An efficient security system based on cancelable face recognition with blockchain over cognitive IoT

This paper presents a new authentication framework for cancelable face recognition biometrics. In recent years, biometric plays a pivotal role in Cognitive Internet of Things (C-IoT) security. The face trait solves a lot of security issues; it increases the resistance of these systems against severe authentication attacks especially in smart IoT-based applications. The proposed scheme runs in two phases; The first phase (enrollment), in which the face image of the person is hashed using SHA256, is passed to an algorithm to extract the main features. Then the image is decomposed using a wavelet transform algorithm. Wavelet transform is applied to hide the face details from the image. This step is done to protect the image against being stolen or modified by external hackers. The image, hash, and wavelet transform are passed through the raspberry Pi to be stored in a database. A block is added to the private blockchain with these files. In the second phase (authentication) the test image is hashed, passed to the system to extract the features, and compared with the stored images in the database. Then the decision is stored in a new block in the chain, whether the two images are identical or not. The simulation results based demonstrate that the proposed cancelable biometric system gives a very reliable and secure performance for securing the IoT applications.