Abstract—In this paper we present an improved steganographic algorithm based on the modified LSB method in four different wavelet transform domains, viz, Wavelet (wavcdf97), Slantlet, Double Density Dual Tree Discrete Wavelet transform and Double Density Dual Tree Complex Wavelet, respectively. The characteristics of the Complex Wavelet Transform that provides shift invariance and offers better directional selectivity makes it a better choice for data hiding over the classical discrete wavelet transform. For achieving the security and robustness, the original message is first converted into binary cell array with the help of self-synchronizing variable length codes, viz., T-codes before the embedding. The use of T-codes does not only compress the message but also provides self-synchronization at decoding stage. The secret message so obtained is embedded in the insensitive high sub-bands obtained from the cover image after applying these transform. The metric used for evaluating the visual quality of stego-image is PSNR. The experimental results show that the dual tree complex wavelet transform is better option than the Wavelet, Wavelet-like transform, i.e., Slantlet transform and Double Density Dual Tree DWT for data hiding in terms of visual quality and embedding capacity, though poor to external attacks such as Gaussian.

Index Terms—Steganography, DWT, SLT, DD DT DWT, DT CWT, T-codes, PSNR.

Sushil Kumar is with Department of Mathematics, Rajdhani College, University of Delhi, New Delhi, India (e-mail: skazad@rajdhani.du.ac.in).
S. K. Muttoo is with Department of Computer Science, University of Delhi, Delhi, India (e-mail: skmuttoo@cs.du.ac.in).