English

Model-based occlusion disentanglement for image-to-image translation

Computer Vision and Pattern Recognition 2020-07-21 v2 Machine Learning Image and Video Processing

Abstract

Image-to-image translation is affected by entanglement phenomena, which may occur in case of target data encompassing occlusions such as raindrops, dirt, etc. Our unsupervised model-based learning disentangles scene and occlusions, while benefiting from an adversarial pipeline to regress physical parameters of the occlusion model. The experiments demonstrate our method is able to handle varying types of occlusions and generate highly realistic translations, qualitatively and quantitatively outperforming the state-of-the-art on multiple datasets.

Keywords

Cite

@article{arxiv.2004.01071,
  title  = {Model-based occlusion disentanglement for image-to-image translation},
  author = {Fabio Pizzati and Pietro Cerri and Raoul de Charette},
  journal= {arXiv preprint arXiv:2004.01071},
  year   = {2020}
}

Comments

ECCV 2020

R2 v1 2026-06-23T14:36:56.962Z