IEEE Geoscience and Remote Sensing Magazine - March 2019 - 105

training in today's powerful statistical learning algorithms.
In particular, we introduced a track for 3D estimation from
a single image, a task that would have seemed out of reach
without the recent advances in deep learning.
get the Data, anD enter the contest
The 2019 Data Fusion Contest is cast as a three-phase
benchmark. It started on 7 January with the release of
training data (with reference labels) and evaluation data
(without reference) to allow participants to elaborate their
best strategy and train their algorithms. One month later,
on 7 February, the evaluation server opened to let participants measure their performances on the evaluation data
and rank themselves in a public leaderboard. The definitive
test phase is currently in process: on 7 March, we released
the final test data, and participants have until 22 March to
submit 3D semantic maps along with a short description of
their approach.
To register for the contest and download the data, participants visit the IADF TC website (http://www.grss-ieee
.org/community/technical-committees/data-fusion/datafusion-contest). There, they get information about how
to proceed, links to download the data, and access to
the evaluation server. They are required to read and accept
the contest terms and conditions. Questions and comments about the data and the contest can be submitted to
the LinkedIn group of the IADF TC (http://www.grss-ieee
.org/community/technical-committees/data-fusion/datafusion-contest/).
awarDees, awarDs, anD prIzes
This year, eight teams will receive awards: the winner (bestranked submission) of each track and the second-best team
of each track. They will be invited to submit a manuscript to
the 2019 IEEE International Geoscience and Remote Sensing Symposium (IGARSS) in Yokohama, Japan, formatted
according to the standard guidelines and templates. It will
describe the method used to enter the Data Fusion Contest
and the results. All eight papers will be presented by the
winners in a long invited session dedicated to the outcome
of the contest at IGARSS 2019. They will also be included
in the technical program and in the IGARSS 2019 proceedings. The winning teams will receive their awards in July
2019 at IGARSS 2019. The awards are as follows:
◗ The four winners and four second-ranked teams will
present their papers in the 2019 Data Fusion Contest invited session.
◗ The same four winners and four second-ranked teams
will be awarded an IEEE Certificate of Recognition. The
awards ceremony will take place at the Technical Committee and Chapter Chairs' Dinner at IGARSS 2019.
◗ The four winners (first-ranked result in each track) will
also be awarded a special prize.

MARCH 2019

ieee Geoscience and remote sensing magazine

◗ The three winners of tracks 1, 2, and 3 will coauthor

an article that will summarize the outcomes of the
contest concerning semantic 3D reconstruction from
space, which will be submitted to IEEE Journal of Selected Topics in Applied Earth Observations in Remote
Sensing (JSTARS).
◗ The winner and the second-ranked team of track 4 will
coauthor an article that will summarize the outcomes
of the contest concerning semantic 3D point-cloud classification, which will be submitted to JSTARS.
For both journal articles,
a maximum of three authors
per team will be included. To
thIs year, the Data
maximize impact and profusIon contest centers
mote the potential of current
on a topIc of growIng
multisource remote sensing
technologies, the open-access
Importance: 3D earth
option will be used for these
observatIon.
submissions. The GRSS will
cover the costs related to the
open-access fees and to the
winning teams' participation in the Technical Committee
and Chapter Chairs' Dinner at IGARSS 2019. It will also
sponsor the prizes for the winning teams.
acknowleDgments
The contest is being organized in collaboration with IARPA and the Applied Physics Laboratory (APL) at Johns Hopkins University (JHU). WorldView3 satellite imagery is
provided courtesy of DigitalGlobe, and lidar data are
provided courtesy of IARPA. The IADF TC chairs would
like to thank IARPA and JHU's APL for providing the data,
HakJae Kim (IARPA) for help and support, JHU's APL staff
for preparing the data, and the GRSS for continuously supporting the annual Data Fusion Contest through funding
and resources.
references
[1] N. Yokoya et al., "Open data for global multimodal land use classification: Outcome of the 2017 IEEE GRSS data fusion contest,"
IEEE J. Sel. Topics Appl. Earth Observ. Remote Sens., vol. 11, no. 5,
pp. 1363-1377, 2018. [Online]. Available: https://ieeexplore
.ieee.org/document/8338367
[2] B. Le Saux, N. Yokoya, R. Hänsch, and S. Prasad, "Advanced
multisource optical remote sensing for urban land use and land
cover classification," IEEE Geosci. Remote Sens. Mag., vol. 6, no. 4,
pp. 85-89, 2018. [Online]. Available: https://ieeexplore.ieee
.org/document/8573970
[3] M. Bosch, K. Foster, G. Christie, S. Wang, G. D. Hager, and M.
Brown, "Semantic stereo for incidental satellite images," in Proc.
Winter Conf. Appl. Comput. Vision, 2019.
grs

105

http://www.grss-ieee.org/community/technical-committees/data-fusion/data-fusion-contest http://www.grss-ieee.org/community/technical-committees/data-fusion/data-fusion-contest http://www.grss-ieee.org/community/technical-committees/data-fusion/data-fusion-contest http://www.grss-ieee.org/community/technical-committees/data-fusion/data-fusion-contest/ http://www.grss-ieee.org/community/technical-committees/data-fusion/data-fusion-contest/ http://www.grss-ieee.org/community/technical-committees/data-fusion/data-fusion-contest/ https://ieeexplore.ieee.org/document/8338367 https://ieeexplore.ieee.org/document/8338367 https://www.ieeexplore.ieee.org/document/8573970 https://www.ieeexplore.ieee.org/document/8573970

IEEE Geoscience and Remote Sensing Magazine - March 2019

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - March 2019