A Computation-Efficient Block-Based Scrambling Technique with Random Control Parameters

In this paper, an image scrambling technique based on chaotic systems is proposed. Scrambling is performed in transform domain to overcome the drawbacks of spatial domain scrambling. The transformed image is first decomposed into blocks of coefficients which are scrambled using 2D chaotic map; this process is repeated for various levels. At each level, the control parameters of scrambling are randomly generated through another 2D chaotic system to make the process of scrambling key-dependent. Experimental analyses show that the technique is efficient and provides good scrambling effect, high security with less computation. It is also shown that discrete wavelet transform implementation has better performance than discrete cosine transform for a noisy transmission channel