Image Processing to Facilitate Histological Evaluation of Tissue Specimens Stained With Perl's Prussian Blue

Image-processing techniques were developed to assist histological analysis of rat spleen specimens stained for iron with Perl's Prussian blue. The stained areas in the splenic tissue images were segmented on the basis of color attributes. A series of image features was extracted to describe the subjective concept of blueness, an important attribute in the histological evaluation of blue-stained tissues. Feature extraction was based on statistical analysis. A neural network model was developed to predict pathologists' scores from selected image features. The model predicted pathologists' scores with an R 2 value of 0.86.