Robust Spatial-spread Deep Neural Image Watermarking

TL;DR

This paper introduces a robust deep neural network-based image watermarking method that spreads information across the spatial domain, achieving high resistance to various distortions like JPEG compression and blurring.

Contribution

It presents a novel end-to-end neural network approach with adversarial training and noiser layers, enhancing robustness and generalization against multiple image attacks.

Findings

High robustness against JPEG compression, blurring, and resizing.

Effective end-to-end neural network model for watermark embedding and recovery.

Proposed differentiable JPEG approximation for training.

Abstract

Watermarking is an operation of embedding an information into an image in a way that allows to identify ownership of the image despite applying some distortions on it. In this paper, we presented a novel end-to-end solution for embedding and recovering the watermark in the digital image using convolutional neural networks. The method is based on spreading the message over the spatial domain of the image, hence reducing the "local bits per pixel" capacity. To obtain the model we used adversarial training and applied noiser layers between the encoder and the decoder. Moreover, we broadened the spectrum of typically considered attacks on the watermark and by grouping the attacks according to their scope, we achieved high general robustness, most notably against JPEG compression, Gaussian blurring, subsampling or resizing. To help us in the models training we also proposed a precise differentiable approximation of JPEG.