Early Skin Cancer Detection Using CNN-ABCD Rule Based Feature Extraction Classification and K-Means Clustering algorithm through Android Mobile Application
M. Salomi, Gunashekar Daram, Sonti Sri Harshitha
Abstract
Our proposed system focuses on the development of a skin cancer prediction model using machine learning algorithms. Skin cancer is the most common cancer disease in almost all the country. Current estimates are more than a million cases are witnessing skin cancer in their lifetime. As per the statistics, the north region of India stands first with 100,000 skin for both males and females with 1.62lakhs and 1.21lakhs, respectively. This proposed system's aim is to create an accurate and efficient diagnostic mobile application to classify skin cancer based on a dataset of skin lesion images obtained from the Kaggle data source. Advanced machine learning techniques such as ABCD Feature rule is employed in the proposed model to achieve high predictive accuracy, and the model outcome will undergo comprehensive performance analysis, evaluating metrics such as sensitivity, specificity and accuracy. The significance is underscored by the growing global incidence of skin cancer, as early detection is crucial for improving patient outcomes. The versatile tool such as an android mobile application generated by the proposed model can be deployed in diverse clinical settings, aiding dermatologists and medical professionals in making timely and informed decisions for accurate diagnosis and effective treatment plans. The research methodology is based on qualitative analysis where various kinds of literature are being reviewed based on machine learning.