Scholarlink

<p>Ensuring the security and robustness of digital image watermarking has become increasingly important due to the growing risks of image tampering, copyright infringement, and unauthorised access. A key challenge in this domain is achieving a reliable balance between imperceptibility and resilience against diverse image-processing attacks. This paper proposes a secure and adaptive watermarking framework that integrates Arnold scrambling with a three-level discrete wavelet transform (3L-DWT). The watermark is first encrypted using an iteration-controlled Arnold map and then decomposed via 3L-DWT, where its LL<InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(_3\)</EquationSource> </InlineEquation> coefficients are embedded into the highest-entropy block of the host image’s LL<InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(_3\)</EquationSource> </InlineEquation> subband. A genetic algorithm (GA) is used to optimise the embedding strength, achieving an effective trade-off between visual quality and robustness. Watermark extraction applies the inverse transform and decryption operations, enabling reliable recovery even under significant distortions. Extensive experiments demonstrate that the proposed method offers superior imperceptibility and robustness relative to state-of-the-art techniques, exhibiting resistance to common signal-processing attacks such as noise, filtering, and blurring. Performance evaluation using different metrics across multiple host-watermark combinations confirms the effectiveness and generalisation capability of the proposed framework.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A machine-learning and GA-optimized encrypted LL\(_3\) watermarking scheme in the DWT domain

  • 期刊论文

  • 发表日期：

    2026年5月29日

查看全文
  • Jannatul Ferdush,
  • Mahbuba Begum,
  • Mohammad Shorif Uddin

摘要

Ensuring the security and robustness of digital image watermarking has become increasingly important due to the growing risks of image tampering, copyright infringement, and unauthorised access. A key challenge in this domain is achieving a reliable balance between imperceptibility and resilience against diverse image-processing attacks. This paper proposes a secure and adaptive watermarking framework that integrates Arnold scrambling with a three-level discrete wavelet transform (3L-DWT). The watermark is first encrypted using an iteration-controlled Arnold map and then decomposed via 3L-DWT, where its LL \(_3\) coefficients are embedded into the highest-entropy block of the host image’s LL \(_3\) subband. A genetic algorithm (GA) is used to optimise the embedding strength, achieving an effective trade-off between visual quality and robustness. Watermark extraction applies the inverse transform and decryption operations, enabling reliable recovery even under significant distortions. Extensive experiments demonstrate that the proposed method offers superior imperceptibility and robustness relative to state-of-the-art techniques, exhibiting resistance to common signal-processing attacks such as noise, filtering, and blurring. Performance evaluation using different metrics across multiple host-watermark combinations confirms the effectiveness and generalisation capability of the proposed framework.