A Smartphone Application for Skin Lesion Detection and Classification with Deep Learning Algorithms

Mary, XA ORCID: https://orcid.org/0000-0002-7284-1541, Winifred, P, Evangeline, CS ORCID: https://orcid.org/0000-0002-1072-6724, Kumaravelu, VB ORCID: https://orcid.org/0000-0002-1778-9891, Karthik, C, Chowdhury, S, Rabie, K ORCID: https://orcid.org/0000-0002-9784-3703, Thu, BT and Tung, VD (2022) A Smartphone Application for Skin Lesion Detection and Classification with Deep Learning Algorithms. International Journal of Intelligent Systems and Applications in Engineering, 10 (3). pp. 34-39.

Preview

Published Version Available under License Creative Commons Attribution Share Alike. Download (413kB) | Preview

Abstract

The Skin Lesion (SL) classification has recently received a lot of attention. Because of the significant resemblance between these skin lesions, physicians spend a lot of time analyzing them. A Deep Learning (DL) based automated categorization system can help clinicians recognize the type of SL and improve the patient's health. In this research, DL approaches such as VGG-16, ResNet-50 and customized model are employed to detect the SL using a smartphone application. These models are trained on the SL classification dataset from the International Skin Imaging Collaboration (ISIC) 2019. The customized model over fits the other two models with a validation accuracy of 86.21%, whereas the validation accuracy of VGG-16 and ResNet-50 is 85.15% and 84.82%, respectively. Physicians will save time and have a higher precision rate in the automatic classification of SL utilizing DL.