Similarity analysis of histopathology cell structures using fuzzy rough sets

In this article, a method of histopathology image recognition, based on image similarity is presented. The image similarity is interpreted in terms of fuzzy rough sets. Approximations of fuzzy sets are used for investigation how close (in terms of fuzzy rough sets) is a considered histopathology image to a reference image information, which enables HER2 image recognition. The proposed approach was tested over real clinical data of HER-2/neu breast cancer histopathology images.