Secure 2D barcodes based on visual cryptography

Abstract

In this thesis we investigate the application of visual cryptography to enforce 2D barcode confidentiality. The approach of visual cryptography is appealing because 2D barcodes are indeed images, but it presents non-trivial challenges. First of all, visual cryptography aims at sharing a unique secret between two or more participants while here we would like to provide each participant with a visual decryption key that can be used to recover the participant's distinguished secret; moreover, visual cryptography assumes that images are pre-shared among participants which is not practical in many situations. We contribute to the state-of-the-art by implementing and experimenting with two extensions to visual cryptography proposed together with the colleague Nicolò Ghiotto. The first scheme allows for generating a single visual ciphertext that can be decrypted by each participant’s visual key, giving a different (plaintext) 2D barcode. This scheme presents a practical limitation: the visual key is a random image that is very hard to scan and needs to be pre-shared. The second scheme extends the first one by permitting to derive the visual key from a password using what we call a visual key derivation function. This removes the limitation of the first scheme but introduces a scalability issue for what concerns the number of participants. We present experiments and discuss limitations with respect to real use cases.