No BSD License
22 Jan 2009
23 Jan 2009)
An implementaion of the very classical optical flow method of Horn & Schunck.
Watch this File
An implementaion of the very classical optical flow method of Horn & Schunck according to their paper:
Horn, B.K.P., and Schunck, B.G., Determining Optical Flow, AI(17), No. 1-3, August 1981, pp. 185-203
very good job
What type of velocity are these vectors representing on the image? Are they pixel velocities? If so, how do I go about averaging them and converting these values to something useful such as speed?
thanks for sharing!
Uhm, Thanks for your code Sharing. I would like convert 2D video into 3D video by using optical flow. How can i use those optical flows in order to convert into depth map?
May I have a look?Ha Ha
Hi. Thanks for your nice code. Can you explain How did you find [-1 1; -1 1],... in computeDerivates.m file? thanks
Are you sure that the derivatives in the computeDerivates.m file are correct? Matlab stores images with the y-axis in the first dimension and the x-axis in the second, so I think that fx and fy might be the wrong way around...
how can i change the optical flow from pixel by pixel to 16x16?
Now includes the computeDerivatives function.