Optimization algorithm for restoring an all-focused micromechanical structure image

This paper presents an optimization algorithm of a technique that generates an all-focused image from several differently focused images. The targeted characteristic of differently focused images is that each has only some in-focus regions and the rest of the image is out-of-focus. Using a sharpness function to defect the focused areas causes errors due to the fact that the out-of-focus objects are bigger than their actual size. Therefore, detection of feature edges is spatially incorrect. In addition, focused areas with low contrast cannot be easily distinguished from the out-of-focus regions. Recently we proposed a method to overcome these difficulties. The input images are first enhanced and then divided into different block sizes. The focused blocks are defined using a gradient function followed by a threshold and then a voting algorithm determines the best-focused pixels. Here, an optimization algorithm has been developed which speeds up the image restoration process by 57%.