IEEE Geoscience and Remote Sensing Magazine - December 2023 - 118

research efforts toward improving the quality-efficiency performance
of real-time semantic segmentation deep learning
methods in remote sensing image analysis.
◗ How can we design a specialized efficient neural network backbone
to improve the balance between accuracy and speed for accurate
and fast segmentation of remotely sensed images in real time?
Exploring and adapting existing lightweight networks like
MobileNet or EfficientNet via an automated architecture
search approach for more suitable real-time segmentation
methods in remote sensing may be an interesting challenge.
◗ How can we improve the domain generalization performance as
well as the UDA of real-time semantic segmentation in remote
sensing? The segmentation quality of the continent-wise
domain generalization is far behind that of the supervised
semantic segmentation. Future work may adopt strategies
in related tasks that have been proven successful, such as
self-training and curriculum learning to improve the accuracy
of real-time continent-wise domain generalization and
continent-wise UDA.
◗ How can we solve the variations in resolution and the problem of
different sensor types of remote sensing imagery? Remote sensing
imagery often comes from different sensors (optical or
synthetic aperture radar) at different resolutions (meter or
submeter), which can make it difficult for the algorithm to
differentiate between different land cover classes. Exploring
novel approaches to process images from different sensors
at various resolutions in real-time segmentation represents
an intriguing research direction in remote sensing
image understanding.
◗ How can we improve the robustness of real-time semantic segmentation
algorithms to different imaging conditions such as weather
conditions, seasonality, and temporal changes? Remote sensing
images can be affected by various environmental factors,
and developing efficient semantic segmentation networks
that can adapt to these factors with a better balance between
accuracy and speed in real time is an important challenge.
◗ How can we improve the scalability of real-time semantic segmentation
algorithms for large-scale remote sensing applications?
Remote sensing images can cover vast areas, and processing
such large datasets (e.g., county wide or province wide)
in real time could be a significant challenge. Developing
efficient scalable algorithms that can segment large-scale
remote sensing data in real time is essential for provincewide
or county-wide remote sensing applications, such as
forest fire monitoring.
AUTHOR INFORMATION
Clifford Broni-Bediako (clifford.broni-bediako@riken.jp)
received his B.Sc. degree in computer science from the
Kwame Nkrumah University of Science and Technology, Kumasi,
Ghana, in 2003, his M.Phil. degree in computer science
from the University of Ghana, Accra, Ghana, in 2018,
and his Ph.D. degree in engineering from Soka University,
Tokyo, Japan, in 2022. Since 2022, he has been a postdoctoral
researcher with the RIKEN Center for Advanced Intelligence
Project, the Geoinformatics Team, Tokyo 1030027,
118
Japan. His current research revolves around deep learning,
particularly efficient deep neural networks, automated neural
architecture search, and remote sensing image understanding
with deep learning algorithms. He is a Member of IEEE.
Junshi Xia (junshi.xia@riken.jp) received his B.S. degree
in geographic information systems and his Ph.D. degree in
photogrammetry and remote sensing from the China University
of Mining and Technology, Xuzhou, China, in 2008
and 2013, respectively, and his Ph.D. degree in image processing
from the Grenoble Images Speech Signals and Automatics
Laboratory, Grenoble Institute of Technology, Grenoble,
France, in 2014. From 2014 to 2015, he was a visiting scientist
with the Department of Geographic Information Sciences,
Nanjing University, Nanjing, China. From 2015 to 2016,
he was a postdoctoral research fellow with the University
of Bordeaux, Bordeaux, France. From 2016 to 2018, he was
the Japan Society for the Promotion of Science Postdoctoral
Overseas Research Fellow with the University of Tokyo,
Tokyo, Japan. Since 2018, he has been a research scientist
with the RIKEN Center for Advanced Intelligence Project, the
Geoinformatics Team, Tokyo 1030027, Japan. His research
interests include multiple classifier systems in remote sensing,
hyperspectral remote sensing image processing, and
deep learning in remote sensing applications. He was the
recipient of the first place prize in the IEEE Geoscience and
Remote Sensing Society Data Fusion Contest organized by
the Image Analysis and Data Fusion Technical Committee
in 2017. Since 2019, he has been an associate editor of IEEE
Geoscience and Remote Sensing Letters, Remote Sensing, and
Frontiers in Remote Sensing and guest editor of Remote Sensing
and IEEE Journal of Selected Topics in Applied Earth Observations
and Remote Sensing. He is a Senior Member of IEEE.
Naoto Yokoya (yokoya@k.u-tokyo.ac.jp) received his
M.Eng. and Ph.D. degrees from the Department of Aeronautics
and Astronautics, The University of Tokyo, Tokyo, Japan,
in 2010 and 2013, respectively. He was an assistant professor
at the University of Tokyo from 2013 to 2017 and a lecturer
from 2020 to 2022. From 2015 to 2017, he was an Alexander
von Humboldt Fellow with the German Aerospace Center,
Oberpfaffenhofen, Germany, and the Technical University
of Munich, Munich, Germany. He is currently an associate
professor at the University of Tokyo and a team leader at
the RIKEN Center for Advanced Intelligence Project, Tokyo
1030027, Japan, where he leads the Geoinformatics Team.
His research is focused on the development of image processing,
data fusion, and machine learning algorithms for understanding
remote sensing images with applications to disaster
management and environmental monitoring. He won first
place in the 2017 IEEE Geoscience and Remote Sensing Society
(GRSS) Data Fusion Contest organized by the Image
Analysis and Data Fusion Technical Committee (IADF TC).
He was chair from 2019 to 2021, a cochair of the IEEE GRSS
IADF TC from 2017 to 2019, and also secretary of the IEEE
GRSS All Japan Joint Chapter from 2018 to 2021. He was an
associate editor of IEEE Journal of Selected Topics in Applied
Earth Observations and Remote Sensing from 2018 to 2021. He
IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE DECEMBER 2023
IEEE Geoscience and Remote Sensing Magazine - December 2023

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - December 2023
