Interactive diagnosis of nevus' classification using FWS model & feature-based analysis

Abstract

A Computer-Assisted Diagnostic System for Decision-making in early detection of the nevus as either benign or dysplastic or melanocytic, was developed with dynamic features of Analysis and design. The system, presents a Feature-based Classification Module combined with a new model called Fourier-Wavelet-Statistical (FWS), supported by the heuristic knowledge for Image Analysis. The design of the system is divided into 3 stages: The FWS model, Classification Module & Interactive Interface. The FWS Model was Comprised of 3 Parallel sub-blocks to which the image is fed, viz., the Edge-Axis algorithm, Multi-level Discrete Wavelet Transforms (MDWT) Process Algorithm & the Statistical Node. The Classification module was designed with a Classifier & Knowledge Data Base (KDB) that follows a Heuristic approach in identifying the nevus as of specific type. The Interactive Interface block is a challenging concept to the designer in handling the collective perspectives of experts, practitioners, and engineers. The design model is efficient, in improving the decision-making strategy and overall accuracy of the system developed; in order to assist a general practitioner in identifying the typical skin lesion under a suggestive and supportive environment.