Low-complexity index assignments for secure quantization

Abstract

Focusing on the design of scalar quantizers for secure communication in a wiretap environment we propose a low-complexity encoding method that minimizes the end-to-end distortion between legitimate terminals while maximizing the distortion at the eavesdropping receiver. The key idea is to generate confusion by means of randomized index assignments. A simple iterative optimization scheme is shown to yield low distortion mappings with noticeable privacy benefits.