I have some doubts related to some of the optical flow implementations.

  1. How to compute angular error for optical flow algorithms? I know the formula for computing,

I have the computed flow field $\vec{v}_c$. How to get the exact flow field $\vec{v}_e$, ie the actual movement of pixels between frames ?

  1. I am trying to implement Nagel's Paper for optical flow computation based on the discussion given in here. What I don't get is how to compute the weight matrix $W$. $W$ is given as

$W=(I_x^2+I_y^2+2\delta)^{-1}\begin{pmatrix}I_y^2+\delta &-I_xI_y\\ -I_xI_y & I_x^2+\delta \end{pmatrix}$

Assuming that the image dimensions are $m\times n$ the first matrix on the RHS has dimension $m\times n$, but the second block matrix is of dimension $2m\times 2n$. What am I doing wrong ?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.