A Diagnose system for Skin Lesion using Deep Convolution Neural Network (DCNN)

Authors

  • J. Ranjani, Preetha. S, Janaki G , Deepika Sree D

Abstract

A Skin lesion is an abnormal lump, bump, sore or colored area on the skin. Skin lesion identification is a key step towards dermatological diagnosis. The proposed system uses multiple pre-processing methods to increase the accuracy of the automated labelling of the images collected. It characterizes the skin lesion by image acquisition, pre-processing and segmentation techniques to define the skin features. The pre-processed images are segmented using Gaussian Mixture Model (GMM). The features are extracted by digital image processing method called Principal Component Analysis (PCA). The extracted features are then classified using the Deep Convolution Neural Network (DCNN). Thus, deep learning helps in increasing the accuracy of the automated initial diagnosis and characterization.

Downloads

Published

2020-05-18

Issue

Section

Articles