Using Deep Learning Techniques and Inferential Speech Statistics for AI Synthesised Speech Recognition

TL;DR

This paper presents a CNN and BiRNN-based model that effectively distinguishes AI-synthesized speech from real speech and identifies the synthesis source, addressing security threats posed by advanced audio generation technologies.

Contribution

The study introduces a novel deep learning model combining CNN and BiRNN for both detection and source identification of AI-synthesized speech, outperforming existing methods.

Findings

Achieved 1.9% error rate in classifying real vs. synthesized speech.

Detected synthesis architecture with 97% accuracy.

Outperformed state-of-the-art approaches in speech classification.

Abstract

The recent developments in technology have re-warded us with amazing audio synthesis models like TACOTRON and WAVENETS. On the other side, it poses greater threats such as speech clones and deep fakes, that may go undetected. To tackle these alarming situations, there is an urgent need to propose models that can help discriminate a synthesized speech from an actual human speech and also identify the source of such a synthesis. Here, we propose a model based on Convolutional Neural Network (CNN) and Bidirectional Recurrent Neural Network (BiRNN) that helps to achieve both the aforementioned objectives. The temporal dependencies present in AI synthesized speech are exploited using Bidirectional RNN and CNN. The model outperforms the state-of-the-art approaches by classifying the AI synthesized audio from real human speech with an error rate of 1.9% and detecting the underlying architecture with an accuracy of 97%.