Scholarlink

Digital image authentication is critical in forensics to detect forgeries that compromise image integrity, such as copy-move attacks. Traditional methods like SIFT and SURF often struggle with complex or distant pixel patterns. This research proposes an optimized CNN model for detecting various tampering types–including duplication, splicing, and noise inconsistencies–achieving 98% accuracy, 98% precision, and 97% recall. The approach outperforms conventional techniques, offering a robust and reliable solution for digital image authentication.

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

Duplication Fraudulent Detection of Tampered Image Using Optimized CNN

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Debjani Chakraborty,
  • Biswajit Halder,
  • Sourav Saha,
  • Nabanita Das

摘要

Digital image authentication is critical in forensics to detect forgeries that compromise image integrity, such as copy-move attacks. Traditional methods like SIFT and SURF often struggle with complex or distant pixel patterns. This research proposes an optimized CNN model for detecting various tampering types–including duplication, splicing, and noise inconsistencies–achieving 98% accuracy, 98% precision, and 97% recall. The approach outperforms conventional techniques, offering a robust and reliable solution for digital image authentication.