Dam-based rolling ball with fuzzy-rough constraints, a new background subtraction algorithm for image analysis in microscopy

Light microscopy measurements provide information on the concentration of molecules in cells, tissues, as well as whole mount. Microscope images as we acquire them are not ideal; images specifically often suffer from an uneven background due to illumination flaws. Consequently, for proper images analysis, background correction methods are required. Currently, the "rolling ball" method is often used for illumination correction. This method, however, is far from correct. It is less effective and robust in terms of noise elimination and it generates artificial artefacts in the image. In this paper we investigate correction methods and we propose a new method that employs a combination of fuzzy and rough set theory. The method constructs and constraints an improved version for the "rolling ball". Our results show that we obtained a better solution for background correction for light microscopy images. In addition, the method works more efficient compared to the existing approach.