Image Inpainting with Variational Autoencoders

Autor: Deesamutara, Sanhanat, Maggiore, Federico
Jazyk: angličtina
Rok vydání: 2020
Předmět:
DOI: 10.5281/zenodo.4383315
Popis: This work presents the training methodology of variational autoencoder in order to restore the missing part of a corrupted image and question the requirement of homogeneity in the implementation and difference situation of corrupted data.
Databáze: OpenAIRE