Scholarlink

Digital forensics is the process of collecting, analyzing, and interpreting computer evidence in order to detect cyber crimes. One way to subvert spoofed propaganda is through the use of digital forensics, which can identify and authenticate where synthetic material originally came from. You see with deepfake content (deceptively real-looking but synthetic video and picture derived from A.I. bots) on the one hand there are advantages to it and on the other disadvantages of social media platforms. The video features that look like real ones are created by a class of methods called deep learning. If successful, the research could be used to identify deepfake content and limit its dissemination to prevent the spread of misinformation. The article describes a way of identifying deepfakes using Long Short Term Memory (LSTM) for detecting deepfake and Modified XceptionNet architecture for recognizing deepfake videos and photos. For training and testing the constructed model, a different benchmark dataset of video and image will be used. In order to accurately assess the performance, transparency and interpretability of the model, it is planned to test with dataset.

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

Identification of Synthetic Streaming Based Media Applying Deep Learning Models in Digital Forensics

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • S. S. Nagamuthu Krishnan,
  • Karthika Selvaraj,
  • Varsha Sivasubramani

摘要

Digital forensics is the process of collecting, analyzing, and interpreting computer evidence in order to detect cyber crimes. One way to subvert spoofed propaganda is through the use of digital forensics, which can identify and authenticate where synthetic material originally came from. You see with deepfake content (deceptively real-looking but synthetic video and picture derived from A.I. bots) on the one hand there are advantages to it and on the other disadvantages of social media platforms. The video features that look like real ones are created by a class of methods called deep learning. If successful, the research could be used to identify deepfake content and limit its dissemination to prevent the spread of misinformation. The article describes a way of identifying deepfakes using Long Short Term Memory (LSTM) for detecting deepfake and Modified XceptionNet architecture for recognizing deepfake videos and photos. For training and testing the constructed model, a different benchmark dataset of video and image will be used. In order to accurately assess the performance, transparency and interpretability of the model, it is planned to test with dataset.