top of page
추상적 인 구체

DeepVoice Detection

DeepVoice detection involves using deep learning to identify synthetic voices, distinguishing them from human ones.  Our system leverages deep neural networks to analyze unique audio traits such as tonal fluctuations and speech patterns that are challenging for AI to mimic accurately.

DeepVoice Detection System

Web demo.png

Web-based DeepVoice Detection

This web-based DeepVoice detection system employs advanced deep learning techniques to distinguish between human and synthetic voices, offering immediate feedback through a user-friendly drag-and-drop interface.

  • Supervised Contrastive Learning (SCL): Enhances detection accuracy by contrasting real and synthetic voice features.

  • Breathing-Talking-Silence Encoding (BTS-E): Focuses on analyzing breathing patterns and speech cadence to detect anomalies.

  • User-Friendly Interface: Enables easy audio file upload and real-time graphical results display.

Demo Video

Demo Video

Watch Now

To request a demo

46, Sadang-ro, Dongjak-gu

Seoul, Korea 


Thanks for submitting!

bottom of page