%0 Book Section
%9 OS CH : Chapitres d'ouvrages scientifiques
%A Jarry, R.
%A Chaumont, M.
%A Berti-Equille, Laure
%A Subsol, G
%T Assessment of CNN-based methods for poverty estimation from satellite images
%B Pattern Recognition : proceedings, part VII
%C Cham
%D 2021
%E Del Bimbo, A.
%E Cucchiara, R.
%E Sclaroff, S.
%E Farinella, G.M.
%E Mei, T.
%E Bertini, M.
%E Escalante, H.J.
%E Vezzani, R.
%L fdi:010085560
%G ENG
%I Springer
%@ 978-3-030-68786-1
%N 12667
%P 550-565
%R 10.1007/978-3-030-68787-8_40
%U https://www.documentation.ird.fr/hor/fdi:010085560
%> https://www.documentation.ird.fr/intranet/publi/2023-01/010085560.pdf
%W Horizon (IRD)
%X One of the major issues in predicting poverty with satellite images is the lack of fine-grained and reliable poverty indicators. To address this problem, various methodologies were proposed recently. Most recent approaches use a proxy (e.g., nighttime light), as an additional information, to mitigate the problem of sparse data. They consist in building and training a CNN with a large set of images, which is then used as a feature extractor. Ultimately, pairs of extracted feature vectors and poverty labels are used to learn a regression model to predict the poverty indicators.First, we propose a rigorous comparative study of such approaches based on a unified framework and a common set of images. We observed that the geographic displacement on the spatial coordinates of poverty observations degrades the prediction performances of all the methods. Therefore, we present a new methodology combining grid-cell selection and ensembling that improves the poverty prediction to handle coordinate displacement.
%S Lecture Notes in Computer Science
%B ICPR.International Workshops and Challenges
%8 2021/01/10-15
%$ 126 ; 122