Noisy Supervision for Correcting Misaligned Cadaster Maps Without Perfect Ground Truth Data

Nicolas Girard 1, 2 Guillaume Charpiat 3 Yuliya Tarabalka 1, 2
2 TITANE - Geometric Modeling of 3D Environments
CRISAM - Inria Sophia Antipolis - Méditerranée
3 TAU - TAckling the Underspecified
LRI - Laboratoire de Recherche en Informatique, UP11 - Université Paris-Sud - Paris 11, Inria Saclay - Ile de France, CNRS - Centre National de la Recherche Scientifique : UMR8623
Abstract : In machine learning the best performance on a certain task is achieved by fully supervised methods when perfect ground truth labels are available. However, labels are often noisy, especially in remote sensing where manually curated public datasets are rare. We study the multi-modal cadaster map alignment problem for which available annotations are mis-aligned polygons, resulting in noisy supervision. We subsequently set up a multiple-rounds training scheme which corrects the ground truth annotations at each round to better train the model at the next round. We show that it is possible to reduce the noise of the dataset by iteratively training a better alignment model to correct the annotation alignment.
Liste complète des métadonnées

https://hal.inria.fr/hal-02065211
Contributor : Nicolas Girard <>
Submitted on : Tuesday, March 12, 2019 - 3:22:49 PM
Last modification on : Friday, March 22, 2019 - 6:44:52 PM

Files

girard.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02065211, version 1
  • ARXIV : 1903.06529

Citation

Nicolas Girard, Guillaume Charpiat, Yuliya Tarabalka. Noisy Supervision for Correcting Misaligned Cadaster Maps Without Perfect Ground Truth Data. 2019. ⟨hal-02065211⟩

Share

Metrics

Record views

30

Files downloads

102