Patents by Inventor Nelson H. Morgan
Nelson H. Morgan has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Publication number: 20110153326Abstract: A system and method for extracting acoustic features and speech activity on a device and transmitting them in a distributed voice recognition system. The distributed voice recognition system includes a local VR engine in a subscriber unit and a server VR engine on a server. The local VR engine comprises a feature extraction (FE) module that extracts features from a speech signal, and a voice activity detection module (VAD) that detects voice activity within a speech signal. The system includes filters, framing and windowing modules, power spectrum analyzers, a neural network, a nonlinear element, and other components to selectively provide an advanced front end vector including predetermined portions of the voice activity detection indication and extracted features from the subscriber unit to the server. The system also includes a module to generate additional feature vectors on the server from the received features using a feed-forward multilayer perceptron (MLP) and providing the same to the speech server.Type: ApplicationFiled: February 9, 2011Publication date: June 23, 2011Applicant: QUALCOMM INCORPORATEDInventors: HARINATH GARUDADRI, HYNEK HERMANSKY, LUKAS BURGET, PRATIBHA JAIN, SACHIN KAJAREKAR, SUNIL SIVADAS, STEPHANE N. DUPONT, MARIA CARMEN BENITEZ ORTUZAR, NELSON H. MORGAN
-
Patent number: 7089178Abstract: A distributed voice recognition system and method for obtaining acoustic features and speech activity at multiple frequencies by extracting high frequency components thereof on a device, such as a subscriber station and transmitting them to a network server having multiple stream processing capability, including cepstral feature processing, MLP nonlinear transformation processing, and multiband temporal pattern architecture processing. The features received at the network server are processed using all three streams, wherein each of the three streams provide benefits not available in the other two, thereby enhancing feature interpretation. Feature extraction and feature interpretation may operate at multiple frequencies, including but not limited to 8 kHz, 11 kHz, and 16 kHz.Type: GrantFiled: April 30, 2002Date of Patent: August 8, 2006Assignee: Qualcomm Inc.Inventors: Harinath Garudadri, Sunil Sivadas, Hynek Hermansky, Nelson H. Morgan, Charles C. Wooters, Andre Gustavo Adami, Maria Carmen Benitez Ortuzar, Lukas Burget, Stephane N. Dupont, Frantisek Grezl, Pratibha Jain, Sachin Kajarekar, Petr Motlicek
-
Publication number: 20030204394Abstract: A distributed voice recognition system and method for obtaining acoustic features and speech activity at multiple frequencies by extracting high frequency components thereof on a device, such as a subscriber station and transmitting them to a network server having multiple stream processing capability, including cepstral feature processing, MLP nonlinear transformation processing, and multiband temporal pattern architecture processing. The features received at the network server are processed using all three streams, wherein each of the three streams provide benefits not available in the other two, thereby enhancing feature interpretation. Feature extraction and feature interpretation may operate at multiple frequencies, including but not limited to 8 kHz, 11 kHz, and 16 kHz.Type: ApplicationFiled: April 30, 2002Publication date: October 30, 2003Inventors: Harinath Garudadri, Sunil Sivadas, Hynek Hermansky, Nelson H. Morgan, Charles C. Wooters, Andre Gustavo Adami, Maria Carmen Benitez Ortuzar, Lukas Burget, Stephane N. Dupont, Frantisek Grezl, Pratibha Jain, Sachin Kajarekar, Petr Motlicek
-
Publication number: 20030004720Abstract: A system and method for extracting acoustic features and speech activity on a device and transmitting them in a distributed voice recognition system. The distributed voice recognition system includes a local VR engine in a subscriber unit and a server VR engine on a server . The local VR engine comprises a feature extraction (FE) module that extracts features from a speech signal, and a voice activity detection module (VAD) that detects voice activity within a speech signal. The system includes filters, framing and windowing modules, power spectrum analyzers, a neural network, a nonlinear element, and other components to selectively provide an advanced front end vector including predetermined portions of the voice activity detection indication and extracted features from the subscriber unit to the server .Type: ApplicationFiled: January 28, 2002Publication date: January 2, 2003Inventors: Harinath Garudadri, Hynek Hermansky, Lukas Burget, Pratibha Jain, Sachin Kajarekar, Sunil Sivadas, Stephane N. Dupont, Maria Carmen Benitez Ortuzar, Nelson H. Morgan
-
Patent number: 6308155Abstract: An automatic speech recognition apparatus and method with a front end feature extractor that improves recognition performance under adverse acoustic conditions are disclosed. The inventive feature extractor is characterized by a critical bandwidth spectral resolution, an emphasis on slow changes in the spectral structure of the speech signal, and adaptive automatic gain control. In one the feature extractor includes a feature generator configured to compute short-term parameters of the speech signal, a filter system configured to filter the time sequences of the short-term parameters, and a normalizer configured to normalize the filtered parameters with respect to one or more previous values of the filtered parameters.Type: GrantFiled: May 25, 1999Date of Patent: October 23, 2001Assignee: International Computer Science InstituteInventors: Brian E. D. Kingsbury, Steven Greenberg, Nelson H. Morgan
-
Patent number: 5537647Abstract: A method and system are provided for alleviating the harmful effects of convolutional and additive noise in speech, such as due to the environmental noise and linear spectral modification, based on the filtering of time trajectories of an auditory-like spectrum in a particular spectral domain.Type: GrantFiled: November 5, 1992Date of Patent: July 16, 1996Assignees: U S West Advanced Technologies, Inc., International Computer Science InstituteInventors: Hynek Hermansky, Nelson H. Morgan
-
Patent number: 5450522Abstract: A method and system are provided for alleviating the harmful effects of convolutional distortions of speech, such as the effect of a telecommunication channel, on the performance of an automatic speech recognizer (ASR). The technique is based on the filtering of time trajectories of an auditory-like spectrum derived from the Perceptual Linear Predictive (PLP) method of speech parameter estimation.Type: GrantFiled: August 19, 1991Date of Patent: September 12, 1995Assignees: U S West Advanced Technologies, Inc., International Computer Science InstituteInventors: Hynek Hermansky, Nelson H. Morgan, Philip D. Kohn
-
Patent number: 5163133Abstract: A computer system utilizes a parallel data broadcast architecture, called "PDB" is adapted to be connected to a host computer system. In one digital embodiment, the PDB system includes an interface and global memory which is connected to the host computer system and to three common buses. The buses, in parallel, are, respectively, a data broadcast bus, an instruction bus and an output bus, each bus being connected to a plurality of computational units. Each computational unit includes an arithmetic processor, which preferably, in this embodiment, is a digital signal processor (a special-purpose integrated circuit microprocessor), a local memory and a local program memory. The computational units receive the input data simultaneously and in parallel; transform the data, depending on the parameters received from the instruction bus; and communicate their results, by multiplexing, to the output bus.Type: GrantFiled: February 17, 1989Date of Patent: November 10, 1992Assignee: SAM Technology, Inc.Inventors: Nelson H. Morgan, Alan S. Gevins
-
Patent number: 5020538Abstract: A magnetoencaphalogram (MEG) detects the neuromagnetic brain signals of a subject by bringing a set of magnetic sensors, preferably SQUIDs (superconducting quantum interference device), close to the scalp of the subject. The adverse effect of environmental magnetic noise is reduced by using a second set of magnetic sensors which are placed near the subject. The system includes a bandpass filter, to divide the signals from the various sensors into predetermined frequency bands, and a computer to provide a least squares noise component estimate of the noise from each brain wave sensor and to calculate therefrom a filtered brain wave signal having reduced noise.Type: GrantFiled: November 14, 1989Date of Patent: June 4, 1991Assignee: SAM Technology, Inc.Inventors: Nelson H. Morgan, Alan S. Gevins
-
Patent number: 4736751Abstract: A method and system is provided for the analysis of a human subject's brain wave activity on a statistical basis using a digital computer. The location of portions of the subject's brain and the location of at least 32, and for example 265, external scalp sensors are digitally recorded and stored in computer memory. The subject receives a set of stimuli which evoke brain waves (evoked potential EP or evoked magnetic fields EF) which, along with the location data, are statistically analyzed to indicate the brain sites giving rise to the activity. The brain site activity, and the time interrelationships of brain waves as they progress between brain sites, are displayed on a three-dimensional model or other three-dimensional perspective display.Type: GrantFiled: December 16, 1986Date of Patent: April 12, 1988Assignee: EEG Systems LaboratoryInventors: Alan S. Gevins, Nelson H. Morgan, Douglas S. Greer
-
Patent number: 4338581Abstract: Room acoustics simulation is achieved by digitally defining delay and weighting parameters and generating sampled analog signal responses. In the early portion of a response term an accurate impulse response is provided and subsequently high echo density is simulated using sampled analog signal Schroeder sections. A sampled data analog signal device is used which employs charge transfer devices as programmable delay media, multiplying digital to analog converters to generate weighting of the analog signal and a general purpose microprocessor as a parameter calculator. The early impulse response characteristic is accomplished by providing a time domain finite impulse response (FIR) section in a transversal filter arrangement which directly feeds to an ultimate output and to parallel comb filter sections which in turn input to at least one all pass section of the so-called Schroeder configuration. The charge transfer devices are employed as programmable delays to propagate the analog signals.Type: GrantFiled: May 5, 1980Date of Patent: July 6, 1982Assignee: The Regents of the University of CaliforniaInventor: Nelson H. Morgan