On Missing Data Treatment for Degraded Video and Film Archives: A Survey and a New Bayesian Approach

Abstract

Abstract-Image sequence restoration has been steadily gaining in importance with the increasing prevalence of visual digital media. The demand for content increases the pressure on archives to automate their restoration activities for preservation of the cultural heritage that they hold. There are many defects that affect archived visual material and one central issue is that of Dirt and Sparkle, or "Blotches." Research in archive restoration has been conducted for more than a decade and this paper places that material in context to highlight the advances made during that time. The paper also presents a new and simpler Bayesian framework that achieves joint processing of noise, missing data, and occlusion.

Ten Years of Digital Visual Restoration Systems

“…Large DFDs in both directions generally indicate defects. The idea can be articulated under a Bayesian framework [1]. Joyeux et al [3] introduce morphological operators to good effect for small defects.…”

“…In the late 1980's, a BBC Research and Development team led by Richard Storey [1] was the first to consider that degraded broadcast television material could be digitally restored. They concentrated on noise and Dirt removal without motion compensation.…”

“…A taxonomy of defects that have been treated in archive restoration research has been presented in [1]. This paper highlights some salient points that have prevented fully automatic restoration from becoming a reality and hence presents the challenges that remain to the community.…”

A Bayesian Probabilistic Framework for Rain Detection

“…However, both spatial and temporal methods rely on image/video spatial and temporal redundancy. Inspired by [27,28], we build a Bayesian framework to formulate rain or snow detection, which involves long-term temporal constraints and prior distribution of rain or snow. In order to characterize rain detection, we try to harmonize the spatial and temporal considerations into our new Bayesian framework to make full use of the image/video redundant information.…”

“…In this paper, the motion character of rain or snow is assumed to be Pathological Motion (PM), which is introduced in [27,28]. Before presenting the details of our method, we would like to summarize the novel contribution of our paper, which include:…”