FACE : Fast mean Absolute difference with Confidence propagation for Extended depth of field

Bright field microscopes with low depth of field provide good magnification and image quality. However, samples generally have a thickness greater than the depth of field, thus being not entirely observable within the viewing distances intervals and preventing the creation of sharp images. It is possible to gradually vary the observation distance to create an image stack where all areas of the image can appear sharp, but all these areas are not on the same image. To allow observation of a complete sample, it is then necessary to process this stack of images to create a new and unique image where the sample appears clear at all pixels. This process is commonly called extended depth-of-field method.

The FACE (Fast mean Absolute difference with Confidence propagation for Extended depth of field) algorithm allows sharp image to be created for all pixels. It consists of a local sharpness estimator which offers a compromise between speed and robustness to noise by the use of calculation of absolute differences and local means, combined with two fast post-processing steps. The reduction of artefacts is obtained by propagating the sharpness information from pixels where the sharpness estimator is considered as reliable to pixels where it is not reliable. FACE algorithm is particularly well adapted for reconstruction of thin objects with sharp objects-background transitions. If you use this plugin, please cite the paper below where more details about the method are also available.

To learn how this tool is used at IFPEN to help understand the production of advanced biofuels using enzymatic processes, follow the second link below.