Back to results
US 2018/0197535 A1Published application
Systems and Methods for Human Speech Training
AssigneeBoard of Regents, The University of Texas System
InventorsMd. Motasim BELLAH, Jodi M. TOMMERDAHI, Mohammad Raziul HASAN, Samir M. IQBAL
Life of this patent
Estimating the term…
Abstract
Language learning and speech training techniques are implemented to provide automated and real-time quantitative feedback to the user. Sound samples produced by a speaker are transcribed and analyzed against database sound samples to: provide speech therapy, compute the accuracy of a speaker's pronunciation, aid in learning a foreign language, and help members of the deaf community learn to communicate with spoken language.
Claims20
A method for human speech training, comprising:
- receiving a sound sample from a human speaker;
- comparing the received sound sample against sound samples in a database by performing a mathematical analysis to determine a percent match between the received sound sample and one of the sound samples in the database;
- using the percent match as quantitative feedback to the human speaker in real time;
- and
- notifying the human speaker that higher percent match indicates higher pronunciation accuracy.
2 more elements
A system for speech training, comprising a processor programmed to:
- receive a sound sample from a human speaker;
- compare the received sound sample against sound samples in a database by performing a mathematical analysis to determine a percent match between the received sound sample and one of the sound samples in the database;
- use the percent match as quantitative feedback to the human speaker in real time;
- and
- notify the human speaker that higher percent match indicates higher pronunciation accuracy.
2 more elements
Dependencies are inferred from the claim text.