImagePoints, vFixationImagePoints] = extractFeatures(fixationImage, fixationImagePoints);

[fOriginalImagePoints, vOriginalImagePoints] = extractFeatures(originalImage, originalImagePoints);

indexPairs = matchFeatures(fFixationImagePoints, fOriginalImagePoints) ;

fixationImageMatchedPoints = vFixationImagePoints(indexPairs(:, 1));

originalImageMatchedPoints = vOriginalImagePoints(indexPairs(:, 2));

Hopefully someone here can give me a clue on how to use the RANSAC or give me a different approache.

Thanks in advance.

Answer 1

[Puru Kathuria]

Hi,

You might want to look at this video tutorial which explains feature detection, extraction, and matching followed by an estimation of the geometric transformation using the RANSAC algorithm.