Bayesian Deep Learning with Monte Carlo Dropout for Qualification of Semantic Segmentation

Clément Dechesne (1) , Pierre Lassalle (1) , Sébastien Lefèvre (2)

1 CNES - Centre National d'Études Spatiales [Toulouse]
2 OBELIX - Observation de l’environnement par imagerie complexe

Despite the intense development of deep neural networks for computer vision, and especially semantic segmentation, their application to Earth Observation data remains usually below accuracy requirements brought by real-life scenarios. Even if well-known deep learning methods produce excellent results, they tend to be over-confident and cannot assess how relevant their predictions are. In this work, a Bayesian deep learning method, based on Monte Carlo Dropout, is proposed to tackle semantic segmentation of aerial and satellite images. Bayesian deep learning can provide both a semantic segmentation and uncertainty maps. Based on the popular U-Net architecture, our model achieves semantic segmentation with high accuracy, e.g. F1-score and overall accuracy respectively reaching 90.84% and 93.22% on a public standard dataset. Uncertainty maps, also derived from our model, show a strong interest in qualitative evaluation of the segmentation and in the improvement of the database.

Mots clés: Deep learning; Semantic segmentation; Bayesian network; Optical imagery; Uncertainty estimation

Domaines: Traitement des images [eess.IV]

Clément Dechesne, Pierre Lassalle, Sébastien Lefèvre. Bayesian Deep Learning with Monte Carlo Dropout for Qualification of Semantic Segmentation. IGARSS 2021 - IEEE International Geoscience and Remote Sensing Symposium, Jul 2021, Brussels, Belgium. ⟨hal-03379980⟩