Visual Object Tracking Using PCA Correlation Filters

Abstract

Accurate translation and robust scale estimation are two challenging problems for visual object tracking.Many existing trackers use some feature extraction methods and the exhaustive scale methods to solve above two problems, respectively.This paper continues to discuss above problems in the tracking-by-detection framework.It proposes an efficient tracker that applies Principal-Component-Analysis (PCA) features to learn the PCA correlation filters, which predicts the location of the target more accurately.Furthermore, our proposed tracker keeps the good performance for the scale variation by using an accurate scale estimation method.Experimental results show that our proposed tracker has a better accuracy for predicting the location of the target and a higher percent in the average overlap precision than some other methods on the 30 benchmark sequences with scale variation.