Diagnosis System for Hepatocellular Carcinoma Based on Fractal Dimension of Morphometric Elements Integrated in an Artificial Neural Network

Figure 4

Overview of the study protocol. Tissue samples from the liver resection pieces of 49 patients with either HCC (21) or LM (28) undergo hematoxylin staining and CD31/34 immunohistochemistry. RGB images are converted in the HSV space and elements are semiautomatically segmented with the calculation of FDs for each element, either cell nuclei or vascular axels. Elements below a 10-pixel threshold are automatically excluded, and the remaining data is fed to a first ANN which classifies the image as either malignant or benign. All malignant images are further classified by a 2nd ANN into either HCC or LM. (The RGB and HSV images provided as examples are reproduced from http://commons.wikimedia.org/wiki/User:SharkD and were originally licensed under the Creative Commons Attribution-Share Alike 3.0 Unported license.)