The goal of flat-field correction is to remove artifacts from 2-D images that are caused by variations in the pixel-to-pixel sensitivity of the detector and/or by distortions in the optical path. The filter requires two images for the input - source image, which represents acquisition of some objects (using microscope, for example), and background image, which is taken without any objects presented. The source image is corrected using the formula: src = bgMean * src / bg, where src - source image's pixel value, bg - background image's pixel value, bgMean - mean value of background image.

Note:If background image is not provided, then it will be automatically generated on each filter run from source image. The automatically generated background image is produced running Gaussian Blur on the original image with (sigma value is set to 5, kernel size is set to 21). Before blurring the original image is resized to 1/3 of its original size and then the result of blurring is resized back to the original size.

Note:The class processes only grayscale (8 bpp indexed) and color (24 bpp) images.

Sample usage:

CopyC#

// create filter
FlatFieldCorrection filter = new FlatFieldCorrection( bgImage );
// process image
filter.ApplyInPlace( sourceImage );

Source image:

Background image:

Result image: