Patents by Inventor Wei Ping

Wei Ping 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: 11002440
    Abstract: A light source module includes a transparent conductive substrate and plural light source groups. The transparent conductive substrate includes a substrate base, plural positive electrodes, plural negative electrodes and plural gaps. The plural positive electrodes and the plural negative electrodes are installed on the substrate base. Each light source group includes plural light-emitting elements. The plural positive electrodes are separately arranged. Consequently, the resistances of the light-emitting elements are substantially equal.
    Type: Grant
    Filed: July 27, 2020
    Date of Patent: May 11, 2021
    Assignee: PRIMAX ELECTRONICS LTD.
    Inventors: Wei-Ping Chan, Wei-Chiang Huang, Chuan-Tai Hsiao, Xiu-Mei Pan
  • Publication number: 20210123996
    Abstract: A gradient coil for an MRI apparatus is disclosed, comprising a main coil layer and a shielding coil layer. The shielding coil layer is arranged around the main coil layer, which includes an X, Y, and Z main coil set, and an X and Y end shielding coil set. The X end shielding coil set is connected in series with the X main coil set and used to form a shielding magnetic field in a direction opposite to the X direction. The Y end shielding coil set is connected in series with the Y main coil set and used to form a shielding magnetic field in the opposite direction to the Y direction. In the Z direction, the X end shielding coil set and Y end shielding coil set are disposed outside an imaging region enclosed by the main coil layer.
    Type: Application
    Filed: October 28, 2020
    Publication date: April 29, 2021
    Applicant: Siemens Healthcare GmbH
    Inventors: Zhen Hua Yue, Wei Ping Peng
  • Publication number: 20210118423
    Abstract: According to various embodiments, an end-to-end TTS framework can integrate a watermarking process into the training of the TTS framework, which enables watermarks to be imperceptible within a synthesized/cloned audio segment generated by the TTS framework. The watermarks added in such a matter are statistically undetectable to prevent authorized removal. According to an exemplary method of training the TTS framework, a TTS neural network model and a watermarking neural network mode in the TTS framework are trained in an end to end manner, with the watermarking being part of the optimization process of the TTS framework. During the training, neuron values of the TTS neural network model are adjusted based on training data to prepare one or more spaces for adding a watermark in a synthesized audio segment to be generated by the TTS framework.
    Type: Application
    Filed: October 21, 2019
    Publication date: April 22, 2021
    Inventors: Wei PING, Zhenyu ZHONG, Yueqiang CHENG, Xing LI, Tao WEI
  • Publication number: 20210110810
    Abstract: Described herein are embodiments of an end-to-end text-to-speech (TTS) system with parallel wave generation. In one or more embodiments, a Gaussian inverse autoregressive flow is distilled from an autoregressive WaveNet by minimizing a novel regularized Kullback-Leibler (KL) divergence between their highly-peaked output distributions. Embodiments of the methodology computes the KL divergence in a closed-form, which simplifies the training process and provides very efficient distillation. Embodiments of a novel text-to-wave neural architecture for speech synthesis are also described, which are fully convolutional and enable fast end-to-end training from scratch. These embodiments significantly outperform the previous pipeline that connects a text-to-spectrogram model to a separately trained WaveNet. Also, a parallel waveform synthesizer embodiment conditioned on the hidden representation in an embodiment of this end-to-end model were successfully distilled.
    Type: Application
    Filed: December 21, 2020
    Publication date: April 15, 2021
    Applicant: Baidu USA LLC
    Inventors: Wei PING, Kainan PENG, Jitong CHEN
  • Publication number: 20210111220
    Abstract: A method for manufacturing light-shielded cameras includes forming a plurality of camera dies by dicing a camera wafer stack including (a) an image sensor wafer having a plurality of image sensors, (b) a cover glass bonded to the image sensor wafer, and (c) a lens wafer bonded to the cover glass and having a plurality of lenses, to form a plurality of camera dies. The method further includes, prior to the step of dicing, (i) from a first side of the image sensor wafer facing away from the cover glass and at least partly covered by an opaque layer, pre-cutting a sensor-cover wafer stack that includes the image sensor wafer and the cover glass, and (ii) depositing an opaque material in the pre-cuts. The method also includes, after the step of dicing, applying an opaque coating to second-side surfaces, of the camera dies, formed by the step of dicing.
    Type: Application
    Filed: October 11, 2019
    Publication date: April 15, 2021
    Inventors: Teng-Sheng CHEN, Wei-Ping CHEN, Ying CHUNG, Chen-Yu HUNG
  • Publication number: 20210103159
    Abstract: A passive speckle-suppressing diffuser includes a microlens array for diffusing a light field originating from one or more coherent light beams, and a diffractive optical element mounted in series with the microlens array and having a pixelated thickness distribution, characterized by a spatial variation across the diffractive optical element, to impose a spatially varying phase shift on the light field. The pixelated thickness distribution may be configured such that the spatially varying phase shift suppresses speckle of the light field while minimizing introduction of distinct diffraction structure.
    Type: Application
    Filed: October 7, 2019
    Publication date: April 8, 2021
    Inventors: Shih-Hsin HSU, Jau-Jan DENG, Wei-Ping CHEN
  • Publication number: 20210090547
    Abstract: WaveFlow is a small-footprint generative flow for raw audio, which may be directly trained with maximum likelihood. WaveFlow handles the long-range structure of waveform with a dilated two-dimensional (2D) convolutional architecture, while modeling the local variations using expressive autoregressive functions. WaveFlow may provide a unified view of likelihood-based models for raw audio, including WaveNet and WaveGlow, which may be considered special cases. It generates high-fidelity speech, while synthesizing several orders of magnitude faster than existing systems since it uses only a few sequential steps to generate relatively long waveforms. WaveFlow significantly reduces the likelihood gap that has existed between autoregressive models and flow-based models for efficient synthesis. Its small footprint with 5.91M parameters makes it 15 times smaller than some existing models. WaveFlow can generate 22.05 kHz high-fidelity audio 42.
    Type: Application
    Filed: August 5, 2020
    Publication date: March 25, 2021
    Applicant: Baidu USA LLC
    Inventors: Wei PING, Kainan PENG, Kexin ZHAO, Zhao SONG
  • Patent number: 10950520
    Abstract: An electronic package is provided. A heat dissipator is bonded via a thermal interface layer to an electronic component disposed on a carrier. The heat dissipator has a concave-convex structure to increase a heat-dissipating area of the thermal interface layer. Therefore, the heat dissipator has a better heat-dissipating effect.
    Type: Grant
    Filed: August 6, 2019
    Date of Patent: March 16, 2021
    Assignee: Siliconware Precision Industries Co., Ltd.
    Inventors: Yu-Lung Huang, Chee-Key Chung, Chang-Fu Lin, Kuo-Hua Yu, Wen-Shan Tsai, En-Li Lin, Kaun-I Cheng, Wei-Ping Wang
  • Publication number: 20210056557
    Abstract: A method for verifying the authenticity of a point of sale prior to the initiation of a payment transaction using unique merchant-specific data and blockchain registration includes: receiving identification data from a point of sale device; receiving data comprising a permissioned blockchain, the data including a plurality of blockchain data values, each data value including at least a set of merchant data; verifying the point of sale device by identifying, in the plurality of blockchain data values, a set of merchant data that matches the received identification data; and outputting a notification to a user of the mobile communication device if verification of the point of sale device fails, or transmitting payment credentials stored in the mobile communication device to the point of sale device if verification of the point of sale device is successful.
    Type: Application
    Filed: August 20, 2019
    Publication date: February 25, 2021
    Applicant: MASTERCARD INTERNATIONAL INCORPORATED
    Inventors: Prashant SHARMA, Amy WU, Philip Wei Ping YEN, Rajat MAHESHWARI
  • Patent number: 10931100
    Abstract: A current limiting circuit for controlling current from a power supply to a load having a capacitance includes an inductor, a transistor coupled in a current path with the inductor, and a control circuit. The transistor includes a control terminal. The control circuit is coupled to sense a voltage across the inductor and coupled to the control terminal of the transistor. The control circuit is configured to turn off the transistor when the voltage across the inductor is greater than a threshold to restrict current from a power supply, and turn on the transistor when a defined parameter is met to allow current from the power supply to charge the load capacitance. Other example current limiting circuits are also disclosed.
    Type: Grant
    Filed: December 26, 2017
    Date of Patent: February 23, 2021
    Assignee: Astec International Limited
    Inventors: Wei Ping Li, Wei Jia Yan, Wen Jian Liao, Xin Zhang
  • Publication number: 20210049999
    Abstract: Described herein are systems and methods for augmenting neural speech synthesis networks with low-dimensional trainable speaker embeddings in order to generate speech from different voices from a single model. As a starting point for multi-speaker experiments, improved single-speaker model embodiments, which may be referred to generally as Deep Voice 2 embodiments, were developed, as well as a post-processing neural vocoder for Tacotron (a neural character-to-spectrogram model). New techniques for multi-speaker speech synthesis were performed for both Deep Voice 2 and Tacotron embodiments on two multi-speaker TTS datasets—showing that neural text-to-speech systems can learn hundreds of unique voices from twenty-five minutes of audio per speaker.
    Type: Application
    Filed: November 2, 2020
    Publication date: February 18, 2021
    Applicant: Baidu USA LLC
    Inventors: Sercan O. ARIK, Gregory DIAMOS, Andrew GIBIANSKY, John MILLER, Kainan PENG, Wei PING, Jonathan RAIMAN, Yanqi ZHOU
  • Patent number: 10903088
    Abstract: An electronic package and a method for fabrication the same are provided. The method includes: disposing an electronic component on a substrate; forming an encapsulant layer on the substrate to encapsulate the electronic component; and forming a shielding layer made of metal on the encapsulant layer. The shielding layer has an extending portion extending to a lateral side of the substrate along a corner of the encapsulant layer, without extending to a lower side of the substrate. Therefore, the present disclosure prevents the shielding layer from coming into contact with conductive pads disposed on the lower side of the substrate and thereby avoids a short circuit from occurrence.
    Type: Grant
    Filed: June 20, 2017
    Date of Patent: January 26, 2021
    Assignee: Siliconware Precision Industries Co., Ltd.
    Inventors: Shu-Chi Chang, Wei Ping Wang, Hsien-Lung Hsiao, Kaun-I Cheng
  • Patent number: 10900493
    Abstract: A luminous fan includes a fan frame, a main circuit board, a light guide element and a light-guiding impeller. The main circuit board is disposed on a bracket of the fan frame and includes a first light-emitting element. The light guide element is located over the main circuit board, and includes a light input surface and light output surface. A light guide post is disposed on the light input surface. A reflective structure is formed on the light output surface. The light-guiding impeller is pivotally coupled to the bracket. The first light-emitting element emits a first light beam. After the first light beam is transmitted to the reflective structure through the light guide post, the first light beam is reflected by the at least one reflective structure and diffused toward a middle region of the light guide element.
    Type: Grant
    Filed: March 25, 2019
    Date of Patent: January 26, 2021
    Assignee: PRIMAX ELECTRONICS LTD.
    Inventors: Wei-Chiang Huang, Wei-Ping Chan
  • Patent number: 10896669
    Abstract: Described herein are systems and methods for augmenting neural speech synthesis networks with low-dimensional trainable speaker embeddings in order to generate speech from different voices from a single model. As a starting point for multi-speaker experiments, improved single-speaker model embodiments, which may be referred to generally as Deep Voice 2 embodiments, were developed, as well as a post-processing neural vocoder for Tacotron (a neural character-to-spectrogram model). New techniques for multi-speaker speech synthesis were performed for both Deep Voice 2 and Tacotron embodiments on two multi-speaker TTS datasets—showing that neural text-to-speech systems can learn hundreds of unique voices from twenty-five minutes of audio per speaker.
    Type: Grant
    Filed: May 8, 2018
    Date of Patent: January 19, 2021
    Assignee: Baidu USA LLC
    Inventors: Sercan O. Arik, Gregory Diamos, Andrew Gibiansky, John Miller, Kainan Peng, Wei Ping, Jonathan Raiman, Yanqi Zhou
  • Patent number: 10885833
    Abstract: A display device is disposed within a casing of an electronic device. The display device includes a display part, at least one light source, a projection part and an imaging structure. The display part is disposed on the casing and exposed outside. The at least one light source is disposed within the casing, and emits at least one light beam. The projection part is arranged between the at least one light source and the display part, and has an image region. After the at least one light beam passes through the projection part, a projection pattern corresponding to the image region is generated. The imaging structure is disposed on the display part. The imaging structure receives the projection pattern, so that the projection pattern is shown on the display part.
    Type: Grant
    Filed: July 22, 2019
    Date of Patent: January 5, 2021
    Assignee: PRIMAX ELECTRONICS LTD.
    Inventors: Han-Pin Kao, Wei-Chiang Huang, Wei-Ping Chan
  • Patent number: 10872543
    Abstract: A display device includes a display plate, a micro light source, a first reflecting element and a second reflecting element. The display plate has a pattern. The micro light source emits a light beam. The first reflecting element is arranged beside a first side of the display plate. A first portion of the light beam is reflected by the first reflecting element. A second portion of the light beam is transmitted through the first reflecting element. The second reflecting element is arranged beside a second side of the display plate to reflect the light beam. The light beam is repeatedly reflected by the first reflecting element and the second reflecting element, so that a multiple-luminous effect of the pattern is provided.
    Type: Grant
    Filed: May 28, 2019
    Date of Patent: December 22, 2020
    Assignee: PRIMAX ELECTRONICS LTD.
    Inventors: Wei-Chiang Huang, Wei-Ping Chan
  • Patent number: 10872596
    Abstract: Described herein are embodiments of an end-to-end text-to-speech (TTS) system with parallel wave generation. In one or more embodiments, a Gaussian inverse autoregressive flow is distilled from an autoregressive WaveNet by minimizing a novel regularized Kullback-Leibler (KL) divergence between their highly-peaked output distributions. Embodiments of the methodology computes the KL divergence in a closed-form, which simplifies the training process and provides very efficient distillation. Embodiments of a novel text-to-wave neural architecture for speech synthesis are also described, which are fully convolutional and enable fast end-to-end training from scratch. These embodiments significantly outperform the previous pipeline that connects a text-to-spectrogram model to a separately trained WaveNet. Also, a parallel waveform synthesizer embodiment conditioned on the hidden representation in an embodiment of this end-to-end model were successfully distilled.
    Type: Grant
    Filed: February 15, 2019
    Date of Patent: December 22, 2020
    Assignee: Baidu USA LLC
    Inventors: Wei Ping, Kainan Peng, Jitong Chen
  • Publication number: 20200376471
    Abstract: Catalyst compositions comprising a zeolite and a mesoporous support or binder are disclosed. The mesoporous support or binder comprises a mesoporous metal oxide having a particle diameter of greater than or equal to 20 ?m at 50% of the cumulative pore size distribution (d50). Also disclosed are processes for producing a mono-alkylated aromatic compound (e.g., ethylbenzene or cumene) which exhibit improved yield of the mono-alkylated aromatic compound using alkylation catalysts comprising one or more of these catalyst compositions.
    Type: Application
    Filed: March 19, 2018
    Publication date: December 3, 2020
    Inventors: Matthew S. Ide, Brett T. Loveless, Doron Levin, William W. Lonergan, Matthew J. Vincent, Wei-Ping Tai, Federico Barrai
  • Publication number: 20200381747
    Abstract: A formulation of an interconnect of a fuel cell includes chrome powder and chrome-iron alloy powder. A ratio of a sum of chrome in the chrome powder and the chrome-iron alloy powder is in a range between 80% in weight and 95% in weight, and a ratio of a sum of iron in the chrome powder and the chrome-iron alloy powder is in a range between 5% in weight and 20% in weight.
    Type: Application
    Filed: March 24, 2020
    Publication date: December 3, 2020
    Inventors: SHING YIN LIAO, WEI HSUN HSU, WEI PING HUANG, HUEI LONG LEE, CHIU LUNG CHU
  • Patent number: 10796686
    Abstract: Described herein are embodiments of a fully-convolutional attention-based neural text-to-speech (TTS) system, which various embodiments may generally be referred to as Deep Voice 3. Embodiments of Deep Voice 3 match state-of-the-art neural speech synthesis systems in naturalness while training ten times faster. Deep Voice 3 embodiments were scaled to data set sizes unprecedented for TTS, training on more than eight hundred hours of audio from over two thousand speakers. In addition, common error modes of attention-based speech synthesis networks were identified and mitigated, and several different waveform synthesis methods were compared. Also presented are embodiments that describe how to scale inference to ten million queries per day on one single-GPU server.
    Type: Grant
    Filed: August 8, 2018
    Date of Patent: October 6, 2020
    Assignee: Baidu USA LLC
    Inventors: Sercan O. Arik, Wei Ping, Kainan Peng, Sharan Narang, Ajay Kannan, Andrew Gibiansky, Jonathan Raiman, John Miller