Tools
Tools
Gupta, Sarthak (2023) Skin Lesion Classification Based on Various Machine Learning Models Explained by Explainable Artificial Intelligence. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (1MB) | Preview |
Preview |
PDF (Configuration manual)
Download (1MB) | Preview |
Abstract
Over the past few years, many lives have been lost due to the spread of various deadly diseases. There has also been a big rise in deaths from skin cancer, and most of these people could have lived longer if the cancer had been detected at an early stage. Therefore, a skin lesion classification model is proposed in this project, which will help in classification of skin lesions. A major problem with using artificial intelligence in the medical sector is a lack of transparency. Therefore, Explainable Artificial Intelligence (XAI) is used to explain the proposed classification model built using deep learning and machine learning models such as Convolutional Neural Networks (CNN) and Extreme Gradient Boosting (XGB Classifier). Two explainable artificial methods, such as Local Interpretable Model-Agnostic Explanations (LIME) and SHapley Additive Explanations (SHAP), are used to explain the proposed models. The Convolutional Neural Networks (CNN) model performed on the augmented images produced an accuracy of 75.94%, which outperformed the other two models performed in this research. The explainable artificial intelligence method (SHAP) was used to explain the attribute importance, and the Local Interpretable Model-Agnostic Explanation (LIME) was used to explain how the model interprets images to predict its class.
Actions (login required)
 |
View Item |