A New Robust Technique for Stabilizing Brightness Fluctuations in Image Sequences

Localised deflicker of moving images

“…and non linear in old footage [5,8,1]. Ideally the mapping t should be estimated at every pixel site but this would require too much computation.…”

“…Ideally the mapping t should be estimated at every pixel site but this would require too much computation. Since the icker artefact is spatially smooth, the solution adopted in the literature is to interpolate the mapping using 2D polynomial [7,4], cosine [2] or spline [5] functions. Using splines for instance yields the following interpolation [5]:…”

“…There is thus a trade-o between the amount of de icker that can be handled and the propagation of errors in the restored sequence. The key solution to icker compensation is actually to consider the problem as a ltering problem [2,5]. Instead of locking the brightness to the previous frame, it should be sought to average the brightness levels between the current frame and its past and future neighbouring frames.…”

“…Instead of locking the brightness to the previous frame, it should be sought to average the brightness levels between the current frame and its past and future neighbouring frames. The ltering idea [5] can be simpli ed as follows:…”

Motion Driven Tonal Stabilization

“…As we will see, the tonal stabilization algorithm presented in this paper draws on a first raw motion estimation between frames to compute a global color correction. While very few works have attempted to correct color fluctuations in videos, several global [24], [25] or local [26], [27] approaches have been proposed to remove high frequency brightness fluctuations, also known as flicker. In practice, the causes and effects of flicker in videos are quite different from those of tonal instabilities.…”