Optimization of Melanoma Skin Cancer Detection through Data Magnification, Filter Preprocessing, Image Enhancement, and Convolutional Techniques

Melanoma skin cancer is one of the most aggressive forms of cancer, requiring early detection to improve patient outcomes. This study evaluates three image processing methods—Laplacian, Box Blur, and Edge Detection—used in melanoma detection, analyzing their performance using Mean Squared Error (MSE) and Structural Similarity Index (SSIM) metrics. Among these, Box Blur demonstrated the best overall performance with the lowest average MSE (104.16), indicating minimal distortion in the processed images. Additionally, it achieved the highest SSIM score (0.851), suggesting that it best preserved the structural integrity of the images, making it the most effective in maintaining both quality and important diagnostic details. In contrast, Edge Detection produced the highest MSE (108.02) and a negative SSIM score (-0.016), significantly distorting image structure and making it less suitable for melanoma detection. Laplacian, while moderate in performance, did not outperform Box Blur, with an MSE of 106.99 and an SSIM of 0.175. These results highlight Box Blur as the most reliable technique for melanoma image analysis, ensuring both clarity and structural preservation. By effectively enhancing diagnostic features and reducing errors, Box Blur offers a valuable tool for clinicians aiming to improve diagnostic accuracy in melanoma detection.