reference:Hign-Speed Tracking with Kernelzied Correlation Filters

questions:

The core componet of most modern trackers is a discriminative classifier, tasked with distingushing between the target and the surrounding environment. To cope with natural image changes, this classifier is typically trained with translated and scaled sample patches. Such sets of samples are riddled with redundancies--any overlapping pixels are constrained to be the same.

solutions:

we proposed an analytic model for datasets of thousands of translated patches. By showing that the resulting data matrix is circulant, we can diagonalize it with the discrete Fourier transform, reducing both storage and compution by several orders of magnitude. Interestingly,

linear regression our  formutlation=a correlation filter

　　which is used by some of the fastest competitive trackers.

for kernel regression,

kernel regression=a new kernelized correlation filter(KCF)

　　which unlike other kernel algorithms has the exact same complexity as its linear counterpart.

Building on ti ,we also propose a fast multi-channel extension of linear correlation filters, via a linear kernel, which we call dual correlation filter(DCT).

see, as the topic demonstrates--high-speed tracking, focus on storage and computation.