公开(公告)号：US12136483B2

公开(公告)日：2024-11-05

申请号：US18627705

申请日：2024-04-05

Applicant: Lunit Inc.

Inventor： Jong Chan Park ,  Dong Geun Yoo ,  Ki Hyun You ,  Hyeon Seob Nam ,  Hyun Jae Lee ,  Sang Hyup Lee

IPC: G16H30/20 ,  G06T7/00 ,  G06V30/166 ,  G16H30/40

Abstract: The present disclosure relates to a medical image analysis method using a processor and a memory which are hardware. The method includes generating predicted second metadata for a medical image by using a prediction model, and determining a processing method of the medical image based on one of first metadata stored corresponding to the medical image and the second metadata.

公开(公告)号：US11978548B2

公开(公告)日：2024-05-07

申请号：US17426336

申请日：2020-05-22

Applicant: Lunit Inc.

Inventor： Jong Chan Park ,  Dong Geun Yoo ,  Ki Hyun You ,  Hyeon Seob Nam ,  Hyun Jae Lee ,  Sang Hyup Lee

IPC: G16H30/20 ,  G06F18/214 ,  G06N20/00 ,  G06T7/00 ,  G06T7/70 ,  G06V10/70 ,  G06V30/166 ,  G16H30/40

CPC classification number: G16H30/20 ,  G06F18/214 ,  G06N20/00 ,  G06T7/0012 ,  G06T7/70 ,  G06V10/70 ,  G06V30/166 ,  G16H30/40 ,  G06T2207/20081 ,  G06T2207/20084 ,  G06T2207/30004

Abstract: The present disclosure relates to a medical image analysis method using a processor and a memory which are hardware. The method includes generating predicted second metadata for a medical image by using a prediction model, and determining a processing method of the medical image based on one of first metadata stored corresponding to the medical image and the second metadata.

公开(公告)号：US10824944B2

公开(公告)日：2020-11-03

申请号：US16676694

申请日：2019-11-07

Applicant: Lunit Inc.

Inventor： Hyun Jae Lee

IPC: G06N3/08 ,  G06F17/18 ,  G06T3/00 ,  G06N20/00

Abstract: A method of recalibrating a feature data of each channel generated by a convolution layer of a convolution neural network is provided. According to some embodiments, since an affine transformation is applied to the feature data of each channel independently of the feature data of the other channel, the overall number of parameters defining the affine transformation is minimized. As a result, the amount of computations required in performing the feature data recalibration can be reduced.