Scholarlink

Digital copyright protection for audio requires effective watermarking techniques to ensure content authenticity and copyright security. A deep learning-based audio watermarking system provides high signal fidelity, imperceptibility, and resistance to attacks. It combines (CNN) and (LSTM) networks in an encoding-decoding framework to embed watermarks without degrading quality. The process inserts an imperceptible watermark into the audio, which remains retrievable even after compression, noise, or resampling. Experiments show that this method outperforms existing ones in terms of imperceptibility, Signal-to-Noise Ratio (SNR), and robustness. It offers a secure, flexible solution for audio copyright protection, expanding digital media security opportunities.

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

High-Fidelity Audio Watermarking: An End-to-End Deep Learning Approach for an Imperceptible and Robust Embedding

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Vani Hiremani,
  • Spandan Chavan,
  • Divyansh Kumar,
  • Atharva Gondhali,
  • Trisha Boda,
  • Sashikala Mishra

摘要

Digital copyright protection for audio requires effective watermarking techniques to ensure content authenticity and copyright security. A deep learning-based audio watermarking system provides high signal fidelity, imperceptibility, and resistance to attacks. It combines (CNN) and (LSTM) networks in an encoding-decoding framework to embed watermarks without degrading quality. The process inserts an imperceptible watermark into the audio, which remains retrievable even after compression, noise, or resampling. Experiments show that this method outperforms existing ones in terms of imperceptibility, Signal-to-Noise Ratio (SNR), and robustness. It offers a secure, flexible solution for audio copyright protection, expanding digital media security opportunities.