Neural network classification of blood vessels and tubules based on haralick features evaluated in histological images of kidney biopsy

Abstract

In this paper, we present a Computer Aided Diagnosis that implements a supervised approach to discriminate vessels versus tubules that are two different types of structural elements in images of biopsy tissue. In particular, in this work we formerly describe an innovative preliminary step to segment region of interest, then the procedure to extract from them significant features and finally present and discuss the Back Propagation Neural Network binary classifier performance that shows Precision 91 % and Recall 91 %.