Abstract: The algorithm of maximum entropy method (MEM) proposed by Hollis et al. is introduced. This algorithm can be accelerated by the method suggested by David et al. in 1997. Some restoration examples of real blur images are shown,including motion-blurred images and defocus-blurred images that are common in forensic practice. The performance of this method is effective and feasible in forensic area.

Key words: maximum entropy, image restoration, image processing

摘要： 首先介绍了由Hollis等人提出的最大熵图像恢复算法,该方法经过DavidS.C.Biggs等提出的方法加速后可以大大缩短处理时间。列举运用该加速算法恢复一些实际模糊图像的例子,其中包括运动模糊图像和离焦模糊图像,这些类型的模糊图像都是司法实践中常见的。良好的恢复结果证明了这一方法的有效性及其在司法领域中应用的可行性。

关键词: 最大熵, 图像恢复, 图像处理