Tissue Segmentation Based on Two-Way ANOVA and Gram-Schmidt Orthogonal Polynomial Approximation

Abstract

This paper presents a method for segmenting tissues in images, regardless of their type or the nature of the
material—whether from magnetic resonance imaging (MRI) or other sources. The approach relies on two
numerical techniques: orthogonal polynomial approximation and the two-way ANOVA (anova2) statistical
method. The study involves two types of images: a normal (healthy) image without impurities or tumors, and a
deformed image, either artificially modified or representing a brain with a tumor. By applying both methods,
tissue structures can be distinguished and enhanced, providing researchers and medical professionals with clearer
image interpretations in a shorter time. The results demonstrate that while the anova2 method offers higher
accuracy for material images, the orthogonal polynomial approximation proves more effective and faster for
medical images.