A robust and imperceptible n-Ary based image steganography in DCT domain for secure communication

Generally, spatial domain and transform domain steganography techniques are used to achieve the covert communication. The main drawback of embedding in frequency domain is rounding error appeared in the embedded secret message. Information is lost when integer pixel values are restored from the frequency domain, causing the embedded message to be distorted. A minor distortion of the embedded message renders extraction impossible and results in loss of embedded secret message. In the present study, a novel imperceptible Discrete Cosine Transform (DCT) and chaotic map based steganographic algorithm for secure communication has been proposed. The proposed method performs an adaptive n-Ary secret data embedding in the higher frequency DCT coefficients and removes the distortion. Additionally, the security of secret data is enhanced using chaotic maps to select the random DCT coefficients for embedding. The experimental results confirm that the proposed technique has high embedding capacity (EC), excellent stego-image quality and capability to withstand against malicious users. To examine security of the proposed method, the key space and key sensitivity parameter were used. Maximum EC = 425,780 bits was achieved with Peak Signal to Noise Ratio (PSNR) = 32.2218, 29.8804, 32.2507, 33.3327, 32.3963 and 31.4936 dB in case of Lena, Baboon, Jetplane, Elaine, Pepper and Boat respectively. The evaluated results show that the proposed method outperforms the existing steganography techniques in terms of EC and PSNR. Additionally, quality of stego image is identified by universal image quality index (Q), Bit Error Rate (BER) and Structural Similarity Index (SSIM).