Convolutional neural network and 2D logistic-adjusted-Chebyshev-based zero-watermarking of color images

Abstract Robust zero-watermarking is a protection of copyright approach that both effective and distortion-free, it has grown into core research on the subject digital watermarking. This paper proposes revolutionary for color images using convolutional neural networks (CNN) 2D logistic-adjusted Chebyshev map (2D-LACM). In this algorithm, we first extracted deep feature maps from an original image pre-trained VGG19. These were then fused featured image, owner's watermark sequence was incorporated XOR operation. Finally, 2D-LACM encrypts scrambles binary matrix to ensure security. The experimental results show proposed algorithm performs well in terms imperceptibility robustness. BER values watermarks below 0.0044 NCC above 0.9929, while average PSNR attacked 33.1537 dB. Also, superior other algorithms robustness conventional processing geometric attacks.