Abstract:
In frequency domain based image steganographic
methods, especially Discrete Cosine Transformation (DCT), the
selection of rounded quantized DCT coefficients in each block for
data embedding is still an open issue. This paper presents a new
data hiding technique based on the DCT coefficients and
modified quantization table values. Embedding strength of each
coefficients are determined by the mathematical formula which
compare the DCT coefficient and appropriate quantization table
value in order then the secrete bits are embedded in frequency
components of the quantized DCT coefficients using least
significant bit (LSB) method to enable large embedding capacity
without image degradation. Further, we investigate this
technique for the different cover images in different quality
factors by applying scaled and modified quantization tables. A
comparison of image quality and embedding capacity between
the proposed method and some of other existing methods in
literature are analysed. The performance of our proposed
method produce satisfactory results compared with other
methods.
