| 摘要 |
<p>The method involves classifying different areas of picture according to their relative focusing, and applying local indiscriminating deconvolution with small number of iterations to estimate local out-of-focus states of the areas of picture. The areas of picture are classified in a defined number of relative classes, based on the results of local indiscriminating deconvolution. The areas are ranked, according to the focusing levels, such that segmented picture is output and the individual class of segmentation corresponds to relative classes of focusing.</p> |
