DCT coefficient dependent quantization table modification steganographic algorithm

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.