Robust Unsupervised Multi-Object Tracking in Noisy Environments
Physical processes, camera movement, and unpredictable environmental conditions like the presence of dust can induce noise and artifacts in video feeds. We observe that popular unsupervised MOT methods are dependent on noise-free inputs. We show that the addition of a small amount of artificial rand...
Main Authors: | , , , , , |
---|---|
Format: | Article in Journal/Newspaper |
Language: | unknown |
Published: |
arXiv
2021
|
Subjects: | |
Online Access: | https://dx.doi.org/10.48550/arxiv.2105.10005 https://arxiv.org/abs/2105.10005 |
id |
ftdatacite:10.48550/arxiv.2105.10005 |
---|---|
record_format |
openpolar |
spelling |
ftdatacite:10.48550/arxiv.2105.10005 2023-05-15T15:33:29+02:00 Robust Unsupervised Multi-Object Tracking in Noisy Environments Yang, C. -H. Huck Chhabra, Mohit Liu, Y. -C. Kong, Quan Yoshinaga, Tomoaki Murakami, Tomokazu 2021 https://dx.doi.org/10.48550/arxiv.2105.10005 https://arxiv.org/abs/2105.10005 unknown arXiv https://dx.doi.org/10.1109/icip42928.2021.9506029 Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode cc-by-4.0 CC-BY Computer Vision and Pattern Recognition cs.CV Artificial Intelligence cs.AI Machine Learning cs.LG Multimedia cs.MM Neural and Evolutionary Computing cs.NE FOS Computer and information sciences article-journal Article ScholarlyArticle Text 2021 ftdatacite https://doi.org/10.48550/arxiv.2105.10005 https://doi.org/10.1109/icip42928.2021.9506029 2022-03-10T14:30:48Z Physical processes, camera movement, and unpredictable environmental conditions like the presence of dust can induce noise and artifacts in video feeds. We observe that popular unsupervised MOT methods are dependent on noise-free inputs. We show that the addition of a small amount of artificial random noise causes a sharp degradation in model performance on benchmark metrics. We resolve this problem by introducing a robust unsupervised multi-object tracking (MOT) model: AttU-Net. The proposed single-head attention model helps limit the negative impact of noise by learning visual representations at different segment scales. AttU-Net shows better unsupervised MOT tracking performance over variational inference-based state-of-the-art baselines. We evaluate our method in the MNIST-MOT and the Atari game video benchmark. We also provide two extended video datasets: ``Kuzushiji-MNIST MOT'' which consists of moving Japanese characters and ``Fashion-MNIST MOT'' to validate the effectiveness of the MOT models. : Accepted to IEEE ICIP 2021 Article in Journal/Newspaper Attu DataCite Metadata Store (German National Library of Science and Technology) |
institution |
Open Polar |
collection |
DataCite Metadata Store (German National Library of Science and Technology) |
op_collection_id |
ftdatacite |
language |
unknown |
topic |
Computer Vision and Pattern Recognition cs.CV Artificial Intelligence cs.AI Machine Learning cs.LG Multimedia cs.MM Neural and Evolutionary Computing cs.NE FOS Computer and information sciences |
spellingShingle |
Computer Vision and Pattern Recognition cs.CV Artificial Intelligence cs.AI Machine Learning cs.LG Multimedia cs.MM Neural and Evolutionary Computing cs.NE FOS Computer and information sciences Yang, C. -H. Huck Chhabra, Mohit Liu, Y. -C. Kong, Quan Yoshinaga, Tomoaki Murakami, Tomokazu Robust Unsupervised Multi-Object Tracking in Noisy Environments |
topic_facet |
Computer Vision and Pattern Recognition cs.CV Artificial Intelligence cs.AI Machine Learning cs.LG Multimedia cs.MM Neural and Evolutionary Computing cs.NE FOS Computer and information sciences |
description |
Physical processes, camera movement, and unpredictable environmental conditions like the presence of dust can induce noise and artifacts in video feeds. We observe that popular unsupervised MOT methods are dependent on noise-free inputs. We show that the addition of a small amount of artificial random noise causes a sharp degradation in model performance on benchmark metrics. We resolve this problem by introducing a robust unsupervised multi-object tracking (MOT) model: AttU-Net. The proposed single-head attention model helps limit the negative impact of noise by learning visual representations at different segment scales. AttU-Net shows better unsupervised MOT tracking performance over variational inference-based state-of-the-art baselines. We evaluate our method in the MNIST-MOT and the Atari game video benchmark. We also provide two extended video datasets: ``Kuzushiji-MNIST MOT'' which consists of moving Japanese characters and ``Fashion-MNIST MOT'' to validate the effectiveness of the MOT models. : Accepted to IEEE ICIP 2021 |
format |
Article in Journal/Newspaper |
author |
Yang, C. -H. Huck Chhabra, Mohit Liu, Y. -C. Kong, Quan Yoshinaga, Tomoaki Murakami, Tomokazu |
author_facet |
Yang, C. -H. Huck Chhabra, Mohit Liu, Y. -C. Kong, Quan Yoshinaga, Tomoaki Murakami, Tomokazu |
author_sort |
Yang, C. -H. Huck |
title |
Robust Unsupervised Multi-Object Tracking in Noisy Environments |
title_short |
Robust Unsupervised Multi-Object Tracking in Noisy Environments |
title_full |
Robust Unsupervised Multi-Object Tracking in Noisy Environments |
title_fullStr |
Robust Unsupervised Multi-Object Tracking in Noisy Environments |
title_full_unstemmed |
Robust Unsupervised Multi-Object Tracking in Noisy Environments |
title_sort |
robust unsupervised multi-object tracking in noisy environments |
publisher |
arXiv |
publishDate |
2021 |
url |
https://dx.doi.org/10.48550/arxiv.2105.10005 https://arxiv.org/abs/2105.10005 |
genre |
Attu |
genre_facet |
Attu |
op_relation |
https://dx.doi.org/10.1109/icip42928.2021.9506029 |
op_rights |
Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode cc-by-4.0 |
op_rightsnorm |
CC-BY |
op_doi |
https://doi.org/10.48550/arxiv.2105.10005 https://doi.org/10.1109/icip42928.2021.9506029 |
_version_ |
1766364042918625280 |