An improved method for hiding data in a maze

Abstract

It is common to use image, audio, video streams as host media in steganography. In this paper, we will select mazes as host media to avoid the problem "robustness" concerned by classic host media. The original idea of embedding data in a maze is proposed by Niwayama et al. Their method has two disadvantages. One is very small embedding capacity; the other is that the embedded maze is not perfect. Here, we present an improved algorithm to increase the embedding capacity. Meanwhile, we can still preserve the property "perfect", that is, we can keep the imperceptible property required in steganography.