摘要： 污点噪声是老电影中常见的问题,它们常表现为黑斑或亮斑。提出了一种有效的方法来去除这种噪声并恢复丢失的数据。通过分析时空域亮度的连续性来检测污点噪声,方法是选出每个检测点的最平滑的方向,估计其未受污染的概率,对该点是否为污点进行软判决,用一个由该概率自适应控制的时空滤波器来进行污点移除数据恢复,引入了一种保护机制来应对连续性分析失败的情况。实验表明这种方法是有效、实用且易于实现的。

关键词: 污点噪声, 自适应滤波, 电影

Abstract: Blotches are common artifacts in old film sequences that manifest themselves as disturbing dark or bright spots. This paper presents a practical method for removing blotches and recovering their missing data, detects the blotch by analysing the spatiotemporal anisotropic brightness continuity. The approach selects the smoothest direction for each examined pixel, and puts out the incorruption probability that the examined pixel may not be corrupted by blotches. The incorruption probability helps to soft-decide whether the examined pixel is belong to blotch or not. Then it uses a filter whose response is adaptively controlled according to a rule defined by the incorruption probability. Furthermore, it introduces a protection method improving robustness against errors of the anisotropic-continuity analysis. Experiments show that this method is efficient, practical and easy to implement.

中图分类号: 

• TP391.41