Open AccessConvolutional neural network for classification of skin cancer based on image data using google colab
Author(s) -
Iqbal Kharisudin,
Afif Nurul Hidayati,
Arief Agoestanto,
MT Muhammad Mashuri
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1968/1/012015
Subject(s) - convolutional neural network , confusion matrix , skin cancer , computer science , artificial neural network , deep learning , artificial intelligence , confusion , software , cancer , image (mathematics) , pattern recognition (psychology) , machine learning , medicine , psychology , psychoanalysis , programming language
Climate change causes the world’s weather to become hotter and has an impact on human health. The direct impact that can be seen is the increase in skin cancer cases due to rising temperatures. This study aims to perform digital image data classification modeling by implementing the Convolutional Neural Network (CNN) method in skin cancer cases using Google Colab software. Research on deep learning applications for identifying and classification image data has been carried out in many recent articles. We used secondary skin cancer image data obtained by a dermoscopy consisting of malignant and benign skin cancer. From 3297, there are 1,800 images of benign skin cancer and 1,497 images of malignant skin cancer. For modeling purposes, it was divided into 2967 training data and 330 testing data. The training process uses variations of the epoch and learning rate to determine the best results. The accuracy value obtained is 99.60% and the validation accuracy value is 92.12%. These results were obtained using 100 epochs and a learning rate of 0.00001. Based on the prediction results using a confusion matrix for testing data, the accuracy value is 90%.