Scholarlink

Deepfake audio using shallow convolutional neural networks (CNNs), a crucial development given the increasing prevalence of AI-generated fake audio content. By focusing on the nuances in the frequency domain, temporal dynamics, and unique artifacts of synthetic audio, the proposed shallow CNN model is trained on a comprehensive mix of authentic and manipulated audio samples. Preliminary results affirm the model’s capability to accurately discern between real and fake audios across a variety of scenarios, underscoring its potential as a scalable and efficient tool in maintaining audio authenticity and combating misinformation in the digital age.

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

Deepfake Audio Detection Using Shallow Convolutional Neural Network

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Prema Sahane,
  • Durgesh Badole,
  • Suhas Chavare,
  • Chaitanya Kale,
  • Siddhesh Walunj

摘要

Deepfake audio using shallow convolutional neural networks (CNNs), a crucial development given the increasing prevalence of AI-generated fake audio content. By focusing on the nuances in the frequency domain, temporal dynamics, and unique artifacts of synthetic audio, the proposed shallow CNN model is trained on a comprehensive mix of authentic and manipulated audio samples. Preliminary results affirm the model’s capability to accurately discern between real and fake audios across a variety of scenarios, underscoring its potential as a scalable and efficient tool in maintaining audio authenticity and combating misinformation in the digital age.