TL;DR
This paper introduces a fast, kernelized correlation filter for object tracking that leverages circulant matrices and Fourier transforms to achieve high speed and accuracy, outperforming existing methods.
Contribution
It develops a novel Kernelized Correlation Filter (KCF) with the same complexity as linear filters, enabling efficient multi-channel tracking and outperforming top trackers.
Findings
KCF runs at hundreds of frames per second.
KCF and DCF outperform top trackers on benchmark videos.
The approach is simple to implement and open-source.
Abstract
The core component of most modern trackers is a discriminative classifier, tasked with distinguishing 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. Based on this simple observation, we propose 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 computation by several orders of magnitude. Interestingly, for linear regression our formulation is equivalent to a correlation filter, used by some of the fastest competitive trackers. For kernel regression, however, we derive a new Kernelized Correlation Filter…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
