Scholarlink

<p>The Internet of Things (IoT) has immense importance in the development of numerous smart systems across key sectors like agriculture, healthcare, homes, and education, aimed at boosting efficiency and enhancing our daily lives. Although certain IoT systems handle sensitive data, requiring additional measures to ensure traceability and authenticity. In a surveillance system, the receiver must verify that the data originates from the right sender and not from an unauthorized third party. This paper presents a secure watermarking framework for surveillance video frames through a novel smart data-hiding approach. The framework includes two core modules, the embedder and the extractor, which communicate over an untrusted channel that is vulnerable to attacks. The embedding process uses Modified Independent Component Analysis (M-ICA) to identify optimal embedding regions. Additionally, we introduce a new quantization method, C-QIM, coupled with tensor-QR decomposition to hide the watermark bits with greater imperceptibility and robustness while requiring less computation. At the receiver side, extraction and verification ensure accurate watermark recovery and authentication. The proposed framework is evaluated against H.264 compression, noise, and filtering attacks. Experimental results demonstrate the highest robustness in identifying optimal embedding areas while maintaining fast execution, authenticity, and image fidelity.</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

Surveillance video frames based on smart data hiding against H.264 compression

  • 期刊论文

  • 发表日期：

    2026年6月2日

查看全文
  • Abdallah Soualmi,
  • Adel Alti,
  • Lamri Laouamer

摘要

The Internet of Things (IoT) has immense importance in the development of numerous smart systems across key sectors like agriculture, healthcare, homes, and education, aimed at boosting efficiency and enhancing our daily lives. Although certain IoT systems handle sensitive data, requiring additional measures to ensure traceability and authenticity. In a surveillance system, the receiver must verify that the data originates from the right sender and not from an unauthorized third party. This paper presents a secure watermarking framework for surveillance video frames through a novel smart data-hiding approach. The framework includes two core modules, the embedder and the extractor, which communicate over an untrusted channel that is vulnerable to attacks. The embedding process uses Modified Independent Component Analysis (M-ICA) to identify optimal embedding regions. Additionally, we introduce a new quantization method, C-QIM, coupled with tensor-QR decomposition to hide the watermark bits with greater imperceptibility and robustness while requiring less computation. At the receiver side, extraction and verification ensure accurate watermark recovery and authentication. The proposed framework is evaluated against H.264 compression, noise, and filtering attacks. Experimental results demonstrate the highest robustness in identifying optimal embedding areas while maintaining fast execution, authenticity, and image fidelity.