Securing Skin Cancer Detection using CNN Secured with Graphical Password

Abstract

Skin cancer is measured as a major contributor to the causes of deaths around the world. The deathrate due to this disease is above all other skin related consolidated malignancies. In this paper, we utilize a dataset of dermoscopic images collected from different documents labelled and validated by professional dermatologists. In our work, we manually trained a resource controlled CNN model using transfer learning for binary (0/1) classification of skin lesions into two classes i.e. benign or malignant classes. CNN model is used because recent advancements in image processing using Deep Convolutional Neural Networks (CNN) have led numerous researchers to use them for skin lesion classification which concluded that CNN performed on par with expert dermatologists. Graphical passwords method widely used for authentication in today’s mobile computing environment, this methodology was announced to enhance security element and overcome the weaknesses of existing passwords methods. In this paper we consider the method of user authentication using graphical password system which is based on stenographical methods. The study describes the generalized model of user authorization in secure systems with utilization of computerized watermarks (digital watermarks),a definitive model of the graphic password system, and lists the foremost requirements for digital watermarks to be utilized in the graphics systems of user authorization for protected information and telecommunication resources.