Object detectors trained on large-scale RGB datasets are being extensively
employed in real-world applications. However, these RGB-trained models suffer a
performance drop under adverse illumination and lighting conditions. Infrared
(IR) cameras are robust under such conditions and can be helpful in real-world
applications. Though thermal cameras are widely used for military applications
and increasingly for commercial applications, there is a lack of robust
algorithms to robustly exploit the thermal imagery due to the limited
availability of labeled thermal data. In this work, we aim to enhance the
object detection performance in the thermal domain by leveraging the labeled
visible domain data in an Unsupervised Domain Adaptation (UDA) setting. We
propose an algorithm agnostic meta-learning framework to improve existing UDA
methods instead of proposing a new UDA strategy. We achieve this by
meta-learning the initial condition of the detector, which facilitates the
adaptation process with fine updates without overfitting or getting stuck at
local optima. However, meta-learning the initial condition for the detection
scenario is computationally heavy due to long and intractable computation
graphs. Therefore, we propose an online meta-learning paradigm which performs
online updates resulting in a short and tractable computation graph. To this
end, we demonstrate the superiority of our method over many baselines in the
UDA setting, producing a state-of-the-art thermal detector for the KAIST and
DSIAC datasets.
Description
Meta-UDA: Unsupervised Domain Adaptive Thermal Object Detection using Meta-Learning
%0 Journal Article
%1 vs2021metauda
%A VS, Vibashan
%A Poster, Domenick
%A You, Suya
%A Hu, Shuowen
%A Patel, Vishal M.
%D 2021
%K test
%T Meta-UDA: Unsupervised Domain Adaptive Thermal Object Detection using
Meta-Learning
%U http://arxiv.org/abs/2110.03143
%X Object detectors trained on large-scale RGB datasets are being extensively
employed in real-world applications. However, these RGB-trained models suffer a
performance drop under adverse illumination and lighting conditions. Infrared
(IR) cameras are robust under such conditions and can be helpful in real-world
applications. Though thermal cameras are widely used for military applications
and increasingly for commercial applications, there is a lack of robust
algorithms to robustly exploit the thermal imagery due to the limited
availability of labeled thermal data. In this work, we aim to enhance the
object detection performance in the thermal domain by leveraging the labeled
visible domain data in an Unsupervised Domain Adaptation (UDA) setting. We
propose an algorithm agnostic meta-learning framework to improve existing UDA
methods instead of proposing a new UDA strategy. We achieve this by
meta-learning the initial condition of the detector, which facilitates the
adaptation process with fine updates without overfitting or getting stuck at
local optima. However, meta-learning the initial condition for the detection
scenario is computationally heavy due to long and intractable computation
graphs. Therefore, we propose an online meta-learning paradigm which performs
online updates resulting in a short and tractable computation graph. To this
end, we demonstrate the superiority of our method over many baselines in the
UDA setting, producing a state-of-the-art thermal detector for the KAIST and
DSIAC datasets.
@article{vs2021metauda,
abstract = {Object detectors trained on large-scale RGB datasets are being extensively
employed in real-world applications. However, these RGB-trained models suffer a
performance drop under adverse illumination and lighting conditions. Infrared
(IR) cameras are robust under such conditions and can be helpful in real-world
applications. Though thermal cameras are widely used for military applications
and increasingly for commercial applications, there is a lack of robust
algorithms to robustly exploit the thermal imagery due to the limited
availability of labeled thermal data. In this work, we aim to enhance the
object detection performance in the thermal domain by leveraging the labeled
visible domain data in an Unsupervised Domain Adaptation (UDA) setting. We
propose an algorithm agnostic meta-learning framework to improve existing UDA
methods instead of proposing a new UDA strategy. We achieve this by
meta-learning the initial condition of the detector, which facilitates the
adaptation process with fine updates without overfitting or getting stuck at
local optima. However, meta-learning the initial condition for the detection
scenario is computationally heavy due to long and intractable computation
graphs. Therefore, we propose an online meta-learning paradigm which performs
online updates resulting in a short and tractable computation graph. To this
end, we demonstrate the superiority of our method over many baselines in the
UDA setting, producing a state-of-the-art thermal detector for the KAIST and
DSIAC datasets.},
added-at = {2021-10-09T05:14:44.000+0200},
author = {VS, Vibashan and Poster, Domenick and You, Suya and Hu, Shuowen and Patel, Vishal M.},
biburl = {https://www.bibsonomy.org/bibtex/2271f2678bb1235d621626229fbfaa4db/markbruns},
description = {Meta-UDA: Unsupervised Domain Adaptive Thermal Object Detection using Meta-Learning},
interhash = {8d9d86254c7758c47bc7cd151ded37cf},
intrahash = {271f2678bb1235d621626229fbfaa4db},
keywords = {test},
note = {cite arxiv:2110.03143Comment: Accepted to WACV 2022},
timestamp = {2021-10-09T05:14:44.000+0200},
title = {Meta-UDA: Unsupervised Domain Adaptive Thermal Object Detection using
Meta-Learning},
url = {http://arxiv.org/abs/2110.03143},
year = 2021
}