公开(公告)号：US09408556B2

公开(公告)日：2016-08-09

申请号：US13990007

申请日：2010-12-01

Applicant: Ping Wang ,  Di Wang ,  Le Wang ,  Jin Yu ,  Kai Yu ,  Yishan Wang ,  Lin Wang ,  Cong Zhao

Inventor： Ping Wang ,  Di Wang ,  Le Wang ,  Jin Yu ,  Kai Yu ,  Yishan Wang ,  Lin Wang ,  Cong Zhao

IPC: A61B5/08 ,  G01N33/00 ,  G01N33/497 ,  A61B5/097

CPC classification number: A61B5/082 ,  A61B5/097 ,  G01N33/0047 ,  G01N33/497

Abstract: The present embodiments disclose a device including a module for sampling, separating and enriching a detected object, an exhaled breath condensates (EBCs) detection module and a combined volatile organic compounds (VOCs) detection module. The sampling module is connected with the EBCs detection module via a syringe pump for sample injection and is connected with the combined VOCs detection module by a capillary separation column. EBCs and VOCs in human exhaled breath are simultaneously sampled, separated and condensed; the heavy metal ions, cell factors, etc. in the collected EBCs are detected with a light addressable potentiometric sensor (LAPS); the condensed VOCs can be quantitatively detected by the combined VOCs detection module with a high sensitivity; and a heating rod and a platinum resistor can be conveniently replaced because a separated outlet heating piece is designed in the combined VOCs detection module.

Abstract translation: 本实施例公开了一种包括用于对检测到的物体进行取样，分离和富集的模块，呼出呼吸冷凝物（EBC）检测模块和组合的挥发性有机化合物（VOC））检测模块的装置。 采样模块通过注射泵与EBC检测模块连接，用于注射样品，并通过毛细管分离柱与组合的VOC检测模块连接。 人呼吸中的EBC和VOCs被同时取样，分离和浓缩; 用光寻址电位传感器（LAPS）检测收集的EBC中的重金属离子，细胞因子等; 可以通过组合的VOC检测模块以高灵敏度定量检测浓缩的VOC; 并且可以方便地更换加热棒和铂电阻器，因为在组合的VOC检测模块中设计了分离的出口加热片。

公开(公告)号：US08873843B2

公开(公告)日：2014-10-28

申请号：US13479256

申请日：2012-05-23

Applicant: Shenghuo Zhu ,  Chang Huang ,  Kai Yu

Inventor： Shenghuo Zhu ,  Chang Huang ,  Kai Yu

IPC: G06K9/62

CPC classification number: G06K9/6276 ,  G06K9/6201

Abstract: A nearest-neighbor-based distance metric learning process includes applying an exponential-based loss function to provide a smooth objective; and determining an objective and a gradient of both hinge-based and exponential-based loss function in a quadratic time of the number of instances using a computer.

Abstract translation: 基于最近邻的距离度量学习过程包括应用基于指数的损失函数来提供平滑的目标; 以及使用计算机在实例数量的二次时间中确定基于铰链和指数的损失函数的目标和梯度。

公开(公告)号：US08774085B2

公开(公告)日：2014-07-08

申请号：US12218276

申请日：2008-07-14

Applicant: Angeliki Alexiou ,  Federico Boccardi ,  Kai Yu

Inventor： Angeliki Alexiou ,  Federico Boccardi ,  Kai Yu

IPC: H04W4/00

CPC classification number: H04B7/022 ,  H04B7/026 ,  H04B7/04 ,  H04B7/0691 ,  H04B7/15592

Abstract: A method is provided, in a wireless communications network comprising a source node, a destination node and at least one relay node, of selecting spatial subchannels for use. The method comprising the steps of: spatially decomposing channels into spatial subchannels; and selecting a subset of the subchannels for use that at least approximately maximises predicted throughput rate.

Abstract translation: 提供了一种在包括源节点，目的地节点和至少一个中继节点的无线通信网络中选择空间子信道以供使用的方法。 该方法包括以下步骤：将信道空间分解为空间子信道; 以及选择使用的子信道的子集，所述子信道至少近似地使预测吞吐率最大化。

公开(公告)号：US08751846B2

公开(公告)日：2014-06-10

申请号：US13193929

申请日：2011-07-29

Applicant: Kai Yu ,  Wangsheng Xie

Inventor： Kai Yu ,  Wangsheng Xie

IPC: G06F1/00 ,  G06F1/32

CPC classification number: H03K17/693 ,  H03K2217/0018 ,  H03K2217/0036 ,  Y10T307/696

Abstract: In the field of electronic technologies, a power supply selector and a power supply selection method are provided. The power supply selector includes: a first selection module, configured to select a power supply from multiple candidate power supplies; a control module, coupled to the first selection module, and configured to use the power supply selected by the first selection module as a power supply, and compare voltages of the multiple candidate power supplies to generate a control signal of each candidate power supply; and a second selection module, coupled to the control module, and configured to select a power supply for output in the multiple candidate power supplies under the control of the control signal of each candidate power supply. The technical solution is used to select a power supply from multiple candidate power supplies.

Abstract translation: 在电子技术领域中，提供电源选择器和电源选择方法。 电源选择器包括：第一选择模块，被配置为从多个候选电源中选择电源; 控制模块，耦合到第一选择模块，并且被配置为使用由第一选择模块选择的电源作为电源，并且比较多个候选电源的电压以产生每个候选电源的控制信号; 以及耦合到控制模块的第二选择模块，并且被配置为在每个候选电源的控制信号的控制下选择用于在多个候选电源中输出的电源。 该技术解决方案用于从多个候选电源中选择电源。

公开(公告)号：US08374979B2

公开(公告)日：2013-02-12

申请号：US12814449

申请日：2010-06-12

Applicant: Xi Zhou ,  Jinjun Wang ,  Fengjun Lv ,  Kai Yu

Inventor： Xi Zhou ,  Jinjun Wang ,  Fengjun Lv ,  Kai Yu

IPC: G06F15/18

CPC classification number: G06K9/469 ,  G06K9/6892

Abstract: Systems and methods are disclosed to perform image parsing on one or more images by identifying a set of similar regions from each image; assigning one or more region labels to each region and generating multiple hypotheses for region label assignment; and detecting class, location and boundary of each object in the image, wherein object classification, detection and segmentation are performed jointly during image parsing.

Abstract translation: 公开了系统和方法，以通过从每个图像识别一组相似的区域来对一个或多个图像执行图像解析; 为每个区域分配一个或多个区域标签，并为区域标签分配生成多个假设; 并检测图像中每个对象的类别，位置和边界，其中在图像解析期间共同执行对象分类，检测和分割。

公开(公告)号：US08345962B2

公开(公告)日：2013-01-01

申请号：US12277504

申请日：2008-11-25

Applicant: Kai Yu ,  Wei Xu ,  Yihong Gong

Inventor： Kai Yu ,  Wei Xu ,  Yihong Gong

IPC: G06K9/62

CPC classification number: G06K9/6256 ,  G06N3/08

Abstract: A method and system for training a neural network of a visual recognition computer system, extracts at least one feature of an image or video frame with a feature extractor; approximates the at least one feature of the image or video frame with an auxiliary output provided in the neural network; and measures a feature difference between the extracted at least one feature of the image or video frame and the approximated at least one feature of the image or video frame with an auxiliary error calculator. A joint learner of the method and system adjusts at least one parameter of the neural network to minimize the measured feature difference.

Abstract translation: 一种用于训练视觉识别计算机系统的神经网络的方法和系统，使用特征提取器提取图像或视频帧的至少一个特征; 使用在神经网络中提供的辅助输出近似图像或视频帧的至少一个特征; 并且利用辅助误差计算器测量提取的图像或视频帧的至少一个特征与图像或视频帧的近似的至少一个特征之间的特征差异。 该方法和系统的联合学习者调整神经网络的至少一个参数以最小化测量的特征差异。

公开(公告)号：US20120308122A1

公开(公告)日：2012-12-06

申请号：US13479256

申请日：2012-05-23

Applicant: Shenghuo Zhu ,  Chang Huang ,  Kai Yu

Inventor： Shenghuo Zhu ,  Chang Huang ,  Kai Yu

IPC: G06K9/62

CPC classification number: G06K9/6276 ,  G06K9/6201

Abstract: A nearest-neighbor-based distance metric learning process includes applying an exponential-based loss function to provide a smooth objective; and determining an objective and a gradient of both hinge-based and exponential-based loss function in a quadratic time of the number of instances using a computer.

Abstract translation: 基于最近邻的距离度量学习过程包括应用基于指数的损失函数来提供平滑的目标; 以及使用计算机在实例数量的二次时间中确定基于铰链和指数的损失函数的目标和梯度。

公开(公告)号：US08233711B2

公开(公告)日：2012-07-31

申请号：US12822424

申请日：2010-06-24

Applicant: Jinjun Wang ,  Fengjun Lv ,  Kai Yu

Inventor： Jinjun Wang ,  Fengjun Lv ,  Kai Yu

IPC: G06K9/00

CPC classification number: G06K9/4671 ,  G06K9/6223

Abstract: Systems and methods are disclosed for classifying an input image by detecting one or more feature points on the input image; extracting one or more descriptors from each feature point; applying a codebook to quantize each descriptor and generate code from each descriptor; applying spatial pyramid matching to generate histograms; and concatenating histograms from all sub-regions to generate a final representation of the image for classification.

Abstract translation: 公开了通过检测输入图像上的一个或多个特征点来对输入图像进行分类的系统和方法; 从每个特征点提取一个或多个描述符; 应用码本量化每个描述符并从每个描述符生成代码; 应用空间金字塔匹配来生成直方图; 以及从所有子区域连接直方图以生成用于分类的图像的最终表示。

公开(公告)号：US20110123965A1

公开(公告)日：2011-05-26

申请号：US12951135

申请日：2010-11-22

Applicant: Kai Yu

Inventor： Kai Yu

IPC: G09B19/04 ,  G10L21/00

CPC classification number: G09B19/04

Abstract: This invention relates to the field of tonal language speech signal processing. We describe a computer system for characterizing samples of a tonal language. These are analyzed to identify one or more vocal tract characterizing parameters of the user and synthesized speech data is generated by modifying a variation of fundamental frequency with time using a set of standard tones. The synthesized speech data represents the user speaking the tonal language with the modified fundamental frequency. Graphical feedback to guide the user can also be provided.

Abstract translation: 本发明涉及音调语言语音信号处理领域。 我们描述一个用于表征色调语言样本的计算机系统。 分析这些被分析以识别用户的一个或多个声道特性参数，并且通过使用一组标准音调修改基频的变化而产生合成语音数据。 合成的语音数据表示用户说出具有修改的基频的音调语言。 还可以提供用于指导用户的图形反馈。

公开(公告)号：US20100169747A1

公开(公告)日：2010-07-01

申请号：US12621604

申请日：2009-11-19

Applicant: Tsao-Tsen Chen ,  Per Ernstrom ,  Sten Ingemar Sjoberg ,  Kai Yu

Inventor： Tsao-Tsen Chen ,  Per Ernstrom ,  Sten Ingemar Sjoberg ,  Kai Yu

IPC: H03M13/23 ,  H03M13/07 ,  G06F11/10

CPC classification number: H03M13/235 ,  H03M13/036 ,  H03M13/413 ,  H03M13/635 ,  H03M13/6362 ,  H03M13/6375

Abstract: Method and apparatus for generating a set of generator polynomials for use as a tail biting convolutional code to operate on data transmitted over a channel comprises: (0) specifying a constraint and a low code rate for a tail biting convolutional code, where the low rate code is lower than 1/n (n being an integer greater than 4); (1) selecting valid combinations of generator polynomials to include in a pool of potential codes, each valid combination being a potential code of the low rate code; (2) determining first lines of a weight spectrum for each potential code in the pool and including potential codes of the pool having best first lines in a candidate set; (3) determining best codes of the candidate set based on the first L number of lines in the weight spectrum; (4) selecting an optimum code(s) from the best codes; and (5) configuring a circuit(s) of a data transceiver to implement the optimum code(s).

Abstract translation: 用于生成一组生成多项式的方法和装置，其用作尾部卷积卷积码以对通过信道发送的数据进行操作包括：（0）为尾部卷积卷积码指定约束和低码率，其中低速率 代码低于1 / n（n是大于4的整数）; （1）选择生成多项式的有效组合以包括在潜在代码池中，每个有效组合是低速率代码的潜在代码; （2）确定所述池中每个潜在代码的权重谱的第一行，并且包括在候选集合中具有最佳第一行的所述池的潜在代码; （3）基于权重谱中的第一L个行数确定候选集合的最佳代码; （4）从最佳代码中选择最佳代码; 和（5）配置数据收发器的电路以实现最佳代码。