High dynamic range image watermarking method based on Tucker decomposition

Abstract

The invention discloses a high dynamic range image watermarking method based on Tucker decomposition. The method comprises a watermark embedding part and a watermark extracting part, and is characterized by during a watermark embedding process, expressing a host image in a high dynamic range as a form of a third-order tensor; then, utilizing the Tucker 3 decomposition to process the high-dynamic-range host image; taking the obtained first feature map of the core tensor as an embedding carrier of the watermark information, so that the main energy of the high-dynamic-range host image is covered by the first feature map of the core tensor, and the embedded watermark information can be diffused into an R channel, a G channel and a B channel of the high-dynamic-range watermark image along with the inverse transformation of decomposition, and accordingly the robustness is effectively improved. During a watermark extraction process, an original high-dynamic-range host image is not needed, so that the blind extraction of the watermark information is realized. In addition, according to the method, the internal structure of the high-dynamic-range image cannot be damaged, and a better effect is achieved at the aspects of invisibility and embedding capacity.

Description

technical field [0001] The invention relates to an image watermarking method, in particular to a high dynamic range image watermarking method based on Tucker decomposition. Background technique [0002] With the wide application of high dynamic range (High Dynamic Range, HDR) imaging technology in the fields of consumer electronics, image production, virtual reality, remote sensing detection and medical detection, the copyright protection of corresponding high dynamic range images is also attracting increasing attention. . Different from traditional low dynamic range (Low Dynamic Range, LDR) images, high dynamic range images are usually represented and stored in floating-point data, and their dynamic range can generally reach or exceed 9 orders of magnitude, which can not only be used in To a certain extent, the accuracy of scene light information is improved, and it can also bring richer color details and light and dark levels. However, since the existing low dynamic rang...

AI Technical Summary

Problems solved by technology

To sum up, although the existing robust watermarking algorithms for high dynamic range images have achieved certain results in resisting minority tone mapping attacks, such algorithms have not fully considered the following two issues: First, DCT, DWT Traditional data processing methods such as Redundant Discrete Wavelet Transform (RDWT), NSCT and SVD all convert high-dimensional data into vectors to represent, but this will not only cause the dimensionality of data samples to be too high, but also It will also destroy the internal structure of the data itself; second, tone mapping is a nonlinear mapping process, that is, the watermark image will show inconsistent robustness in different brightness regions after being attacked by tone mapping

Images