Automated corruption removal is important in film restoration and typically involves a detection/interpolation step. Current algorithms model the corruption as a binary mixture between the original, clean images and an opaque (dirt) field. This typically causes incomplete removal that manifests as corruption haloes in reconstruction. This paper proposes a new approach by modeling the corruption as a continuous mixture between the two components and generating a solution using a Bayesian framework.
We also present an automated technique for ground-truth generation from infrared scans of corruptions. Previous ground-truth generation efforts require manually inpainting the corrupted regions, which is not scalable.
Our novel ground-truth generation approach, however, allows for the first time large-scale quantitative evaluation of several restoration techniques.
Comparisons with current blotch and line removal techniques show that our proposed corruption removal framework produces more complete removal
and generates less restoration artifacts.
@article{elgharib13,
title = {Blotch and scratch removal in archived film using a semi-transparent corruption model and a ground-truth generation technique},
author = {Elgharib, Mohamed A. and Piti{\'e}, Fran{\c{c}}ois and Kokaram, Anil},
journal = {EURASIP Journal on Image and Video Processing},
number={1},
pages={33},
year={2013}
}
