The Real-time Secure Chat System is a web-based project that allows real-time and secure data transfer that will ensure secure chatting between users with end-to-end encryption and point-to-point security in individual or group conversations with cutting-edge authentication and authorization mechanisms. This real-time chat web application will allow the users to share data with each other in real-time with foolproof security features without being viewed or hijacked by any third party while transmitting from source to destination. All the messages will be end-to-end encrypted using AES 256 Encryption Algorithm and point-to-point configured using JWT, this mechanism ensures the double encryption on the message. Users will be allowed to chat in their rooms and in global. Only the authenticated user /owner will be able to view his chats and can send to others. The database associated with application is also encrypted this means that we as developers of application are not also able to view user chats, only the intended receivers are able to decrypt the message no one else (including developers, agencies, or any middleman) cannot change or see the message by changing database or by hijacking the transmission.

SECURE MEDICAL IMAGING DATA USING CRYPTOGRAPHY WITH CLASSIFICATION

Medical imaging data in today’s healthcare information systems is an essential part of diagnostics. The secure medical imaging data plays a critical role in current time but today it is complex task of maintaining data privacy so the main objective of this study to solve this problem. In this project firstly we secure the MRI images of the brain using cryptography. In this process input images are encrypted & decrypted using public key cryptography and supplied as an input to the pre-trained convolutional neural network such as Alex-net. The model comprises of the 25 layers such as convolutional, batch-normalization, ReLU and max-pooling etc. The classification between the tumor/healthy images has been performed using softmax layer. The performance of the proposed model has been tested on publically available BRATS-2020 Challenging dataset. The proposed model achieved up to the 95% prediction scores that are far better as compared to the latest published research work in this domain.

Educational Website