MFST: Multi-Features Siamese Tracker

TL;DR

MFST introduces a novel multi-feature fusion approach in Siamese tracking, leveraging hierarchical features and multi-CNN models to improve robustness and accuracy over existing methods.

Contribution

The paper proposes a new Siamese tracking algorithm that fuses multiple hierarchical features and uses multi-CNN models for enhanced robustness and performance.

Findings

Outperforms several state-of-the-art trackers.

Achieves higher tracking accuracy.

Utilizes hierarchical feature fusion for robustness.

Abstract

Siamese trackers have recently achieved interesting results due to their balance between accuracy and speed. This success is mainly due to the fact that deep similarity networks were specifically designed to address the image similarity problem. Therefore, they are inherently more appropriate than classical CNNs for the tracking task. However, Siamese trackers rely on the last convolutional layers for similarity analysis and target search, which restricts their performance. In this paper, we argue that using a single convolutional layer as feature representation is not the optimal choice within the deep similarity framework, as multiple convolutional layers provide several abstraction levels in characterizing an object. Starting from this motivation, we present the Multi-Features Siamese Tracker (MFST), a novel tracking algorithm exploiting several hierarchical feature maps for robust deep similarity tracking. MFST proceeds by fusing hierarchical features to ensure a richer and more efficient representation. Moreover, we handle appearance variation by calibrating deep features extracted from two different CNN models. Based on this advanced feature representation, our algorithm achieves high tracking accuracy, while outperforming several state-of-the-art trackers, including standard Siamese trackers. The code and trained models are available at https://github.com/zhenxili96/MFST.