Published Paper Details:

Skin cancer, among the deadliest cancers, poses a significant threat when not detected and treated promptly. Sun exposure accelerates the proliferation of skin cells, leading to its development. Early detection is crucial to prevent its spread to other parts of the body. This study proposes a computerized technique for skin cancer classification, capitalizing deep convolutional neural networks (CNNs) to enhance diagnostic accuracy and efficiency. The dataset encompasses nine distinct types of skin cancer: seborrheic keratosis, actinic keratosis, benign keratosis, nevus, vascular lesions, basal cell carcinoma, dermatofibroma, melanoma, and squamous cell carcinoma, the aim is to develop a CNN model capable of accurately diagnosing and categorizing skin cancer into these classes. By integrating image processing and deep learning techniques, augmented with various image augmentation strategies, the dataset's diversity is enhanced, thereby improving the model's robustness and generalization capability. The eventual aim is to achieve notable performance metrics for classification tasks adopting the CNN approach. The expected outcomes include a weighted average precision of 0.88, a weighted average recall of 0.74, a weighted average f1-score of 0.80, and an overall accuracy of 90.51%. These metrics serve as benchmarks for evaluating the effectiveness and reliability of the suggested CNN methodology in diagnosing skin cancer.