Skin Cancer Clasification Using Region Growing & Recurrent Neural Network

Abstract

Skin cancer is a disease caused by mutations in the skin of cells. Melanoma and non-melanoma skin cancers are the two basic classifications for skin cancer. In addition, it is estimated that 5.9-7.8% of all cancer cases each year involve skin cancer. In Indonesia, 65.5% of skin cancers are basal cell carcinomas, followed by 23.0% squamous cell carcinomas and 7.9% malignant melanomas. When not discovered early, melanoma skin cancer can result in a high fatality rate. Basal cell carcinoma and squamous cell carcinoma are two examples of nonmelanoma skin cancers (NMSCs), which are far more frequent but far less likely to spread and cause mortality. Diagnosis made by an expert or doctor takes a long time and is often inconsistent because the environment and personal conditions influence the expert's condition. To minimize this problem, this research aims to introduce image processing methods for early skin cancer detection, the Region growing method, and artificial neural networks RNN for classification. It is hoped that this method of early cancer detection can be done quickly and does not require much money. This study will use two methods to detect skin cancer: the Region growing method and RNN-LSTM. This research aims to introduce the Region of interest (ROI) method and artificial neural networks to detect skin cancer.

Keywords—Cancer skin, Region of Interest, Region Growing, Recurrent Neural Network, LSTM