知搜-全球专利检索

  1. Login
  2. Sign Up

Search Results

828 records (took 0.027 seconds)

    Image segmentation method and system using GAN architecture
    171.
    发明授权
    Image segmentation method and system using GAN architecture 有权

    公开(公告)号：US12086712B2

    公开(公告)日：2024-09-10

    申请号：US17512463

    申请日：2021-10-27

    Applicant: Korea Electronics Technology Institute

    Inventor： Choong Sang Cho ,  Young Han Lee

    IPC: G06T7/11 ,  G06F18/21 ,  G06N3/045 ,  G06N3/08 ,  G06V10/75

    CPC classification number: G06N3/08 ,  G06F18/2163 ,  G06F18/217 ,  G06N3/045 ,  G06T7/11 ,  G06V10/751 ,  G06T2207/20081 ,  G06T2207/20084

    Abstract: There are provided a method and a system for image segmentation utilizing a GAN architecture. A method for training an image segmentation network according to an embodiment includes: inputting an image to a first network which is trained to output a region segmentation result regarding an input image, and generating a region segmentation result; and inputting the region segmentation result generated at the generation step and a ground truth (GT) to a second network, and acquiring a discrimination result, the second network being trained to discriminate inputted region segmentation results as a result generated by the first network and a GT, respectively; and training the first network and the second network by using the discrimination result. Accordingly, region segmentation performance of a semantic segmentation network regarding various images can be enhanced, and a very small image region can be exactly segmented.

    METHOD AND SYSTEM FOR DETERMINING MODEL OUTPUT DIRECTION ON BASIS OF HEAT DISSIPATION CHARACTERISTIC ANALYSIS FOR STABILIZING OUTPUT OF METAL 3D PRINTING
    173.
    发明公开
    METHOD AND SYSTEM FOR DETERMINING MODEL OUTPUT DIRECTION ON BASIS OF HEAT DISSIPATION CHARACTERISTIC ANALYSIS FOR STABILIZING OUTPUT OF METAL 3D PRINTING 审中-公开

    公开(公告)号：US20240256729A1

    公开(公告)日：2024-08-01

    申请号：US18565797

    申请日：2022-05-25

    Applicant: Korea Electronics Technology Institute

    Inventor： Hwa Seon SHIN ,  Sung Hwan CHUN ,  Hye In LEE ,  Jae Ho SHIN ,  Sung Hun PARK

    IPC: G06F30/17

    CPC classification number: G06F30/17

    Abstract: Provided are a method and a system for determining a model output direction on the basis of heat dissipation characteristic analysis for stabilizing output of metal 3D printing. A method for determining a model output direction on the basis of heat dissipation characteristic analysis according to an embodiment of the present invention comprises: a shape characteristic parameter deriving step in which a model output direction determining system calculates model shape characteristic data according to the output direction of a model that changes in a metal 3D printing output process; a heat data change amount collecting step in which the model output direction determining system collects simulation results regarding residual heat data of the model every time the output direction changes; a heat data change amount analyzing step in which the model output direction determining system analyzes heat dissipation characteristics inside the model on the basis of the collected simulation results, thereby calculating heat flatness after heat dissipation with regard to each output direction; and an output direction determining step in which the model output direction determining system recommends output directions in descending order of heat flatness remaining in the model on the basis of the result of calculating heat flatness after heat dissipation with regard to each output direction in the heat data change amount analyzing step. Accordingly, the amount of remaining heat that changes depending on the output direction is measured through simulation, thereby analyzing heat dissipation characteristics of the model, and output directions appropriate for output stabilization are derived and are proposed to process workers, thereby contributing to output stabilization of metal 3D printing.

    Accelerated processing method for deep learning based-panoptic segmentation using a RPN skip based on complexity
    174.
    发明授权
    Accelerated processing method for deep learning based-panoptic segmentation using a RPN skip based on complexity 有权

    公开(公告)号：US12051230B2

    公开(公告)日：2024-07-30

    申请号：US17623067

    申请日：2020-11-25

    Applicant: Korea Electronics Technology Institute

    Inventor： Kyoung Won Min ,  Ganzorig Gankhuyag ,  Haeng Seon Son ,  Seon Young Lee ,  Young Bo Shim

    IPC: G06V10/26 ,  G06V10/50

    CPC classification number: G06V10/267 ,  G06V10/50

    Abstract: Provided is a deep learning-based panoptic segmentation accelerated processing technique using a complexity-based RPN skip method. An image segmentation system includes: a first processing unit configured to extract dynamic objects in an instance segmentation method by using an extracted feature; a calculation unit configured to control to skip some areas of the feature extracted at the network by the first processing unit, on the basis of complexity of the input image; a second processing unit configured to extract static objects in a semantic segmentation method by using the feature extracted at the network; and a fusion unit configured to fuse a result of extracting by the first processing unit and a result of extracting by the second processing unit. Accordingly, the panoptic segmentation method can be easily performed even in an embedded environment by reducing complexity for panoptic segmentation processing by reducing a calculation burden.

    METHOD FOR DATA INTERWORKING BETWEEN APPLICATION AND IOT DEVICE AND MIDDLEWARE DEVICE FOR PERFORMING THE METHOD
    177.
    发明公开
    METHOD FOR DATA INTERWORKING BETWEEN APPLICATION AND IOT DEVICE AND MIDDLEWARE DEVICE FOR PERFORMING THE METHOD 审中-公开

    公开(公告)号：US20240223662A1

    公开(公告)日：2024-07-04

    申请号：US18397739

    申请日：2023-12-27

    Applicant: Korea Electronics Technology Institute

    Inventor： Tai Gil KWON

    IPC: H04L67/125 ,  G16Y10/75 ,  G16Y40/35

    CPC classification number: H04L67/125 ,  G16Y10/75 ,  G16Y40/35

    Abstract: Proposed are a method for data interworking between an application and an IoT device, and a middleware device for performing the method. The method may include generating matching information between an IoT device and a network protocol used by the IoT device based on IoT device profile information, and analyzing a request message received from an application device to extract IoT device information of a request target. The method may also include generating a request packet conforming to a network protocol used by a target IoT device based on the request message, the IoT device information of the target request, and the matching information, and transmitting the request packet to the target IoT device. The method may further include converting a response packet transmitted by the target IoT device in response to the request packet into a response message, and transmitting the response message to the application device.

    MULTI-CHANNEL MICRO PERIPHERAL NERVE SENSOR FOR NEUROSIGNAL MEASUREMENT
    178.
    发明公开
    MULTI-CHANNEL MICRO PERIPHERAL NERVE SENSOR FOR NEUROSIGNAL MEASUREMENT 审中-公开

    公开(公告)号：US20240215886A1

    公开(公告)日：2024-07-04

    申请号：US18147345

    申请日：2022-12-28

    Applicant: KOREA ELECTRONICS TECHNOLOGY INSTITUTE

    Inventor： Kyusik SHIN ,  Cheolung CHA ,  Yunjae WON

    IPC: A61B5/24

    CPC classification number: A61B5/24 ,  A61B2562/028 ,  A61B2562/12 ,  A61B2562/164 ,  A61B2562/166

    Abstract: A multi-channel micro peripheral nerve sensor for measuring a neurosignal with low noise is proposed. The multi-channel micro peripheral nerve sensor may include a substrate, an analog-digital (AD) converter, and a signal acquisition unit. The substrate may be combined with each cut surface of a severed nerve and include channels into which nerve cells are inserted. The AD converter may be combined with the substrate, be electrically connected to the channels, receive analog neurosignals from the nerve cells, and convert the analog neurosignals into digital signals. The signal acquisition unit may be electrically connected to the AD converter and acquire the digital signals from the AD converter.

Patent Agency Ranking