Comparison of image focus assessment methods for multi-focused image construction

摘要

Image merging is a process of obtaining one image from multiple. The resulting image carries more information about the photographed scene, than each of the originals. Such an image can be more useful when we deal with human or image processing system. Algorithms that performed this task are used in a wide applying in practical: computer vision, robotics, medicine, forensics, etc. The most popular image focus assessment for usage in the task of forming a multi-focused image were considered. The classification of metrics evaluation was done. Experiments on suggested image focus metrics construction were minutely described. The forms of referenced characteristics were proposed. Using correlation analysis to select the best from the point of view of the problem, metrics and estimated time required for their computation were discussed.