MUMBAI, India, Jan. 2 -- Intellectual Property India has published a patent application (202541122126 A) filed by Vellore Institute Of Technology, Vellore, Tamil Nadu, on Dec. 4, 2025, for 'deepfake audio detection using bi-directional gru and wgan-gp.'
Inventor(s) include Dr. Abdul Gaffar H; Mr. Mudit Sultania; Mr. Sarthak Vishal Mahire; and Ms. Pooja Purushottama Raja.
The application for the patent was published on Jan. 2, under issue no. 01/2026.
According to the abstract released by the Intellectual Property India: "The present disclosure provides a computer-implemented method (100) for detecting deepfake audio. The method includes receiving an audio input signal (102), extracting spectral features from the audio input signal to generate spectral features (108), processing the spectral features using a neural network architecture that includes Bi-Directional Gated Recurrent Unit layers (114) or a Wasserstein Generative Adversarial Network with Gradient Penalty, and generating a classification output (118) indicating whether the audio input signal is authentic or synthetic. The method may include applying noise reduction preprocessing (104) and normalizing amplitude (106) prior to extracting Mel spectrogram features (108). The method may also include applying batch normalization (110) to the spectral features, processing through convolutional layers (112), and applying an attention mechanism (116) to temporal features generated by the Bi-Directional Gated Recurrent Unit layers."
Disclaimer: Curated by HT Syndication.
