DFAT

The official implementation of the Information Fusion paper "Exploring Fusion Strategies for Accurate RGBT Visual Object Tracking".

It is supported by the PRCI-Lab (https://github.com/PRCI-Lab)

Introduction

Based on the pysot, DFAT mainly focus on the bias caused by the different imaging mechanism.

Illustration of databias

Download models

The used model can be downloaded from baidu_disk. CODE:nfbe

Download testing datasets

Testing dataset like GTOT, RGBT210, RGBT234 can be downloaded from Chenglong Li's website

VOT-RGBT2019 and VOT-RGBT2020 datasets can be downloaded from baidu_disk. CODE:bcyo

Eval tracker

For VOT toolkit: python_siam_base.py

For testing: python test_gtot.py/test_vot.py

RFN

The RFN block is introduced from RFN-Nest

Contributors

• Zhangyong Tang