Biomedical image processing based on regression models

Abstract

The problem of detecting specific objects or anomalies (OA) in tissue samples of living beings from their digital images is considered. Taking into account that pixel values in images (brightness levels) can be interpreted as sums of values related to physical objects and realizations of additive random variables, regression method (RM) for detection of OA was developed. It is based on mathematical approach of statistical analysis of regression models, coupled with hierarchic clustering method. © 2008 Springer-Verlag.