LSOTB-TIR:A Large-Scale High-Diversity Thermal Infrared Object Tracking Benchmark

TL;DR

This paper introduces LSOTB-TIR, the largest and most diverse thermal infrared object tracking benchmark, with extensive datasets and annotations to facilitate development and evaluation of deep learning TIR trackers.

Contribution

It provides the largest high-diversity TIR tracking dataset with comprehensive annotations and challenge scenarios, enabling fair evaluation and training of deep learning models.

Findings

Deep trackers achieve promising performance on LSOTB-TIR.

Training on LSOTB-TIR significantly improves deep TIR tracker performance.

The benchmark includes over 1,400 sequences and 600K frames for comprehensive evaluation.

Abstract

In this paper, we present a Large-Scale and high-diversity general Thermal InfraRed (TIR) Object Tracking Benchmark, called LSOTBTIR, which consists of an evaluation dataset and a training dataset with a total of 1,400 TIR sequences and more than 600K frames. We annotate the bounding box of objects in every frame of all sequences and generate over 730K bounding boxes in total. To the best of our knowledge, LSOTB-TIR is the largest and most diverse TIR object tracking benchmark to date. To evaluate a tracker on different attributes, we define 4 scenario attributes and 12 challenge attributes in the evaluation dataset. By releasing LSOTB-TIR, we encourage the community to develop deep learning based TIR trackers and evaluate them fairly and comprehensively. We evaluate and analyze more than 30 trackers on LSOTB-TIR to provide a series of baselines, and the results show that deep trackers achieve promising performance. Furthermore, we re-train several representative deep trackers on LSOTB-TIR, and their results demonstrate that the proposed training dataset significantly improves the performance of deep TIR trackers. Codes and dataset are available at https://github.com/QiaoLiuHit/LSOTB-TIR.