Lossless Data Hiding Based on Improved Vq Index Joint Neighboring Coding

Data hiding is a form of steganography to embed secret data into digital media for the purpose of content authentication, annotation and copyright protection. Lossless data hiding is a special kind of data hiding technique that guarantees the digital media can be losslessly restored after the secret data are correctly extracted. In the past several years, several reversible data hiding schemes for Vector Quantization (VQ) compressed images have been proposed. Some embed data during the VQ encoding process such as Side-Match VQ (SMVQ)-based and Modified Fast Correlation VQ (MFCVQ)based methods; others embed data during the VQ index streaming process such as the Joint Neighboring Coding (JNC)-based scheme. This paper presents an improved JNC (IJNC) process for both lossless data hiding and lossless index compression. The JNC process performs data hiding on each index outside the seed area, while the proposed IJNC process performs data hiding on each 2×2 index block without any seed area. Experimental results show that our scheme outperforms three recently proposed schemes, namely JNC-based, SMVQ-based and MFCVQ-based data hiding methods, in terms of the hiding capacity and the stego image quality.