Automated wound identification system based on image segmentation and Artificial Neural Networks

Abstract

A system that can automatically and accurately identify the region of a chronic wound could largely improve conventional clinical practice for the wound diagnosis and treatment. We designed a system that uses color wound photographs taken from the patients, and is capable of automatic image segmentation and wound region identification. Several commonly used segmentation methods are utilized with their parameters fine-tuned automatically to obtain a collection of candidate wound regions. Two different types of Artificial Neural Networks (ANNs), the Multi-Layer Perceptron (MLP) and the Radial Basis Function (RBF) with parameters determined by a cross-validation approach, are then applied with supervised learning in the prediction procedure for the wound identification, and their results are compared. The satisfactory results obtained by this system make it a promising tool to assist in the field of clinical wound evaluation.