System and method for organizing and processing feature based data structures

    公开(公告)号:US10127219B2

    公开(公告)日:2018-11-13

    申请号:US15374479

    申请日:2016-12-09

    IPC分类号: G06F17/27 G06F17/28

    摘要: A method for organizing and processing feature based data structures that can be used in linguistic spell checking and auto-correction, comprising: splitting an original dictionary into sub-dictionaries based on different values of a common feature such as high frequency words; receiving an input text that contains errors; determining a sub-dictionary selection feature from the input human-readable text; selecting the sub-dictionary based on the determined sub-dictionary selection feature; executing a first matching in the selected sub-dictionary, wherein a match is found if a similarity between the characters, words, or phrases in proximity of the errors in the input text and a character, word, or phrase in the sub-dictionary is above a threshold; if a unique match is found, the result is returned as an output to correct the errors; otherwise, executing a second matching with a raised threshold, and repeating the second matching until a unique match is found.

    Multi-stage hybrid analog-to-digital converter

    公开(公告)号:US10103742B1

    公开(公告)日:2018-10-16

    申请号:US15878375

    申请日:2018-01-23

    摘要: A hybrid Analog-to-Digital Converter (ADC) has multiple stages. A first stage and a final stage each use a Successive-Approximation Register (SAR) ADC to generate the Most-Significant-Bits (MSBs) and the Least-Significant-Bits (LSBs) over successive internal cycles. Middle stage(s) use a faster flash ADC with multiple comparators in parallel to generate the middle binary bits, which are then re-converted by a Digital-to-Analog Converter (DAC) and subtracted from the stage's input analog voltage to generate a difference that is amplified by a residual amplifier that outputs an amplified voltage to the next stage. The first stage also has this multiplying DAC structure to convert the MSBs to an amplified voltage to the first of the middle stages. Finally, digital error correction logic removes redundant binary bits between stages. Initial and final SAR stages of 4 and 8 bits with a 4-bit middle stage provide a hybrid ADC of 14-bit precision.

    METHODS AND APPARATUS FOR ON-CHIP DERIVATIVE SPECTROSCOPY

    公开(公告)号:US20180209899A1

    公开(公告)日:2018-07-26

    申请号:US15874956

    申请日:2018-01-19

    IPC分类号: G01N21/27 G01N21/25

    摘要: A derivative spectroscopy system for achieving a tunable resolution of 2 nm or less in resolving spectral components of an input optical signal is provided so as to estimate derivative spectra of the input optical signal based on the resolved spectral components. In the system, a first dispersive-element structure spectrally decomposes the input optical signal into subband signals. A second dispersive-element structure receives part or all of the subband signals and spectrally decomposes the received subband signals to plural spectral components. A material having a temperature-variant refractive index is used to build the second dispersive-element structure, enabling a shift of center wavelength of each spectral component as small as 2 nm of less upon changing a temperature of the second dispersive-element structure. By obtaining three spectral-component sets obtained at three different predetermined temperatures with the center-wavelength shift of 2 nm or less, first- and second-order derivative spectra are obtained with good accuracy.

    Method and System for Enhancing Accuracy in Location and Proximity Determination

    公开(公告)号:US20180098301A1

    公开(公告)日:2018-04-05

    申请号:US15285493

    申请日:2016-10-05

    摘要: A system comprising beacon-generating clusters for enabling a mobile computing device to perform location or proximity determination is provided. Each cluster comprises a primary station for broadcasting a primary beacon, and an offset-assisted station, positioned from the primary station by an offset distance along a direction, for broadcasting a secondary beacon. The offset distance and the direction are unique for all the clusters, causing a jittering area, within which the mobile computing device is unable to determine which one of adjacent primary stations is closer to from detecting the primary beacons, to be reduced by further utilizing the secondary beacons provided by the clusters. Fluctuation in received signal strength indicator (RSSI) can be reduced by having each station, primary or secondary, installed with multiple co-located broadcasting units for utilizing space and time diversity to reduce RSSI fluctuation. In another option, multiple offset-assisted stations are used for each cluster.

    Timing offset estimation in an OFDM-based system by SINR measurement

    公开(公告)号:US09860861B2

    公开(公告)日:2018-01-02

    申请号:US15133237

    申请日:2016-04-20

    IPC分类号: H04W56/00

    CPC分类号: H04W56/001 H04W56/0035

    摘要: A method for timing synchronization of an OFDM signal is useful for a sniffing base station (BS) to establish BS synchronization with another BS in a mobile communication system. The method comprises estimating a timing offset of the signal from a reference sampling instant. In estimating the timing offset, first determine a maximum detection range of an estimable timing offset estimated solely by an observed phase difference between two pre-selected pilot symbols in the signal. Then the timing offset is determined as an integer multiple of the maximum detection range plus a residual timing offset. The multiplying integer is determined from a set of candidate integers. According to a candidate integer under consideration, a portion of an OFDM-signal sample sequence is masked out and a resultant signal to interference plus noise ratio (SINR) is computed. The multiplying integer is determined by identifying the candidate integer having the greatest SINR.

    Method and System for Bit-Depth Reduction in Artificial Neural Networks

    公开(公告)号:US20170270408A1

    公开(公告)日:2017-09-21

    申请号:US15415810

    申请日:2017-01-25

    IPC分类号: G06N3/08 G06N3/04

    CPC分类号: G06N3/08 G06N3/04 G06N3/063

    摘要: A bit-depth optimization engine reduces the hardware cost of a neural network. When training data is applied to a neural network during training routines, accuracy cost and hardware costs are generated. A hardware complexity cost generator generates costs for weights near bit-depth steps where the number of binary bits required to represent a weight decreases, such as from 2N to 2N−1, where one less binary bit is required. Gradients are generated from costs for each weight, and weights near bit-depth steps are easily selected since they have a large gradient, while weights far away from a bit-depth step have near-zero gradients. The selected weights are reduced during optimization. Over many cycles of optimization, a low-bit-depth neural network is generated that uses fewer binary bits per weight, resulting in lower hardware costs when the low-bit-depth neural network is manufactured on an Application-Specific Integrated Circuit (ASIC).

    Multi-location learning-activity state management for distance education

    公开(公告)号:US09679340B2

    公开(公告)日:2017-06-13

    申请号:US14549554

    申请日:2014-11-21

    IPC分类号: G09B3/00 G06Q50/20 G06Q10/00

    CPC分类号: G06Q50/20 G06Q10/00

    摘要: A method for delivering distance education for real clients, each either a teacher client or a student client, at multiple teaching sites and a system using the method are provided. A multi-layer arrangement is used to arrange computing servers to be one first-level server and one or more second-level servers each communicable with the first-level server. Each server serves real clients located in a pre-determined site. Furthermore, each server sets up a virtual client to execute the teacher client's activity command so as to locally generate application-specific data, which are then stored in this server. When a communication link to a student client is re-established after an activity interruption, resynchronization of the student client's learning-activity state with the teacher client's is regained based on the locally-generated application-specific data stored in the server that serves the student client without a need to burden other servers.

    Wavelet-based Image Decolorization and Enhancement
    90.
    发明申请
    Wavelet-based Image Decolorization and Enhancement 有权
    基于小波的图像脱色和增强

    公开(公告)号:US20160379340A1

    公开(公告)日:2016-12-29

    申请号:US14746852

    申请日:2015-06-23

    发明人: Cheung Ming LAI

    摘要: The present invention relates to image processing. More particularly, the present invention provides methods for efficient image decolorization and color image enhancement. The methods of the present invention comprise decolorization in frequency domain, adaptive brightness control for an enhanced grayscale image and color image enhancement. The present invention is able to improve sharpness and fine details in both enhanced grayscale and color images.

    摘要翻译: 本发明涉及图像处理。 更具体地说,本发明提供了有效的图像脱色和彩色图像增强的方法。 本发明的方法包括频域中的脱色,用于增强灰度图像的自适应亮度控制和彩色图像增强。 本发明能够提高增强灰阶和彩色图像的清晰度和细节。