Skin Cancer Detection Using Deep Learning

Abstract

Skin cancer is a life-threatening
condition characterized by the abnormal
growth of skin cells and early detection is
critical for successful treatment. In this
research, deep learning models such as
Dense Convolutional Network
(DenseNet) and VGG-19 Convolutional
Neural Network (CNN) are applied for
accurate classification of four types of
skin cancers melanoma, basal cell
carcinoma, squamous cell carcinoma,
and Merkel cell carcinoma — along with
healthy skin. Both models are trained and
tested on image datasets to differentiate
malignant from benign lesions.
DenseNet’s dense connectivity promotes
feature reuse, while VGG-19’s deep
architecture captures rich hierarchical
patterns. The performance of these
models is evaluated through metrics
including accuracy, sensitivity, and
specificity. Results show improved
classification efficiency in comparison to
traditional methods. The study
contributes to clinical decision support
systems by increasing detection reliability.
Early and accurate identification aids in
reducing mortality rates. Findings
demonstrate the potential of CNN-based
classification in real-world dermatology
diagnostics.