Unsupervised digital movie restoration with spatial models of color

Abstract

In this paper, we present an efficient approach to digital color restoration, based on the idea of recovering the appearance of color rather than the original color signal, since, in most of the cases, for old films the original reference is missing, and new films or digital coding can be subject to severe gamut transformations. This approach is based on the application of algorithms inspired by the capabilities of the Human Vision System of automatically adjust to the variation of color and lightness in the scene. The proposed method allows more unsupervised restoration. We present an overview of the approach, characteristics of this family of algorithms, and restoration examples.

Film and Digital Media: Open Issues and Novel Approaches for Digital Color Film Restoration

“…Retinex models and SCAs algorithms have been found successful for many applications in different domains, among which the cinematographic and photographic restoration. In this field Retinex implementations have been used as kick-off technique in the workflow, but also to perform full restoration at reduced costs [21], [22], [23]. See Figure 2 for some examples.…”

A cockpit of multiple measures for assessing film restoration quality

“…N sets of M pixels randomly sampled around p t with radial density, and averaging the reciprocals of the maximum intensities of the sprays. RSR has proven to be effective for color movie restoration [36] . STRESS inherits from RSR the idea of a local white reference, but implements a very different pixel value stretching.…”

“…Digital Restoration, DR, applies two main steps: the "DR" per se , which minimizes or even removes artifacts like flickering, scratches, dust and stabilizes the scenes ( [14] , and the "Color Correction", which handles colour alteration and fading. While mostly used restoring software employ automatic or semi-automatic algorithms for DR, the color correction task is still mainly manually performed by experts via semi-automatic methods, though unsupervised algorithms have been recently proposed [5,36,37,46] . DR presents several advantages with respect to the analog restoration.…”

Point‐based spatial colour sampling in Milano‐Retinex: a survey

“…Furthermore, these characteristics are suitable for other visualisation purposes. For example, these types of algorithms have been successfully used in the field of cultural heritage, to restore digital movies [5]; in a tool to assess indoor and outdoor advertising [6]; to locally correct the colour in stereoscopic shooting [7], and recently has been tested has a daltonisation method [8,9].…”