MANIFOLD REGULARIZED CORRELATION OBJECT TRACKING

Abstract

In this paper, we propose a manifold regularized correlation tracking method with augmented samples. To make better use of the unlabeled data and the manifold structure of the sample space, a manifold regularization-based correlation filter is introduced, which aims to assign similar labels to neighbor samples. Meanwhile, the regression model is learned by exploiting the block-circulant structure of matrices resulting from the augmented translated samples over multiple base samples cropped from both target and nontarget regions. Thus, the final classifier in our method is trained with positive, negative, and unlabeled base samples, which is a semisupervised learning framework. A block optimization strategy is further introduced to learn a manifold regularization-based correlation filter for efficient online tracking. Experiments on two public tracking data sets demonstrate the superior performance of our tracker compared with the state-of-the-art tracking approaches.

Existing System

Kernelized correlation filter (KCF), Manifold Regularization-Based Tracking

Proposed System

We present a method to predict the regression values of unlabeled samples based on the manifold assumption of feature space under the correlation filtering framework in a semisupervised manner. 2) In order to collect plenty of training samples, we introduce an augmented sample set generation technique from image patches cropped from both target and nontarget regions, leading to a more discriminative regression model for visual object tracking. 3) A block method for solving the introduced optimization problem is presented, and it results in an efficient learning model.

Architecture

BLOCK DIAGRAM