Scholarlink

<p>Visual cryptography often faces a trade-off between reconstruction clarity and encryption strength, particularly under noise and geometric distortions. This work presents Deep Edge-VCS, a dual-domain visual cryptography framework designed to achieve secure and high-fidelity image sharing. The approach leverages edge-enhanced preprocessing and a transformer-based encoding scheme combined with a masked diffusion encryption module for ensuring robust resistance to cryptanalytic attacks. Two non-informative visual shares are generated and they individually reveal nothing but reconstruct the secret accurately when combined. Experiments demonstrate superior reconstruction quality with a PSNR of 66.82 dB and SSIM of 0.9999, along with strong encryption properties indicated by UACI of 33.47% and NPCR of 99.65%. These results confirm that Deep Edge-VCS effectively balances visual quality, security, and computational efficiency, making it well-suited for secure image-sharing applications.</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

Deep Edge-VCS: Dual-Domain Transformer-Based Framework for Secure Visual Cryptography

  • 期刊论文

  • 发表日期：

    2025年10月22日

查看全文
  • N. Sugirtham,
  • R. Sudhakar

摘要

Visual cryptography often faces a trade-off between reconstruction clarity and encryption strength, particularly under noise and geometric distortions. This work presents Deep Edge-VCS, a dual-domain visual cryptography framework designed to achieve secure and high-fidelity image sharing. The approach leverages edge-enhanced preprocessing and a transformer-based encoding scheme combined with a masked diffusion encryption module for ensuring robust resistance to cryptanalytic attacks. Two non-informative visual shares are generated and they individually reveal nothing but reconstruct the secret accurately when combined. Experiments demonstrate superior reconstruction quality with a PSNR of 66.82 dB and SSIM of 0.9999, along with strong encryption properties indicated by UACI of 33.47% and NPCR of 99.65%. These results confirm that Deep Edge-VCS effectively balances visual quality, security, and computational efficiency, making it well-suited for secure image-sharing applications.