Tackling Catastrophic Forgetting and Background Shift in Continual Semantic Segmentation - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2021

Tackling Catastrophic Forgetting and Background Shift in Continual Semantic Segmentation

Résumé

Deep learning approaches are nowadays ubiquitously used to tackle computer vision tasks such as semantic segmentation, requiring large datasets and substantial computational power. Continual learning for semantic segmentation (CSS) is an emerging trend that consists in updating an old model by sequentially adding new classes. However, continual learning methods are usually prone to catastrophic forgetting. This issue is further aggravated in CSS where, at each step, old classes from previous iterations are collapsed into the background. In this paper, we propose Local POD, a multi-scale pooling distillation scheme that preserves long- and short-range spatial relationships at feature level. Furthermore, we design an entropy-based pseudo-labelling of the background w.r.t. classes predicted by the old model to deal with background shift and avoid catastrophic forgetting of the old classes. Finally, we introduce a novel rehearsal method that is particularly suited for segmentation. Our approach, called PLOP, significantly outperforms state-of-the-art methods in existing CSS scenarios, as well as in newly proposed challenging benchmarks.

Dates et versions

hal-03503834 , version 1 (28-12-2021)

Identifiants

Citer

Arthur Douillard, Yifu Chen, Arnaud Dapogny, Matthieu Cord. Tackling Catastrophic Forgetting and Background Shift in Continual Semantic Segmentation. 2021. ⟨hal-03503834⟩
11 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More