High-Speed Tracking with Kernelized Correlation Filters
Linear regression
Circulant matrices
So, we can use circulant matrice X to represent the vector x.
Compute w^∗
However, this formula isn’t used in the code. The following one is used.
kernel trick
k is the kernel function (Gaussian or Polynomial)
then we need to compute
compute α
compute the response map of current frame
z represent the current frame
use circulant matrice K to represent the kernel correlation
summary
- image patch => hog feature
- hog feature => fft
- kf = gaussian_correlation(xf, xf, kernel.sigma); (compute the kernel correlation of the first frame)
- alphaf = yf ./ (kf + lambda); (compute
α - kzf = gaussian_correlation(zf, model_xf, kernel.sigma); (second frame)
response = real(ifft2(model_alphaf .* kzf));
as illustrated in the formula