Since the beginning of this Century we are assisting to a crucial step in the domain of cinema: the transition from analogic to digital. The computed generated imagery substituted the early developed special effects on film and the use of smartphones and digital devices took the place of old Super8 formats. In this context, film restoration passes quite often by the digitization and the use of software for image enhancement and restoration are widely diffused and used also by non-pro users. The digital restoration of colour has widely spread, and digital techniques allow to easier removal of colour casts, frames histograms equalization, dynamic range expansion and saturation adjustment. Despite this easy applicability, when adjusting the colour of a film of historical value, constant supervision by qualified personnel is still required. In our contribution we show and resume our results from the works and projects we have conducted in digital film restoration and we describe our method based on the idea of recovering the original appearance of colours instead of the original colour signal. This approach is based on the application of Spatial Colour Algorithms (SCAs), a family of algorithms inspired by the capabilities of Human Visual System (HSV) of autonomously adjust the variation of lightness and colours in a scene. SCAs are based on the idea that colour sensation does not depend on the pointwise value of the stimulus, but on the spatial arrangement of the stimuli in the scene. In fact, identical levels of radiance at a point can originate completely different colour sensations according to the values in the rest of the image. To reproduce this behaviour SCAs performs spatial comparisons among the pixels in the input image and for this reason SCAs are characterized by a high computational cost and a local and global behaviour. All the SCAs present a common structure: a first step in which each pixel is recomputed according to the spatial distribution of the other pixels in the image, in the second step an adjustment can be added according to the SCA goal. Due to this characteristic, SCAs can be suitable for analog-to-digital dynamic range adjustment or colour enhancement when the fading occurred on a film. Thanks to the SCAs algorithms is possible to conduct a semi-automatic and unsupervised colour restoration on faded frames, with a technique that permit also to non-pro users to operate on film restoration without the need of expensive and complex software. This solution could be an interesting alternative to classical restoration approach in the domain of little archives or institution that aim to valorise and diffuse their collections and records without the need of a complex and long process of restoration. In addition those new methods could open an interesting discussion in how evaluate the quality of a restored film, and how to proceed in a restoration when there are no references.

Algorithms for film digital restoration: unsupervised approaches for film frames enhancement / A. Rizzi, A. Plutino. ((Intervento presentato al 26. convegno Moving Pictures, Living Machines: Automation, Animation and the Imitation of Life in Cinema and Media tenutosi a Gorizia nel 2019.

Algorithms for film digital restoration: unsupervised approaches for film frames enhancement

A. Rizzi;A. Plutino
2019

Abstract

Since the beginning of this Century we are assisting to a crucial step in the domain of cinema: the transition from analogic to digital. The computed generated imagery substituted the early developed special effects on film and the use of smartphones and digital devices took the place of old Super8 formats. In this context, film restoration passes quite often by the digitization and the use of software for image enhancement and restoration are widely diffused and used also by non-pro users. The digital restoration of colour has widely spread, and digital techniques allow to easier removal of colour casts, frames histograms equalization, dynamic range expansion and saturation adjustment. Despite this easy applicability, when adjusting the colour of a film of historical value, constant supervision by qualified personnel is still required. In our contribution we show and resume our results from the works and projects we have conducted in digital film restoration and we describe our method based on the idea of recovering the original appearance of colours instead of the original colour signal. This approach is based on the application of Spatial Colour Algorithms (SCAs), a family of algorithms inspired by the capabilities of Human Visual System (HSV) of autonomously adjust the variation of lightness and colours in a scene. SCAs are based on the idea that colour sensation does not depend on the pointwise value of the stimulus, but on the spatial arrangement of the stimuli in the scene. In fact, identical levels of radiance at a point can originate completely different colour sensations according to the values in the rest of the image. To reproduce this behaviour SCAs performs spatial comparisons among the pixels in the input image and for this reason SCAs are characterized by a high computational cost and a local and global behaviour. All the SCAs present a common structure: a first step in which each pixel is recomputed according to the spatial distribution of the other pixels in the image, in the second step an adjustment can be added according to the SCA goal. Due to this characteristic, SCAs can be suitable for analog-to-digital dynamic range adjustment or colour enhancement when the fading occurred on a film. Thanks to the SCAs algorithms is possible to conduct a semi-automatic and unsupervised colour restoration on faded frames, with a technique that permit also to non-pro users to operate on film restoration without the need of expensive and complex software. This solution could be an interesting alternative to classical restoration approach in the domain of little archives or institution that aim to valorise and diffuse their collections and records without the need of a complex and long process of restoration. In addition those new methods could open an interesting discussion in how evaluate the quality of a restored film, and how to proceed in a restoration when there are no references.
22-mar-2019
Settore INF/01 - Informatica
https://filmforumfestival.it/conference
Algorithms for film digital restoration: unsupervised approaches for film frames enhancement / A. Rizzi, A. Plutino. ((Intervento presentato al 26. convegno Moving Pictures, Living Machines: Automation, Animation and the Imitation of Life in Cinema and Media tenutosi a Gorizia nel 2019.
Conference Object
File in questo prodotto:
File Dimensione Formato  
PlutinoRizzi_Proposal.pdf

accesso aperto

Tipologia: Post-print, accepted manuscript ecc. (versione accettata dall'editore)
Dimensione 598.85 kB
Formato Adobe PDF
598.85 kB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/2434/645021
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact