Novel skin cancer Detection based transfer learning with optimization algorithm using Dermatology Dataset

Abstract

Detecting skin cancer at the preliminary stage is a challenging issue, and is of high significance for the affected patients. Here, Fractional Gazelle Optimization Algorithm_Convolutional Neural Network based Transfer Learning with Visual Geometric Group-16 (FGOA_CNN based TL with VGG-16) is introduced for primary prediction of skin cancer. Initially, input skin data is acquired from the database and it is fed to the data preprocessing. Here, data preprocessing is done by missing value imputation and linear normalization. Once data is preprocessed, the feature selection is done by the proposed FGOA. Here, the proposed FGOA is an integration of Fractional Calculus (FC) and Gazelle Optimization Algorithm (GOA). After that, skin cancer detection is carried out using CNN-based TL with VGG-16, which is trained by the proposed FGOA and it is an integration of FC and GOA. Moreover, the efficiency of the proposed FGOA_ CNN-based TL with VGG-16 is examined based on five various metrics, like accuracy, Positive Predictive Value (PPV), True Positive Rate (TPR), True Negative Rate (TNR), and Negative Predictive Value (NPV) and the outcome of experimentation reveals that the devised work is highly superior and has attained maximal values of metrics is 92.65%, 90.35%, 91.48%, 93.56%, 90.77% respectively.