Scholarlink

<p>Technological breakthroughs and the tendency toward digital multimedia content entail a productive framework for securing ownership and copyright protection. These are currently attainable through a multitude of proposed image watermarking approaches, featuring variable efficacy in terms of robustness and imperceptibility. This research proposes a dual-step technique encompassing: (1) discrete wavelet transform analysis of the host image, and (2) low-low subband-based singular value decomposition (SVD) of the host image to 4 × 4 blocks. Such a process fosters embedding watermark bits into selected coefficients seen in the SVD-emanated V and U elements. The proposed technique works with human visual system metrics to precisely select blocks tailored to embedding. It further employs an optimization algorithm to identify optimal scaling elements for embedding. This algorithm picks parameter values that (1) realizes optimum robustness, and (2) sustains imperceptibility at the desired level. By comparison, teaching–learning-based optimization (TLBO) outperforms artificial bee colony (ABC) in improving the robustness of the proposed technique. As with the experimental results, the proposed technique can effectively produce watermarked images at various levels of peak signal-to-noise ratio (PSNR) without any extraction error. Compared to other groundbreaking techniques, the proposed technique possesses a comparable imperceptibility and is noticeably robust against most attacks. Likewise, it advantageously mitigates the chance of false positive problem occurrence.</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

An Enhanced DWT-SVD-Based Robust Image Watermarking Scheme Optimized for Imperceptibility and Robustness

  • 期刊论文

  • 发表日期：

    2025年1月7日

查看全文
  • Vajiheh Sabeti

摘要

Technological breakthroughs and the tendency toward digital multimedia content entail a productive framework for securing ownership and copyright protection. These are currently attainable through a multitude of proposed image watermarking approaches, featuring variable efficacy in terms of robustness and imperceptibility. This research proposes a dual-step technique encompassing: (1) discrete wavelet transform analysis of the host image, and (2) low-low subband-based singular value decomposition (SVD) of the host image to 4 × 4 blocks. Such a process fosters embedding watermark bits into selected coefficients seen in the SVD-emanated V and U elements. The proposed technique works with human visual system metrics to precisely select blocks tailored to embedding. It further employs an optimization algorithm to identify optimal scaling elements for embedding. This algorithm picks parameter values that (1) realizes optimum robustness, and (2) sustains imperceptibility at the desired level. By comparison, teaching–learning-based optimization (TLBO) outperforms artificial bee colony (ABC) in improving the robustness of the proposed technique. As with the experimental results, the proposed technique can effectively produce watermarked images at various levels of peak signal-to-noise ratio (PSNR) without any extraction error. Compared to other groundbreaking techniques, the proposed technique possesses a comparable imperceptibility and is noticeably robust against most attacks. Likewise, it advantageously mitigates the chance of false positive problem occurrence.