COMPARATIVE ANALYSIS OF DEEP LEARNING AND MACHINE LEARNING APPROACHES FOR AUTOMATED SKIN CANCER DETECTION

Abstract

Skin cancer is a prevalent and potentially life-threatening disease with increasing global incidence rates. Early detection plays a crucial role in improving patient outcomes and survival rates. This thesis presents a comprehensive study on the development and evaluation of an automated approach for skin cancer detection, aiming to assist clinicians in accurate and timely diagnosis. The research focuses on leveraging advanced machine learning and computer vision techniques to analyze dermoscopic images and identify potential malignancies. The dataset utilized consists of a diverse collection of annotated skin lesion images, encompassing various types and stages of skin cancer. The primary objective is to develop a robust and reliable model capable of distinguishing between benign and malignant lesions. The thesis explores different aspects of the automated skin cancer detection pipeline, starting from data preprocessing and augmentation techniques to enhance the model's generalization capabilities. Various feature extraction methods, including handcrafted features and deep learning-based representations, are investigated to capture relevant patterns and discriminative information from the images. Multiple classification algorithms are studied to compare their performance and determine the most effective approach. Evaluation metrics such as accuracy, precision, recall, and area under the receiver operating characteristic curve (AUC-ROC) are utilized to assess the models' diagnostic accuracy and robustness.