A Self-Embedding Watermarking Image Encryption Algorithm Based on LBP and a Two-Dimensional Discrete Chaotic Model

In this paper, a digital image watermarking encryption scheme is proposed to address some security issues of image transmission and information copyright disputes. First, Local Binary Pattern (LBP) is used to obtain the texture information of the plaintext image and is embed into the plaintext image as a watermark to safeguard the copyright of the message. This extraction method is used because the obtained image texture information is highly similar to the plaintext image, which is not easy for the attacker to detect. Second, in the process of embedding the watermark, a combination of Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SVD) is used to decompose the low-frequency sub-band of the image and to embed the watermark in the region, which can improve the robustness of the watermark. Finally, the watermarked image is corrupted twice during the image encryption process. The first time, the image is corrupted using random numbers generated from chaotic sequences. The second time, the pixel positions are corrupted again by performing a remainder operation on the block matrix position index. In the diffusion part, the plaintext pixels are further hidden using a bidirectional XOR method. In this way, the attacker cannot easily access the image information. Thus, the research in this paper can provide some theoretical support to image encryption.