Secret Image Sharing based on Vector Quantization
Lee Shu-Teng Chen, Wei-Kai Su, Ja‐Chen Lin · 2009
This paper proposes an (r, n) threshold secret image sharing scheme based on Vector Quantization (VQ). By using the host images as the VQ codebooks, the VQ indices of the secret image is computed and then shared among the n shadows. The created n shadows and the mixed information of the VQ codebooks are hidden in the original host images to form the n stego images. During the recovery phase, the secret image with VQ quality can be reconstructed by any r of the n stego images. The proposed method is therefore missing-allowable since n-r stego images can be lost in the recovery phase. Experiments show that the qualities of our stego images and recovered images are all acceptable. Moreover, the proposed method is secure because it is unlikely to reveal the secret image if less than r stego images are intercepted.