An efficient edge-based segmentation model with correlation weighted linked feature vector for melanoma detection using VGG16

Melanoma is one of the rarest forms of skin cancer and one of the leading causes of cancer-related deaths. Nonetheless, a number of imaging procedures can be employed to identify the spread of disease to other organs or tissues when a primary diagnosis or clinical suspicion is present. The proposed model makes use of a VGG16, a Convolutional Neural Networks (CNNs) model for accurate detection of melanoma. The model is fine-tuned using a diverse dataset and VGG-16 deep hierarchical feature extraction capabilities to solve class imbalance and image variances. Model accuracy and generalizability are fine-tuned through the use of augmentation and transfer learning techniques. This research proposes a Edge based Segmentation model with Correlation Weighted Linked Feature Vector for Melanoma Detection using VGG16 (ESCWLFV-VGG16) for performing image processing operations and accurate disease detection. During training, hyperparameters are fine-tuned and tested on a dataset and performance metrics like accuracy, precision, recall, and F1 score are evaluated. When compared to competing models, the VGG16 model attained an accuracy of 98.6%. This research is helpful for medical experts for early and accurate detection of melanoma in earlier stages.