Texas Instruments Corporation

Lorin Paul Netsch

Jacek Piotr Stachurski

George R Doddington

David Patrick Magee

Tarkesh Pande

Murtaza Ali

Dan Wang

Barbara J Wheatley

Periagaram K Rajasekaran

Jayant M Naik

Gregory Jan Malysa

Vishu R Viswanathan

Charles T Hemphill

Kaisheng N Yao

Alexis P Bernard

Levent M Arslan

Yeshwant K Muthusamy

Christopher M Kribs

A method of encoding samples in a digital signal is provided that includes receiving a frame of N samples of the digital signal, determining L possible distinct data values in the N samples, determining a reference data value in the L possible distinct data values and a coding order of L−1 remaining possible distinct data values, wherein each of the L−1 remaining possible distinct data values is mapped to a position in the coding order, decomposing the N samples into L−1 coding vectors based on the coding order, wherein each coding vector identifies the locations of one of the L−1 remaining possible distinct data values in the N samples, and encoding the L−1 coding vectors.

Method and system for lossless value-location encoding

A method for sound source localization in a digital system having at least two audio capture devices is provided that includes receiving audio signals from the two audio capture devices, computing a signal-to-noise ratio (SNR) for each frequency band of a plurality of frequency bands in a processing frame of the audio signals, determining a frequency band weight for each frequency band of the plurality of frequency bands based on the SNR computed for the frequency band, computing an estimated time delay of arrival (TDOA) of sound for the processing frame using the frequency band weights, and converting the estimated TDOA to an angle representing sound direction.

Robust estimation of sound source localization

A method of encoding samples in a digital signal is provided that includes receiving a frame of N samples of the digital signal, determining L possible distinct data values in the N samples, determining a reference data value in the L possible distinct data values and a coding order of L&#x2212;1 remaining possible distinct data values, wherein each of the L&#x2212;1 remaining possible distinct data values is mapped to a position in the coding order, decomposing the N samples into L&#x2212;1 coding vectors based on the coding order, wherein each coding vector identifies the locations of one of the L&#x2212;1 remaining possible distinct data values in the N samples, and encoding the L&#x2212;1 coding vectors.

A system includes a frequency modulated continuous wave (FMCW) transceiver, a processor, and a memory. The memory stores program instructions that, when executed by the processor, cause the system to receive a signal representative of an FMCW signal reflected from an object of interest, apply a first Fourier transform to the signal to obtain range data, identify a subset of the range data corresponding to a region of interest, apply a second Fourier transform on the identified subset of the range data to obtain velocity data corresponding to the identified subset of the range data, and identify a gesture performed by the object of interest based on at least a portion of velocity data.

Hidden markov model-based gesture recognition with FMCW radar

Systems and methods are provided for an automated speech recognition system. A microphone records a keyword spoken by a user, and a front end divides the recorded keyword into a plurality of subunits, each containing a segment of recorded audio, and extracts a set of features from each of the plurality of subunits. A decoder assigns one of a plurality of content classes to each of the plurality of subunits according to at least the extracted set of features for each subunit. A quality evaluation component calculates a score representing a quality of the keyword from the content classes assigned to the plurality of subunits.

Quality feedback on user-recorded keywords for automatic speech recognition systems

An example method includes: receiving a test phrase; comparing feature vectors of the test phrase to contents of a first database to generate a first score; comparing the feature vectors of the test phrase to contents of a second database to generate a second score; comparing feature vectors of the contents of the second database to the contents of the first database to generate a third score; comparing the feature vectors of the contents of the second database to a model of the test phrase to generate a fourth score; determining a first difference score based on a difference between the first and second scores; determining a second difference score based on a difference between the third and fourth scores; and generating a difference confidence score based on a lesser of the first and second difference scores.

Multi-phrase difference confidence scoring

A gesture recognition system is shown using a 77 GHz FMCW radar system. The signature of a gesturing hand is measured to construct an energy distribution in velocity space over time. A gesturing hand is fundamentally a dynamical system with unobservable 'state' (i.e. the type of the gesture) which determines the sequence of associated observable velocity-energy distributions, therefore a Hidden Markov Model is used to for gesture recognition. A method for reducing the length of the feature vectors by a factor of 12 is also shown, by re-parameterizing the feature vectors in terms of a sum of Gaussians without decreasing the recognition performance.

Hidden Markov model-based gesture recognition with FMCW radar

A method of encoding samples in a digital signal is provided that includes receiving a plurality of samples of the digital signal, and encoding the plurality of samples, wherein an output number of bits is adapted for coding efficiency when a value in a range of possible distinct data values of the plurality of samples is not found in the plurality of samples.

Adaptive encoding of a digital signal with one or more missing values

Automatic speech recognition verification using a combination of two or more confidence scores based on UV features which reuse computations of the original recognition.

Multiplying confidence scores for utterance verification in a mobile telephone

A method for training acoustic models for a new target language is provided using a phonetic table, which characterizes the phones, used in one or more reference language(s) with respect to their articulatory properties; a phonetic table, which characterizes the phones used in the target language with respect to their articulatory properties; a set of trained monophones for the reference language(s); and a database of sentences in the target language and its phonetic transcription. With these inputs, the new method completely and automatically takes care of the steps of monophone seeding and triphone clustering and machine intensive training steps involved in triphone acoustic training.

Growing community of inventors

Allen, TX, United States of America

Lorin Paul Netsch