Weakly-supervised continual learning for class-incremental segmentation
Computer Vision and Pattern Recognition
2022-06-16 v2
Abstract
Transfer learning is a powerful way to adapt existing deep learning models to new emerging use-cases in remote sensing. Starting from a neural network already trained for semantic segmentation, we propose to modify its label space to swiftly adapt it to new classes under weak supervision. To alleviate the background shift and the catastrophic forgetting problems inherent to this form of continual learning, we compare different regularization terms and leverage a pseudo-label strategy. We experimentally show the relevance of our approach on three public remote sensing datasets. Code is open-source and released in this repository: https://github.com/alteia-ai/ICSS}{https://github.com/alteia-ai/ICSS.
Cite
@article{arxiv.2201.01029,
title = {Weakly-supervised continual learning for class-incremental segmentation},
author = {Gaston Lenczner and Adrien Chan-Hon-Tong and Nicola Luminari and Bertrand Le Saux},
journal= {arXiv preprint arXiv:2201.01029},
year = {2022}
}