Sparse Labeling Augmentation for Dense Models Training
DOI:
https://doi.org/10.26754/jji-i3a.201802674Abstract
This work proposes and validates a simple but effective approach to train dense semantic segmentation models from sparsely labeled data. Data and labeling collection is most costly task of semantic segmentation. Our approach needs only a few pixels per image reducing the human interaction required.
Downloads
Download data is not yet available.
Downloads
Published
2018-05-25
How to Cite
Alonso, Íñigo, & Murillo Arnal, A. C. (2018). Sparse Labeling Augmentation for Dense Models Training. Jornada De Jóvenes Investigadores Del I3A, 6. https://doi.org/10.26754/jji-i3a.201802674
Issue
Section
Artículos (Tecnologías de la Información y las Comunicaciones)