A Comparative Study Of CNN And SVM With Particle Swarm Optimization For Skin Cancer Detection In CT Images

Abstract

This research looks at how well deep learning and traditional machine learning methods can detect skin cancer using CT images. It uses a Convolutional Neural Network (CNN) to automatically extract features and classify images, while a Support Vector Machine (SVM) is fine-tuned with Particle Swarm Optimization (PSO) to improve its accuracy. The performance of both models is measured using Accuracy, Precision, Recall, F1-Score, and AUC-ROC. The results highlight the advantages and limitations of each method in terms of how accurately and efficiently they classify the images.