← Search

Pietro Cerri

2 accepted papers

2020

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

ECCV 2020poster

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 pa…

Cited by 19SourcePDFScholar