The book presents three new methods of improving robustness of spatial and transform domain steganography without compromising with the minimum acceptable perceptual quality of the cover image containing hidden data. The suggested methods of spatial domain hiding perform data hiding into higher nibbles or moderatory bits (i.e. bits '2'' to '5'') of an 8-bit host image pixel using data matching or substitution with pixel gray value error minimization respectively. Both these techniques offer higher robustness as compared to commonly used LSB technique. In addition a new algorithm for achieving higher robustness as compared to generic DCT algorithm has been proposed based on long-tail modeling of mid-frequency DCT coefficients. The work contained also includes study on robust medical data hiding along with compaction using computationally simple Haar wavelet transform for illustrating its feasibility in secure telemedicine application along with a new effective technique of steganalysis based on generating relative frequency histograms.