Scholarlink

<p>This paper presents a robust and blind audio watermarking technique utilizing iterative filtering for adaptively embedding encrypted watermark data into perceptually insignificant regions of the audio signal. This work aims to address the growing challenge of intellectual property (IP) infringement and fill the existing research gap by developing a watermarking method that guarantees audio quality alongside strong resistance to illegal attacks. The objective of this research is to design a watermarking scheme that can jointly ensure robustness, imperceptibility, security, and blind extraction, specifically under advanced signal processing, desynchronization, and Stirmark attacks. The core of the proposed scheme is iterative filtering, which decomposes the host audio signal into various intrinsic mode functions (IMFs) and a low-frequency residue component. The final residue, being crucial for maintaining perceptual quality, is selected for watermark embedding. The residue is segmented into 8 × 8 non-overlapping blocks, and an adaptive threshold is computed for each block based on its statistical features. Watermark embedding is carried out by modifying specific samples using a novel adaptive strategy, wherein the selected sample is adjusted to a quarter fraction of the product of the local threshold and an embedding strength factor, depending on the watermark bit value. Chaotic and DNA encryption is employed to guarantee the security of watermark. Experimental results confirm high imperceptibility (SNR: 37–40&#xa0;dB), perfect recovery (NCC = 1), and robust performance with a payload of 204.8 bps.</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

Robust audio watermarking using adaptive embedding in intrinsic modes

  • 期刊论文

  • 发表日期：

    2025年6月13日

查看全文
  • Subreena Mushtaq,
  • Samrah Mehraj,
  • Shabir A. Parah

摘要

This paper presents a robust and blind audio watermarking technique utilizing iterative filtering for adaptively embedding encrypted watermark data into perceptually insignificant regions of the audio signal. This work aims to address the growing challenge of intellectual property (IP) infringement and fill the existing research gap by developing a watermarking method that guarantees audio quality alongside strong resistance to illegal attacks. The objective of this research is to design a watermarking scheme that can jointly ensure robustness, imperceptibility, security, and blind extraction, specifically under advanced signal processing, desynchronization, and Stirmark attacks. The core of the proposed scheme is iterative filtering, which decomposes the host audio signal into various intrinsic mode functions (IMFs) and a low-frequency residue component. The final residue, being crucial for maintaining perceptual quality, is selected for watermark embedding. The residue is segmented into 8 × 8 non-overlapping blocks, and an adaptive threshold is computed for each block based on its statistical features. Watermark embedding is carried out by modifying specific samples using a novel adaptive strategy, wherein the selected sample is adjusted to a quarter fraction of the product of the local threshold and an embedding strength factor, depending on the watermark bit value. Chaotic and DNA encryption is employed to guarantee the security of watermark. Experimental results confirm high imperceptibility (SNR: 37–40 dB), perfect recovery (NCC = 1), and robust performance with a payload of 204.8 bps.