Progressive Multi-Stage Learning for Discriminative Tracking

TL;DR

This paper introduces a progressive multi-stage learning framework with a self-paced sample selection strategy to improve the robustness of visual tracking by effectively handling intra-class variations.

Contribution

It proposes a novel joint discriminative learning scheme with a time-weighted, detection-guided self-paced strategy for sample selection in visual tracking.

Findings

Demonstrates improved tracking robustness on benchmark datasets.

Effectively handles intra-class variations during online learning.

Outperforms existing methods in accuracy and stability.

Abstract

Visual tracking is typically solved as a discriminative learning problem that usually requires high-quality samples for online model adaptation. It is a critical and challenging problem to evaluate the training samples collected from previous predictions and employ sample selection by their quality to train the model. To tackle the above problem, we propose a joint discriminative learning scheme with the progressive multi-stage optimization policy of sample selection for robust visual tracking. The proposed scheme presents a novel time-weighted and detection-guided self-paced learning strategy for easy-to-hard sample selection, which is capable of tolerating relatively large intra-class variations while maintaining inter-class separability. Such a self-paced learning strategy is jointly optimized in conjunction with the discriminative tracking process, resulting in robust tracking results. Experiments on the benchmark datasets demonstrate the effectiveness of the proposed learning framework.