Scholarlink

Image splicing refers to the process of manipulating digital images by copying, pasting, or superimposing parts of different images to create a composite image. The resulting image is intended to trick the viewer into believing it is a real photo when in fact it is a manipulated image. Image forgery is often used for malicious purposes such as: Fabricating fake news, slandering people, or manipulating evidence in court. There are several methods for detecting fakes in image splicing, including visual inspection, statistical analysis, and Deep learning-based approaches. Visual inspection involves manually examining an image to identify discrepancies in lighting, color, texture, and geometry in different parts of the image. Feature Extraction is done from the Casia dataset. The proposed method described below gives an accuracy of 96.59% with an epoch of 30.

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

Image Splicing Detection: A Deep Learning Based Approach

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Bhavana Kaushik,
  • Smita Jha,
  • Vishwa Pratap,
  • K. Vinay,
  • Vinnel Ettam

摘要

Image splicing refers to the process of manipulating digital images by copying, pasting, or superimposing parts of different images to create a composite image. The resulting image is intended to trick the viewer into believing it is a real photo when in fact it is a manipulated image. Image forgery is often used for malicious purposes such as: Fabricating fake news, slandering people, or manipulating evidence in court. There are several methods for detecting fakes in image splicing, including visual inspection, statistical analysis, and Deep learning-based approaches. Visual inspection involves manually examining an image to identify discrepancies in lighting, color, texture, and geometry in different parts of the image. Feature Extraction is done from the Casia dataset. The proposed method described below gives an accuracy of 96.59% with an epoch of 30.