Skip to Main content Skip to Navigation
Conference papers

Détection des défauts dans les vieux films par apprentissage profond à partir d'une restauration semi-manuelle

Abstract : Detection of defects is the first step in the restoration of old movies. The specificity of our work is to learn the expertise of a film restorer through a pair of sequences consisting of a film with defects, and this same film after semi-manual restoration with the help of a specialized software. In order to detect defects with a minimum of human interaction, and to reduce the time spent on restoration, we feed a U-Netneural network with a sequence of defective images as in-put, in order to detect abnormally high spatio-temporal variations in pixel intensity. The output of the network being the defect mask, we create comparative masks using the differences between the defective and restored versions of the film (those used during restoration not being directly accessible). Our network manages to automatically detect real defects more accurately than manual selections, or even some defects missed by the restoration expert.
Document type :
Conference papers
Complete list of metadata
Contributor : Ccsd Sciencesconf.Org Connect in order to contact the contributor
Submitted on : Thursday, September 9, 2021 - 3:40:00 PM
Last modification on : Thursday, January 20, 2022 - 5:31:34 PM
Long-term archiving on: : Saturday, December 11, 2021 - 7:25:23 AM


Files produced by the author(s)


  • HAL Id : hal-03339640, version 1


Arthur Renaudeau, Travis Seng, Axel Carlier, Fabien Pierre, François Lauze, et al.. Détection des défauts dans les vieux films par apprentissage profond à partir d'une restauration semi-manuelle. 18èmes journées francophones des jeunes chercheurs en vision par ordinateur (ORASIS 2021), Centre National de la Recherche Scientifique [CNRS]; Equipe REVA, IRIT : Institut de Recherche en Informatique de Toulouse, Sep 2021, Saint Ferréol, France. ⟨hal-03339640⟩



Les métriques sont temporairement indisponibles