COMPUTER IMAGE-ANALYSIS OF BRAIN CT IMAGES FOR DISCRIMINATING HYPODENSE CEREBRAL-LESIONS IN CHILDREN

A computer software system was designed for the automatic discriminati on of focal oedemas from local glioses in brain CT examinations. Image analysis methods were applied to the images of 77 CT examinations of children with focal oedemas (42) or local glioses (35). Textural featu res derived from the co-occurrence matrix of the lesion's image and a neural network classifier (the multilayer perceptron) were employed fo r the design of the system. Best classification accuracy (89.6%) was a chieved by two textural features (contrast-difference entropy), one hi dden layer and three hidden nodes of the classifier. The proposed soft ware system provides new textural information and may be of value to t he radiologist in differentiating focal oedemas from local glioses, es pecially in small lesions, where other radiological criteria are not e vident.