An interval type-2 fuzzy edge detection and matrix coding approach for color image adaptive steganography

This paper proposes an adaptive color image steganography technique based on edge detection and matrix coding, by comprehensively considering the human eye sensitivity to edges and the RGB color components. First, non-edge and edge points are detected by computing interval type-2 fuzzy similarity on cover image after noise estimation and filtering. Then, a genetic algorithm is used to optimize the embedding bits (edge and non-edge in RGB channels) according to the capacity of the secret message. Next, a chaotic method and random sequence scrambling are used to prevent the secret message from attacks. Finally, three matrix coding schemes with different embedding efficiency (Least significant bit, Hamming code, and XOR coding) are used to embed the encrypted confidential information. The proposed method hides a large amount of data with good quality of stego image from the human visual system (HVS) and guarantees the confidentiality of communication. Experiments showed that it outperformed several state-of-the-art approaches in terms of payload, mean square error, peak signal-to-noise ratio, and structural similarity. The robustness of the method is also tested by RS steganalysis and pixel difference histogram (PDH) analysis.