Institut d'Électronique et de Télécommunications de Rennes

| ENT |

Accueil du site > Accueil > Production scientifique > Thèses en cours > Équipe VAADER (Dpt IMAGE)

Design and efficient implementation of one-way hash functions based on chaotic maps and neural networks - Conception et mise en oeuvre efficace de fonctions de hachage à sens unique basées sur des cartes chaotiques et réseaux neuronaux.

Doctorant : Nabil ABDOUN.

Directeurs de thèse : Safwan EL ASSAD, Olivier DEFORGES.

Thèse débutée le : 02/02/2015.

Hash functions are one of the most useful primitives in cryptography. Indeed, they play an important role in integrity protection, message authentication and digital signature. Hash function is a one-way function that maps an arbitrary finite length message to a fixedlength quantity. They must fulfill some security properties such as collision resistance or preimage resistance. The typical building blocks of these functions, is the domain extension algorithm proposed by Merkle and Damgard.

Neural networks’ confusion and diffusion properties have been used to design encryption algorithms. In fact, neural networks have also a one-way property. Compared with the simple chaotic map, chaotic neural network has stronger spatiotemporal complexity and mixture. Moreover, most existing chaotic hash functions can only run in a sequential mode. Namely the processing of the current message unit cannot start until the previous one has been processed. In this way, the efficiency has inevitably been reduced greatly. To break through this limitation, we propose in this project to study parallel keyed hash functions based on neural networks and to compare their performance with sequential mode. These structures must achieve the following requirements : high message and key sensitivity, good statistical properties, collision resistance and secure against meet-in-the middle attacks.

This thesis will jointly consider the robustness against cryptanalysis and the computational performance of several Chaos-based Hash functions in order to explore the tradeoff between robustness and performance.

Page du doctorant

Départements et équipes de recherche