Secure Object Detection with Disease Diagnosis in Medical Images Using CNN With Asymmetric Algorithm

Abstract

Security is the most critical issue amid transmission of medical images because it contains sensitive information of patients. Medical image security is an essential method for secure the sensitive data when computerized images and their relevant patient data are transmitted across public networks. Sensitive images carry extensive important information and different features compared to standard images. Medical images have much more sensitive and essential information than any other digital image. Each pixel in the image can be necessary for the diagnosis process, and any deformation can result in a faulty diagnosis. The most robust securing of these images affects an image to the extent that it can be ignored; this is different from insensitive imagery as the border of redundancy is very low. The embedding capacity in medical images is deficient. Existing researchers present different data security techniques as cryptography and data hiding to guarantee data verification. But these approaches take more time and less security in medical image application. So in this project, implement Fragmented based Elliptical curve cryptography with Convolutional neural network algorithm to provide secure disease diagnosis system for medical images. Experimental results shows that the proposed system implemented Lung CT scan images that are collected from Open medical data sources and with high level security.