Segmentasi dan Klasifikasi Gambar Citra pada Kanker Kulit Menggunakan Metode Convolutional Neural Network (CNN) dengan Arsitektur ResNet-50

Skin cancer is a type of cancer whose incidence continues to increase throughout the world. Early and accurate detection is very important to increase the chances of recovery. In this research, we developed a Convolutional Neural Network (CNN) based system using the ResNet-50 architecture for image segmentation and classification of skin cancer images. This method was chosen because ResNet-50 has the ability to overcome the problem of degradation of accuracy in very deep networks through the use of residual learning. This research involves several main stages, namely data collection and preprocessing, training a CNN model with the ResNet-50 architecture, and evaluating model performance. The data used was taken from Kaggle "Melanoma Skin Cancer Dataset Of 10000 Images" which has data for 10615 images which are divided into 2 classes, namely Malignant and Benign images. Data preprocessing includes image augmentation and normalization to improve data quality and model performance. The evaluation results show that the developed model is able to achieve high accuracy in image segmentation and classification of skin cancer images. The ResNet-50 model trained in this research succeeded in achieving an accuracy of 92.00%, precision of 92%, recall of 92%, and F1-score of 92% on the test dataset. These findings demonstrate that the CNN-based approach with ResNet-50 architecture is effective for skin cancer segmentation and classification tasks, and has great potential for application in computer-based diagnostic systems in the medical field.