Scholarlink

<p>With the proliferation of digital imaging and readily available editing tools, ensuring the authenticity of digital images has become a critical challenge in many security-sensitive applications. Detecting and localizing these modifications within the image is a crucial issue. In this paper, we present a novel quadruple self-embedding framework for detecting and localizing the tamper. At first, we divide the cover image into 4 major blocks. Further, these four blocks are subdivided into 4 × 4 blocks that are non-overlapping. A watermark is generated from the image itself by choosing one 4 × 4 block from each four major consecutive blocks. We compute Average Arithmetic Intensity (AAI) for each block to facilitate the localization of tamper. This AAI is then converted to binary form to obtain a Bit Vector (BV), that is 8 bits long for each block. The BV of each block is concatenated to obtain an Enhanced Bit Vector (EBV), which is a 32-bit watermark. To ensure the security of this watermark, we encrypt it using Deoxyribonucleic Acid (DNA) encryption. Further, for mapping, we use Arnold mapping, making it an efficient way for localizing the tamper within the image since it does not use any lookup tables. The experimentation reveals that the proposed framework outdoes the existing state-of-the-art techniques (SOTA) by providing higher PSNR and better localization capability. We have evaluated the framework for Kodak, UCID, and Waterloo grayscale image database. Our method obtained an average PSNR of 44.44 dB for a payload of 2 bits per pixel and an average SSIM value of 0.9898. Further, the framework provides better results for False Positive Rates (FPR), False Negative Rates (FNR), and Tamper Detection Rates (TDR), making it a better candidate for authentication of images.</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 self-embedding fragile watermarking framework based on DNA and Arnold transform

  • 期刊论文

  • 发表日期：

    2026年2月12日

查看全文
  • Ahmad Alkhayyat,
  • M. Ambika,
  • Deeksha Verma,
  • Rajendar Sandiri,
  • Gadug Sudhamsu,
  • Jajneswar Nanda,
  • Vikas Wasson,
  • Murari Devakannan Kamalesh,
  • Mohit Kumar,
  • Anita Gehlot

摘要

With the proliferation of digital imaging and readily available editing tools, ensuring the authenticity of digital images has become a critical challenge in many security-sensitive applications. Detecting and localizing these modifications within the image is a crucial issue. In this paper, we present a novel quadruple self-embedding framework for detecting and localizing the tamper. At first, we divide the cover image into 4 major blocks. Further, these four blocks are subdivided into 4 × 4 blocks that are non-overlapping. A watermark is generated from the image itself by choosing one 4 × 4 block from each four major consecutive blocks. We compute Average Arithmetic Intensity (AAI) for each block to facilitate the localization of tamper. This AAI is then converted to binary form to obtain a Bit Vector (BV), that is 8 bits long for each block. The BV of each block is concatenated to obtain an Enhanced Bit Vector (EBV), which is a 32-bit watermark. To ensure the security of this watermark, we encrypt it using Deoxyribonucleic Acid (DNA) encryption. Further, for mapping, we use Arnold mapping, making it an efficient way for localizing the tamper within the image since it does not use any lookup tables. The experimentation reveals that the proposed framework outdoes the existing state-of-the-art techniques (SOTA) by providing higher PSNR and better localization capability. We have evaluated the framework for Kodak, UCID, and Waterloo grayscale image database. Our method obtained an average PSNR of 44.44 dB for a payload of 2 bits per pixel and an average SSIM value of 0.9898. Further, the framework provides better results for False Positive Rates (FPR), False Negative Rates (FNR), and Tamper Detection Rates (TDR), making it a better candidate for authentication of images.