Type of I.P :
Patent A government authority or licence conferring a right or title for a set period, especially the sole right to exclude others from making, using, or selling an invention.
Copyright The exclusive and assignable legal right, given to the originator for a fixed number of years, to print, publish, perform, film, or record literary, artistic, or musical material.
Trademark A symbol, word, or words legally registered or established by use as representing a company or product.
Others Trade Secerts, Innovative Step, Creative etc
Patent Title : A Method for Speech Watermarking in Speaker Verification
Patent Registration No : To be provided
R&D The invention is at the lab stage or proof concept stage
Pre-Commercial The invention is currently in the process of field testing
Ready For Market Invention is ready to be licensed out or sold
Selling Method :
Exclusive Licensing No person or company other than the named licensee can exploit the relevant intellectual property rights. Importantly, the licensor is also excluded from exploiting the intellectual property rights.
Licensing Authorization to use a particular patent. Licensee may not be the only one licensed to exploit the intellectual property.
Assignment The sale of ownership to the individual assigning the intellectual property.
Product Year :
The present invention relates to a method for speech watermarking in speaker verification, comprising the steps of: embedding watermark data into speech signal at a transmitter; and extracting watermark data from the speech signal at a receiver; characterised by the steps of: selecting frames having least speaker-specific information from the speech signal to carry watermark data; detecting voice activity to detect presence or absence of speaker’s voice in the speech signal; and embedding watermark data into the selected frames of the speech signal according to the presence or absence of the speaker’s voice.
- Energy efficient
- save time
Speakers, online speaker recognition
Speaker recognition system uses in telephone banking and voice mail systems. However, many vulnerability slots are available in remote or online speaker recognition system. One of these vulnerability slots is voice conversion which is known as spoofing attack. In spoofing attack, one speaker’ speech which is produced in source, is modified like target speaker’ speech. Two popular spoofing attack methods include speech synthesis system and human voice mimicking. In speech synthesis system, a mapping function between source and target is trained (GMM-based conversion) or target speaker's speech is directly applied for synthesizer (unit selection method). However in human voice mimicking, a person tries to generate speech like target speaker or target's speech is recorded and then played back (play-back attacks). Although studies have shown that human can easily distinguished between synthesized and natural speech, it is difficult even for human to distinguish play-back attacks. Furthermore, state-of-the-art on spoofing attack in speaker recognition have applied phase to detect differentiate natural from synthetic speech. However, using cosine normalized phase, modified group delay phase and relative phase shift are not sufficient in all condition, spoofing attacks and environment. For example, low-quality speech in telephone channel, channel variability and limited simulated spoofing attacks scenario are the main problems in anti-spoofing studied.
Bid your price Here
Starting Price : USD
Login to Bid