This inventor holds 1 USPTO granted patent. Top assignee: The City University of New York. Active years: 2024.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile
Company Filing History:
Years Active: 2024
Title: Soumi Maiti: Inventor in Speech Signal Processing.
Introduction
Soumi Maiti is an inventor based in Brooklyn, New York. His work includes a patented method for extracting speech from degraded audio signals.
Latest Patents
Maiti is associated with 1 patent: “Method for Extracting Speech from Degraded Signals by Predicting the Inputs to a Speech Vocoder.” The invention uses a prediction model to estimate parameters for an audible signal from degraded audio containing a distorted target signal. Trained to minimize the difference between the target and predicted audio, the model supplies its predicted parameters to a waveform generator, which synthesizes the audible signal through parametric resynthesis.
Career Highlights
Maiti is associated with The City University of New York. His patent focuses on reconstructing speech from degraded signals by predicting the inputs needed by a speech vocoder.
Collaborations
Michael Mandel is identified as Maiti’s coworker.
Conclusion
Soumi Maiti’s patent describes a speech-processing approach that predicts vocoder parameters to synthesize audible speech from degraded audio.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile