Crowdsourcing More Effective Initializations for Single-Target Trackers through Automatic Re-querying

This repository contains the code and data used for the paper "Crowdsourcing More Effective Initializations for Single-target Trackers through Automatic Re-querying," published at the 2021 Conference on Human Factors in Computing Systems (CHI). If you find this work helpful, please cite:

@inproceedings{lemmer_crowdsourcing_2021,
address = {Virtual (Originally Yokohama, Japan)},
title = {Crowdsourcing {More} {Effective} {Initializations} for {Single}-{Target} {Trackers} through {Automatic} {Re}-querying},
booktitle = {Proceedings of the 2021 {Conference} on {Human} {Factors} in {Computing} {Systems}},
publisher = {ACM Press},
author = {Lemmer, Stephan J. and Song, Jean Y. and Corso, Jason J.},
month = may,
year = {2021}
}

This work is separated into two folders: /mturk_tools/ contains files related to the crowdsourcing of initialization bounding boxes, and is derived from the annotation tool here. /tracker/ contains the files related to the analysis of tracker performance using the crowdsourced initializations. Its code is derived from DaSiamRPN. License information for each folder is available in that folder, as the licenses are derived from those of the original work (the annotation tool is BSD 3-Clause, while DaSiamRPN is MIT). Original code in this repository is under the MIT license.