Robust zero-watermarking scheme based on a depthwise overparameterized VGG network in healthcare information security

Huang, Tongyuan; Xu, Jia; Tu, Shixin; Han, Baoru

doi:10.1016/j.bspc.2022.104478

Cited by 15 publications

(11 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Subsequently, we will organize and compare the experimental and analytical results obtained for each image under different attack types and intensities. In this section, we will not only conduct a comprehensive analysis and comparison of the experiments conducted with our designed framework but also extend our analysis to compare the results with other recent methods that utilize zero-watermarking, such as Xing et al [12], Huang et al [13], and Liu et al [14]. This comparative analysis aims to provide a more objective and impartial perspective on the effectiveness of our proposed method in the context of the latest approaches in the field of zero-watermarking.…”

Section: Experimental Analysis and Resultsmentioning

confidence: 99%

“…Next, our proposed method will be compared to the methods proposed by Xing et al [12], Huang et al [13], and Liu et al [14]. All of methods are under the same zero-watermark framework but with different embedding algorithms, we will utilize NC as the evaluation metric and comparing.…”

Section: Resultsmentioning

confidence: 99%

“…Overall, our proposed method demonstrates stability and robustness compared to this method. Comparing [14] on the NC Metric. Table 6 presents the results of comparisons with the zero-watermarking framework proposed by Liu et al [14].…”

Section: Resultsmentioning

confidence: 99%

“…Comparing [14] on the NC Metric. Table 6 presents the results of comparisons with the zero-watermarking framework proposed by Liu et al [14]. The compared methods were tested under predominantly geometric attacks.…”

Section: Resultsmentioning

confidence: 99%

“…Huang et al [13] employs a pre-trained DO-VGG model to extract deep abstract features from medical images and generates the zero-watermark using a perceptual hashing algorithm. Additionally, Liu et al [14] introduced a method that combines Local Binary Patterns (LBP) with Discrete Cosine Transform (DCT). It extracts low-frequency feature vectors from digital images, performs hash sequence transformation, and binarization to embed the watermark.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Robust Zero-Watermarking by Circular Features and 1-D NRDPWT Transformation

Hsiu-Chi Tseng,

Hsiu-Chi Tseng

2024

Journal of Computers

View full text Add to dashboard Cite

<p>This paper introduces a secure and robust zero-watermarking framework that leverages the advantages of zero-watermarking, ensuring non-destructive modification of original images and unlimited capacity. The proposed method enables robust watermark embedding while preserving the original image. It employs a novel feature extraction approach using circular areas based on image radius, enhancing feature resilience. Additionally, applying one-dimensional non-recursive discrete periodized wavelet transform (1-D NRDPWT) converts feature values into phi, contributing to enhanced stability and robustness. Enhanced security is achieved through the use of Shuffle and Pseudo-Random Number Generator (PRNG). Experimental results, evaluated using metrics such as Bit Error Rate (BER) and Normalized Correlation (NC), validate the exceptional performance of this watermarking technique. These findings underscore the framework’s robustness, security, reliability, and integrity against both general and geometric noise attacks, making it a secure and robust solution for modern digital image copyright protection. In summary, our method offers an effective defense against various noise attacks while ensuring the highest watermark quality without compromising the original image. It is a significant advancement in copyright protection applications.</p> <p> </p>

show abstract

Section: Experimental Analysis and Resultsmentioning

confidence: 99%