Scholarlink

<p>Digital video watermarking is widely used to protect copyright and content authentication. This study proposes a video watermarking method that partitions the frames of the video into non-overlapping <i>Blocks</i> and uses a series of integers as watermarks encoded as 2D representations of their self-inverting permutations which partitions further the <i>Block</i> into <i>Cells</i>. The watermark embedding is performed in the frequency domain and modifies the magnitude of the FFT high frequency coefficients of the corresponding <i>Cells</i>. Subsequently, a single hash value is computed from all the video frames, and its binary representation is embedded into the least significant bits (LSBs) of selected audio samples. The method achieves high imperceptibility (PSNR <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\ge \)</EquationSource> </InlineEquation> 35 dB, SSIM <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\approx \)</EquationSource> </InlineEquation> 1), robust watermark extraction (Re <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(\ge \)</EquationSource> </InlineEquation> 0.90, BER <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(\le \)</EquationSource> </InlineEquation> 0.05 in the majority of attack scenarios), and 100% tamper detection through hash-based verification in all attack scenarios.</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 code-based multiple watermarking scheme for the preservation of authenticity of digital video

  • 期刊论文

  • 发表日期：

    2026年3月30日

查看全文
  • Iosif Polenakis,
  • Vasileios Vouronikos,
  • Konstantinos Balis,
  • Stavros D. Nikolopoulos

摘要

Digital video watermarking is widely used to protect copyright and content authentication. This study proposes a video watermarking method that partitions the frames of the video into non-overlapping Blocks and uses a series of integers as watermarks encoded as 2D representations of their self-inverting permutations which partitions further the Block into Cells. The watermark embedding is performed in the frequency domain and modifies the magnitude of the FFT high frequency coefficients of the corresponding Cells. Subsequently, a single hash value is computed from all the video frames, and its binary representation is embedded into the least significant bits (LSBs) of selected audio samples. The method achieves high imperceptibility (PSNR \(\ge \) 35 dB, SSIM \(\approx \) 1), robust watermark extraction (Re \(\ge \) 0.90, BER \(\le \) 0.05 in the majority of attack scenarios), and 100% tamper detection through hash-based verification in all attack scenarios.