Chakrabarty, AmitabhaRahman, Md. TawsifurAzad, Md. Siam SadmanMuhtasim, Ali2024-07-032024-07-03©20232023-05ID 20141027ID 20141002ID 17301163http://hdl.handle.net/10361/23649Cataloged from PDF version of thesis.Includes bibliographical references (pages 47-48).This thesis is submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Computer Science, 2023.Machine learning (ML) for skin lesion identification employs algorithms, notably convolutional neural networks (CNNs), to categorize and detect skin lesions, aiming to enhance early detection and treatment of skin cancer. CNNs, trained on diverse lesion images, excel in learning features for classification, often rivaling dermatologists’ accuracy. Recent studies demonstrate CNNs’ effectiveness, achieving accuracy comparable to or surpassing dermatologists. Ongoing research focuses on addressing challenges like dataset diversity and robust evaluation metrics. Despite obstacles, ML’s potential to enhance early melanoma detection remains significant, promising to save lives through improved diagnosis and treatment. Notably, our research explored a hybrid approach, combining ResNet50v2 and InceptionV3 models trained on GAN-generated data. This innovative strategy achieved a notable 77% accuracy, showcasing promising results in advancing muticlass skin lesion identification accuracy.58 pagesenBrac University theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission.Convolutional neural networkMachine learningCancerResNet50v2Inception V3GANDisease detectionDiagnostic imagingNeural networks (Computer science)Cancer--DiagnosisThesis