Abstract: Systems and methods for user authentication include a microphone generating a corresponding audio input signal, and a voice processing system operable to identify a pass phrase including human speech in the audio input signal, a voice authentication system operable to extract feature characteristics associated with the identified pass phrase, compare the extracted feature characteristics to user voice features, and determine a confidence level of a voice match. The voice authentication system receives a first acoustic code, compares the first acoustic code to a second acoustic code associated with the user, and determines whether the first acoustic code and second acoustic code match. A user is authenticated if the first acoustic code and second acoustic code match and the confidence level exceeds a threshold value. A user device is operable to prompt the user to speak the pass phrase, generate the acoustic code and output the acoustic code through a speaker.

Abstract: Systems and methods for user authentication include a microphone generating a corresponding audio input signal, and a voice processing system operable to identify a pass phrase including human speech in the audio input signal, a voice authentication system operable to extract feature characteristics associated with the identified pass phrase, compare the extracted feature characteristics to user voice features, and determine a confidence level of a voice match. The voice authentication system receives a first acoustic code, compares the first acoustic code to a second acoustic code associated with the user, and determines whether the first acoustic code and second acoustic code match. A user is authenticated if the first acoustic code and second acoustic code match and the confidence level exceeds a threshold value. A user device is operable to prompt the user to speak the pass phrase, generate the acoustic code and output the acoustic code through a speaker.

Abstract: Systems and methods for multi-sourced noise suppression are provided. An example system may receive streams of audio data including a voice signal and noise, the voice signal including a spoken word. The streams of audio data are provided by distributed audio devices. The system can assign weights to the audio streams based at least partially on quality of the audio streams. The weights of audio streams can be determined based on signal-to-noise ratios (SNRs). The system may further process, based on the weights, the audio stream to generate cleaned speech. Each audio device comprises microphone(s) and can be associated with the Internet of Things (IoT), such that the audio devices are Internet of Things devices. The processing can include noise suppression and reduction and echo cancellation. The cleaned speech can be provided to a remote device for further processing which may include Automatic Speech Recognition (ASR).

Abstract: Method and systems for modeling user possession of a mobile device for a user authentication framework are provided. The method includes analyzing sensor data representing information captured from sensor(s) associated with at least one of a plurality of devices, the plurality including the user's mobile device. The method allows for determining, based on the analyzed sensor data, a probability that the user maintains possession of the user's mobile device; and configuring the probability for use in determining whether to require authentication. The determination may include configuring a probabilistic model that includes a first state indicating that the user possesses the mobile device and a second state indicating that the user does not; classifying motions of the mobile device by types, the motions being determined based on the sensor data; and updating probabilities of the two states in response to determining that at least one of the motions has occurred.

Abstract: Systems and methods for multi-sourced noise suppression are provided. An example system may receive streams of audio data including a voice signal and noise, the voice signal including a spoken word. The streams of audio data are provided by distributed audio devices. The system can assign weights to the audio streams based at least partially on quality of the audio streams. The weights of audio streams can be determined based on signal-to-noise ratios (SNRs). The system may further process, based on the weights, the audio stream to generate cleaned speech. Each audio device comprises microphone(s) and can be associated with the Internet of Things (IoT), such that the audio devices are Internet of Things devices. The processing can include noise suppression and reduction and echo cancellation. The cleaned speech can be provided to a remote device for further processing which may include Automatic Speech Recognition (ASR).