Scholarlink

<p>In the realm of secure digital communication, concealing data within images has emerged as an effective strategy to protect information from unauthorized access. Despite significant progress, many existing image steganography techniques either compromise image quality or support limited data capacity, especially when larger payloads are involved. While edge-based methods offer improved visual fidelity by embedding data in visually complex regions, they often depend on conventional edge detection algorithms that struggle to accurately identify intricate edge details and maximize embedding opportunities. This study introduces an improved steganographic approach that combines Laplacian of Gaussian (LoG) edge detection with image dilation to more precisely extract and expand edge regions within the cover image. These enhanced edge areas serve as optimal locations for embedding secret data as the inherent complex visual properties of these areas makes alterations less perceptible. The concealment process embeds the secret image within both edge and non-edge regions of the cover image using a defined X: Y distribution ratio, where larger payload is allocated to edge pixels (Y) than to non-edge pixels (X), thereby balancing capacity and imperceptibility. The complete process involves three phases: generating the edge map through pre-processing, embedding secret data using the adaptive ratio, and retrieving the hidden image in the extraction stage. Simulation results show that the proposed method achieves payloads up to 3.44 bits per pixel (bpp) while maintaining high visual quality, with PSNR exceeding 35 dB and SSIM values above 0.96, demonstrating an effective balance between capacity and imperceptibility.</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

Variable payload-based image steganographic scheme using dilated Laplacian of gaussians edge detection

  • 期刊论文

  • 发表日期：

    2025年7月18日

查看全文
  • Biswajit Patwari,
  • Sudipta Kr Ghosal,
  • Sayani Dhar,
  • Utpal Nandi

摘要

In the realm of secure digital communication, concealing data within images has emerged as an effective strategy to protect information from unauthorized access. Despite significant progress, many existing image steganography techniques either compromise image quality or support limited data capacity, especially when larger payloads are involved. While edge-based methods offer improved visual fidelity by embedding data in visually complex regions, they often depend on conventional edge detection algorithms that struggle to accurately identify intricate edge details and maximize embedding opportunities. This study introduces an improved steganographic approach that combines Laplacian of Gaussian (LoG) edge detection with image dilation to more precisely extract and expand edge regions within the cover image. These enhanced edge areas serve as optimal locations for embedding secret data as the inherent complex visual properties of these areas makes alterations less perceptible. The concealment process embeds the secret image within both edge and non-edge regions of the cover image using a defined X: Y distribution ratio, where larger payload is allocated to edge pixels (Y) than to non-edge pixels (X), thereby balancing capacity and imperceptibility. The complete process involves three phases: generating the edge map through pre-processing, embedding secret data using the adaptive ratio, and retrieving the hidden image in the extraction stage. Simulation results show that the proposed method achieves payloads up to 3.44 bits per pixel (bpp) while maintaining high visual quality, with PSNR exceeding 35 dB and SSIM values above 0.96, demonstrating an effective balance between capacity and imperceptibility.