Analysis of the network pattern in dermatoscopic images

Background/aims: Procedures have been developed for automated analysis of the pigment network in dermatoscopic images. These procedures perform complete network segmentation as well as measurement of the width of network lines and the size and shape of network holes. Methods: The procedures were applied to images of 16 common nevi, 22 dysplastic nevi, and 31 melanomas. Results/Conclusions: There were highly significant differences among the three groups on the measured parameters, and a linear discriminant function composed of mean line width and mean hole area distinguished nevi and melanomas with 80% accuracy.