CLASSIFICATION OF BENIGN MELANOCYTIC SKIN LESION USING ABCD FEATURES AND CONVOLUTIONAL NEURAL NETWORK (CNN)

[ 31 Dec 2020 | vol. 13 | no. 4 | pp. 23-32 ]

About Authors:

Sallar Khan1, Payal Matani2, Rabeesa Shakeel Siddiqui3, Rabiya Tahir4 and Syeda Sara Ashraf5
-1,2,3,4,5Sir Syed University of Engineering and Technology, Pakistan

Abstract:

Skin cancer and its early detection are very critical and crucial, also among humans it is one of the very dangerous form of cancer which is hurriedly spreading globally. While in the current era of technology mobile apps are getting more usage and attention among the users, where mobile based skin cancer detection systems are rarely found and mostly are paid applications. In this research, researchers proposed a real-time detection of skin cancer type: Benign Melanocytic Skin Lesion (BMSL) through android based mobile application, in which we have used ABCD features for feature extraction process and Convolutional Neural Network (CNN) for the classification accuracy purposes. In terms of results, the classification accuracy of 95% has been achieved in the training phase, while we are working on testing phase and results will be shown in our future research.

Keywords:

Skin Cancer, Benign Melanocytic Skin Lesion, Melanoma, Skin Disease, Android Application, Health Care

 

About this Article: