Deep neural network
    1.
    发明授权

    公开(公告)号:US11521047B1

    公开(公告)日:2022-12-06

    申请号:US16391007

    申请日:2019-04-22

    Abstract: A hardware neural network system includes an input buffer for input neurons (Nbin), an output buffer for output neurons (Nbout), and a third buffer for synaptic weights (SB) connected to a Neural Functional Unit (NFU) and a control logic (CP) for performing synapses and neurons computations. The NFU pipelines a computation into stages, the stages including weight blocks (WB), an adder tree, and a non-linearity function.

    IRIS RECOGNITION USING FULLY CONVOLUTIONAL NETWORKS

    公开(公告)号:US20200160079A1

    公开(公告)日:2020-05-21

    申请号:US16685690

    申请日:2019-11-15

    Abstract: A method of accelerated iris recognition includes acquiring an image comprising at least an iris and a pupil, segmenting the iris and the pupil using a fully convolutional network (FCN) model, normalizing the segmented iris, encoding the normalized iris, the normalizing and encoding using a rubber sheet model and 1-D log Gabor filter, and masking the encoded iris.

    POWER MAPPING AND MODELING SYSTEM FOR INTEGRATED CIRCUITS
    4.
    发明申请
    POWER MAPPING AND MODELING SYSTEM FOR INTEGRATED CIRCUITS 审中-公开
    集成电路功率映射与建模系统

    公开(公告)号:US20160124443A1

    公开(公告)日:2016-05-05

    申请号:US14183384

    申请日:2014-02-18

    CPC classification number: G05D23/1917 G06F1/203

    Abstract: This invention relates in general to a system, apparatus, and method comprises one or more mapping and modeling systems used for power estimation, management, and improved efficiencies for the integrated circuit. In addition, the invention provides a feedback telemetry and actuation system, apparatus, and method for controlling the flow speed and temperature of coolant fluid over an integrated circuit

    Abstract translation: 本发明一般涉及系统,装置和方法,包括用于功率估计,管理和集成电路的改进效率的一个或多个映射和建模系统。 另外,本发明提供了一种用于控制集成电路上的冷却剂流体的流速和温度的反馈遥测和致动系统,装置和方法

    Iris recognition using fully convolutional networks

    公开(公告)号:US11113553B2

    公开(公告)日:2021-09-07

    申请号:US16685690

    申请日:2019-11-15

    Abstract: A method of accelerated iris recognition includes acquiring an image comprising at least an iris and a pupil, segmenting the iris and the pupil using a fully convolutional network (FCN) model, normalizing the segmented iris, encoding the normalized iris, the normalizing and encoding using a rubber sheet model and 1-D log Gabor filter, and masking the encoded iris.

    Power mapping and modeling system for integrated circuits

    公开(公告)号:US10175705B2

    公开(公告)日:2019-01-08

    申请号:US14183384

    申请日:2014-02-18

    Abstract: This invention relates to a power mapping and modeling system for integrated circuits. An automated process includes determining a floor plan of the integrated circuit, dividing the integrated circuit into blocks based on the floor plan, pumping oil through an inlet valve, over the integrated circuit, and out an outlet valve to create a simulated oil-based cooling system, injecting a plurality of power pulses across the integrated circuit, applying infrared image techniques to the integrated circuit to create measured thermal images for the simulated oil-based cooling system, setting up a finite-element model (FEM) to estimate entries in a first thermal-to-power modeling matrix for the simulated oil-based cooling system, determining and storing performance monitoring counter (PMC) measurements related to power consumptions of the blocks of the integrated circuit, reconstructing a power dissipation for the simulated oil-based cooling system by using a thermal map for the simulated oil-based cooling system system, using the power dissipation of the simulated oil-based cooling system, the FEM, and the PMC measurements to create a thermal map for a copper based cooling system, and creating a final thermal-to-power model for the integrated circuit from the thermal map for the model copper based cooling system.

Patent Agency Ranking