Patents by Inventor Shuhua Zhang
Shuhua Zhang 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).
-
Patent number: 11869478Abstract: A device includes one or more processors configured to receive an input audio signal. The one or more processors are also configured to process the input audio signal based on a combined representation of multiple sound sources to generate an output audio signal. The combined representation is used to selectively retain or remove sounds of the multiple sound sources from the input audio signal. The one or more processors are further configured to provide the output audio signal to a second device.Type: GrantFiled: March 18, 2022Date of Patent: January 9, 2024Assignee: QUALCOMM IncorporatedInventors: Siddhartha Goutham Swaminathan, Sunkuk Moon, Shuhua Zhang, Erik Visser
-
Patent number: 11806670Abstract: The present disclosure provides a system for extracting long chain dicarboxylic acid, the system comprising: a primary membrane filtration unit, a first crystallization unit, a first separation unit, a first dissolution tank, a secondary membrane filtration unit, a second separation unit, a second crystallization unit and a third separation unit. By the system for extracting long chain dicarboxylic acid of an embodiment of the present invention, the resulted long chain dicarboxylic acid product has a high purity, very low and even no residual alkane residue, and organic solvent-free.Type: GrantFiled: December 15, 2021Date of Patent: November 7, 2023Assignees: Cathay Biotech Inc., CIBT America Inc., Cathay (Jinxiang) Biomaterial Co., Ltd.Inventors: Chen Yang, Shuhua Zhang, Yufeng Yang, Xiucai Liu
-
Publication number: 20230326477Abstract: A device to perform speech enhancement includes one or more processors configured to process image data to detect at least one of an emotion, a speaker characteristic, or a noise type. The one or more processors are also configured to generate context data based at least in part on the at least one of the emotion, the speaker characteristic, or the noise type. The one or more processors are further configured to obtain input spectral data based on an input signal. The input signal represents sound that includes speech. The one or more processors are also configured to process, using a multi-encoder transformer, the input spectral data and the context data to generate output spectral data that represents a speech enhanced version of the input signal.Type: ApplicationFiled: June 14, 2023Publication date: October 12, 2023Inventors: Kyungguen BYUN, Shuhua ZHANG, Lae-Hoon KIM, Erik VISSER, Sunkuk MOON, Vahid MONTAZERI
-
Publication number: 20230298561Abstract: A device includes one or more processors configured to receive an input audio signal. The one or more processors are also configured to process the input audio signal based on a combined representation of multiple sound sources to generate an output audio signal. The combined representation is used to selectively retain or remove sounds of the multiple sound sources from the input audio signal. The one or more processors are further configured to provide the output audio signal to a second device.Type: ApplicationFiled: March 18, 2022Publication date: September 21, 2023Inventors: Siddhartha Goutham SWAMINATHAN, Sunkuk Moon, Shuhua Zhang, Erik Visser
-
Publication number: 20230260525Abstract: A device includes a memory configured to store untransformed ambisonic coefficients at different time segments. The device includes one or more processors configured to obtain the untransformed ambisonic coefficients at the different time segments, where the untransformed ambisonic coefficients at the different time segments represent a soundfield at the different time segments. The one or more processors are configured to apply one adaptive network, based on a constraint that includes preservation of a spatial direction of one or more audio sources in the soundfield at the different time segments, to the untransformed ambisonic coefficients at the different time segments to generate transformed ambisonic coefficients at the different time segments, wherein the transformed ambisonic coefficients at the different time segments represent a modified soundfield at the different time segments, that was modified based on the constraint.Type: ApplicationFiled: April 24, 2023Publication date: August 17, 2023Inventors: Lae-Hoon KIM, Shankar THAGADUR SHIVAPPA, S M Akramus SALEHIN, Shuhua ZHANG, Erik VISSER
-
Patent number: 11721353Abstract: A device includes one or more processors configured to obtain audio signals representing sound captured by at least three microphones and determine spatial audio data based on the audio signals. The one or more processors are further configured to determine a metric indicative of wind noise in the audio signals. The metric is based on a comparison of a first value and a second value. The first value corresponds to an aggregate signal based on the spatial audio data, and the second value corresponds to a differential signal based on the spatial audio data.Type: GrantFiled: December 21, 2020Date of Patent: August 8, 2023Assignee: Qualcomm IncorporatedInventors: S M Akramus Salehin, Lae-Hoon Kim, Hannes Pessentheiner, Shuhua Zhang, Sanghyun Chi, Erik Visser, Shankar Thagadur Shivappa
-
Patent number: 11715480Abstract: A device to perform speech enhancement includes one or more processors configured to obtain input spectral data based on an input signal. The input signal represents sound that includes speech. The one or more processors are also configured to process, using a multi-encoder transformer, the input spectral data and context data to generate output spectral data that represents a speech enhanced version of the input signal.Type: GrantFiled: March 23, 2021Date of Patent: August 1, 2023Assignee: QUALCOMM IncorporatedInventors: Kyungguen Byun, Shuhua Zhang, Lae-Hoon Kim, Erik Visser, Sunkuk Moon, Vahid Montazeri
-
Patent number: 11705147Abstract: Systems, methods and computer-readable media are provided for speech enhancement using a hybrid neural network. An example process can include receiving, by a first neural network portion of the hybrid neural network, audio data and reference data, the audio data including speech data, noise data, and echo data; filtering, by the first neural network portion, a portion of the audio data based on adapted coefficients of the first neural network portion, the portion of the audio data including the noise data and/or echo data; based on the filtering, generating, by the first neural network portion, filtered audio data including the speech data and an unfiltered portion of the noise data and/or echo data; and based on the filtered audio data and the reference data, extracting, by a second neural network portion of the hybrid neural network, the speech data from the filtered audio data.Type: GrantFiled: April 28, 2021Date of Patent: July 18, 2023Assignee: QUALCOMM IncorporatedInventors: Erik Visser, Vahid Montazeri, Shuhua Zhang, Lae-Hoon Kim
-
Patent number: 11676571Abstract: A device for speech generation includes one or more processors configured to receive one or more control parameters indicating target speech characteristics. The one or more processors are also configured to process, using a multi-encoder, an input representation of speech based on the one or more control parameters to generate encoded data corresponding to an audio signal that represents a version of the speech based on the target speech characteristics.Type: GrantFiled: January 21, 2021Date of Patent: June 13, 2023Assignee: QUALCOMM IncorporatedInventors: Kyungguen Byun, Sunkuk Moon, Shuhua Zhang, Vahid Montazeri, Lae-Hoon Kim, Erik Visser
-
Patent number: 11636866Abstract: A device includes a memory configured to store untransformed ambisonic coefficients at different time segments. The device also includes one or more processors configured to obtain the untransformed ambisonic coefficients at the different time segments, where the untransformed ambisonic coefficients at the different time segments represent a soundfield at the different time segments. The one or more processors are also configured to apply one adaptive network, based on a constraint, to the untransformed ambisonic coefficients at the different time segments to generate transformed ambisonic coefficients at the different time segments, wherein the transformed ambisonic coefficients at the different time segments represent a modified soundfield at the different time segments, that was modified based on the constraint.Type: GrantFiled: March 23, 2021Date of Patent: April 25, 2023Assignee: Qualcomm IncorporatedInventors: Lae-Hoon Kim, Shankar Thagadur Shivappa, S M Akramus Salehin, Shuhua Zhang, Erik Visser
-
Publication number: 20220340713Abstract: A method for producing an organopolysiloxane having at least one silicon atom-bonded alkoxysilylalkyl group per molecule is provided. The method comprises the following steps: (i) treating an organopolysiloxane having at least one silicon atom-bonded hydrogen atom per molecule with an alkali salt of carboxylic acid; (ii) removing the alkali salt from the organopolysiloxane; and (iii) reacting the organopolysiloxane with an alkenyl group-containing alkoxysilane in the presence of a platinum-based catalyst. The organopolysiloxane is generally obtained by the method without hydrolyzing any alkoxysilylalkyl groups, and is useful as a surface modifier for various types of fillers.Type: ApplicationFiled: September 3, 2019Publication date: October 27, 2022Inventors: Peng LI, Xiaocong XU, Hongjun MA, Yan ZHENG, Shuhua ZHANG, Yejun WU, Hui ZHU
-
Publication number: 20220310108Abstract: A device to perform speech enhancement includes one or more processors configured to obtain input spectral data based on an input signal. The input signal represents sound that includes speech. The one or more processors are also configured to process, using a multi-encoder transformer, the input spectral data and context data to generate output spectral data that represents a speech enhanced version of the input signal.Type: ApplicationFiled: March 23, 2021Publication date: September 29, 2022Inventors: Kyungguen BYUN, Shuhua ZHANG, Lae-Hoon KIM, Erik VISSER, Sunkuk MOON, Vahid MONTAZERI
-
Publication number: 20220230623Abstract: A device for speech generation includes one or more processors configured to receive one or more control parameters indicating target speech characteristics. The one or more processors are also configured to process, using a multi-encoder, an input representation of speech based on the one or more control parameters to generate encoded data corresponding to an audio signal that represents a version of the speech based on the target speech characteristics.Type: ApplicationFiled: January 21, 2021Publication date: July 21, 2022Applicant: QUALCOMM IncorporatedInventors: Kyungguen BYUN, Sunkuk MOON, Shuhua ZHANG, Vahid MONTAZERI, Lae-Hoon KIM, Erik VISSER
-
Publication number: 20220199100Abstract: A device includes one or more processors configured to obtain audio signals representing sound captured by at least three microphones and determine spatial audio data based on the audio signals. The one or more processors are further configured to determine a metric indicative of wind noise in the audio signals. The metric is based on a comparison of a first value and a second value. The first value corresponds to an aggregate signal based on the spatial audio data, and the second value corresponds to a differential signal based on the spatial audio data.Type: ApplicationFiled: December 21, 2020Publication date: June 23, 2022Inventors: S M Akramus SALEHIN, Lae-Hoon KIM, Hannes PESSENTHEINER, Shuhua ZHANG, Sanghyun CHI, Erik VISSER, Shankar THAGADUR SHIVAPPA
-
Patent number: 11353904Abstract: A voltage regulator system includes a switch system including a power switch to conduct an output current through an inductor based on an input voltage and a switching signal to generate an output voltage at a load. A feedback system generates a PWM signal based on the output voltage and based on a variable reference voltage. A gate driver system generates the switching signal based on the PWM signal. The gate driver system controls the switch system to increase the output voltage at output voltage slopes in each of startup stages during startup of the voltage regulator system. A sampling system samples the output current and the output voltage during the startup of the voltage regulator system to measure each slope of the output voltage slopes at each of the respective startup stages during the startup of the voltage regulator system.Type: GrantFiled: April 30, 2019Date of Patent: June 7, 2022Assignee: TEXAS INSTRUMENTS INCORPORATEDInventors: Kuang-Yao Cheng, Wenkai Wu, Preetam Tadeparthy, Shuhua Zhang
-
Publication number: 20220105469Abstract: The present disclosure provides a system for extracting long chain dicarboxylic acid, the system comprising: a primary membrane filtration unit, a first crystallization unit, a first separation unit, a first dissolution tank, a secondary membrane filtration unit, a second separation unit, a second crystallization unit and a third separation unit. By the system for extracting long chain dicarboxylic acid of an embodiment of the present invention, the resulted long chain dicarboxylic acid product has a high purity, very low and even no residual alkane residue, and organic solvent-free.Type: ApplicationFiled: December 15, 2021Publication date: April 7, 2022Applicants: Cathay Biotech Inc., CIBT America Inc., Cathay (Jinxiang) Biomaterial Co., Ltd.Inventors: Chen Yang, Shuhua Zhang, Yufeng Yang, Xiucai Liu
-
Publication number: 20220106249Abstract: The present disclosure provides a method for extracting long chain dicarboxylic acid, the method comprising: (1) subjecting a long chain dicarboxylic acid fermentation broth to a primary membrane filtration treatment to give a first filtrate; subjecting the first filtrate to decolorization, crystallization, and solid-liquid separation to give a first solid; (2) redissolving the first solid in water to form a solution; subjecting the solution to decolorization, crystallization by acidification, and solid-liquid separation to give a second solid. By the method, the resulted long chain dicarboxylic acid product has a high purity and no residual organic solvent.Type: ApplicationFiled: December 15, 2021Publication date: April 7, 2022Applicants: Cathay Biotech Inc., CIBT America Inc., Cathay (Jinxiang) Biomaterial Co., Ltd.Inventors: Chen Yang, Shuhua Zhang, Yufeng Yang, Xiucai Liu
-
Patent number: 11229881Abstract: The present disclosure provides a method and a system for extracting long chain dicarboxylic acid, the method comprising: (1) subjecting a long chain dicarboxylic acid fermentation broth to a primary membrane filtration treatment to give a first filtrate; subjecting the first filtrate to decolorization, acidification/crystallization, and solid-liquid separation treatments to give a first solid; (2) mixing the first solid, a base and water to form a solution; subjecting the solution to a secondary membrane filtration treatment to give a second filtrate; subjecting the second filtrate to decolorization, acidification/crystallization, and solid-liquid separation treatments to give a second solid; and (3) mixing the second solid and water to form a mixture; subjecting the mixture to a thermostatic treatment at 105-150° C., followed by cooling for crystallization and solid-liquid separation treatment.Type: GrantFiled: May 29, 2020Date of Patent: January 25, 2022Assignees: Cathay Biotech Inc., CIBT America Inc., Cathay (Jinxiang) Biomaterial Co., Ltd.Inventors: Chen Yang, Shuhua Zhang, Yufeng Yang, Xiucai Liu
-
Publication number: 20210343306Abstract: Systems, methods and computer-readable media are provided for speech enhancement using a hybrid neural network. An example process can include receiving, by a first neural network portion of the hybrid neural network, audio data and reference data, the audio data including speech data, noise data, and echo data; filtering, by the first neural network portion, a portion of the audio data based on adapted coefficients of the first neural network portion, the portion of the audio data including the noise data and/or echo data; based on the filtering, generating, by the first neural network portion, filtered audio data including the speech data and an unfiltered portion of the noise data and/or echo data; and based on the filtered audio data and the reference data, extracting, by a second neural network portion of the hybrid neural network, the speech data from the filtered audio data.Type: ApplicationFiled: April 28, 2021Publication date: November 4, 2021Inventors: Erik VISSER, Vahid MONTAZERI, Shuhua ZHANG, Lae-Hoon KIM
-
Publication number: 20210304777Abstract: A device includes a memory configured to store untransformed ambisonic coefficients at different time segments. The device also includes one or more processors configured to obtain the untransformed ambisonic coefficients at the different time segments, where the untransformed ambisonic coefficients at the different time segments represent a soundfield at the different time segments. The one or more processors are also configured to apply one adaptive network, based on a constraint, to the untransformed ambisonic coefficients at the different time segments to generate transformed ambisonic coefficients at the different time segments, wherein the transformed ambisonic coefficients at the different time segments represent a modified soundfield at the different time segments, that was modified based on the constraint.Type: ApplicationFiled: March 23, 2021Publication date: September 30, 2021Inventors: Lae-Hoon KIM, Shankar THAGADUR SHIVAPPA, S M Akramus SALEHIN, Shuhua ZHANG, Erik VISSER