Automatic Tissue Type Classification in Large-Scale Microscopic Images Using Zernike Moments

In this paper, we propose an approach in identification of histological sections of human tissue in large-scale microscopic images on the basis of sample tissue fragments from the image. The method uses pattern recognition properties of Zernike moments in the form of image descriptors consisting of sequences of Zernike moments. The goal is to acquire a robust and precise method that allows for identification of the original source of the tissue fragments in the microscopic images. The approach relies on machine learning to perform the final identification of the tissue subject from the constructed image descriptors. The method is verified by a series of experiments on a set of microscopic slides of histological sections. The results and their analysis are presented in the conclusion of the paper.