公开(公告)号：US11694319B2

公开(公告)日：2023-07-04

申请号：US16938812

申请日：2020-07-24

Applicant: Samsung Display Co., Ltd.

Inventor： Yan Kang ,  Janghwan Lee ,  Shuhui Qu ,  Jinghua Yao ,  Sai MarapaReddy

IPC: G06T7/00 ,  G06N3/08 ,  G06N20/20 ,  G06V10/82 ,  G06V10/70 ,  G06F18/241 ,  G06F18/25 ,  G06N3/045 ,  G06V20/69

CPC classification number: G06T7/0004 ,  G06F18/241 ,  G06F18/251 ,  G06N3/045 ,  G06N3/08 ,  G06N20/20 ,  G06V10/70 ,  G06V10/82 ,  G06V20/69 ,  G06T2207/10056 ,  G06T2207/10116 ,  G06T2207/20081 ,  G06T2207/20084

Abstract: A system for manufacturing defect classification is presented. The system includes a first neural network receiving a first data as input and generating a first output, a second neural network receiving a second data as input and generating a second output, wherein first neural network and the second neural network are trained independently from each other, and a fusion neural network receiving the first output and the second output and generating a classification. The first data and the second data do not have to be aligned. Hence, the system and method of this disclosure allows various type of data that are collected during manufacturing to be used in defect classification.

公开(公告)号：US12136205B2

公开(公告)日：2024-11-05

申请号：US18320866

申请日：2023-05-19

Applicant: Samsung Display Co., Ltd.

Inventor： Yan Kang ,  Janghwan Lee ,  Shuhui Qu ,  Jinghua Yao ,  Sai MarapaReddy

IPC: G06T7/00 ,  G06F18/241 ,  G06F18/25 ,  G06N3/045 ,  G06N3/08 ,  G06N20/20 ,  G06V10/70 ,  G06V10/82 ,  G06V20/69

Abstract: A system for manufacturing defect classification is presented. The system includes a first neural network receiving a first data as input and generating a first output, a second neural network receiving a second data as input and generating a second output, wherein first neural network and the second neural network are trained independently from each other, and a fusion neural network receiving the first output and the second output and generating a classification. The first data and the second data do not have to be aligned. Hence, the system and method of this disclosure allows various type of data that are collected during manufacturing to be used in defect classification.

公开(公告)号：US20210319546A1

公开(公告)日：2021-10-14

申请号：US16938812

申请日：2020-07-24

Applicant: Samsung Display Co., Ltd.

Inventor： Yan Kang ,  Janghwan Lee ,  Shuhui Qu ,  Jinghua Yao ,  Sai MarapaReddy

IPC: G06T7/00 ,  G06K9/62 ,  G06N3/08 ,  G06N3/04 ,  G06N20/20

Abstract: A system for manufacturing defect classification is presented. The system includes a first neural network receiving a first data as input and generating a first output, a second neural network receiving a second data as input and generating a second output, wherein first neural network and the second neural network are trained independently from each other, and a fusion neural network receiving the first output and the second output and generating a classification. The first data and the second data do not have to be aligned. Hence, the system and method of this disclosure allows various type of data that are collected during manufacturing to be used in defect classification.

公开(公告)号：US20210319270A1

公开(公告)日：2021-10-14

申请号：US16938857

申请日：2020-07-24

Applicant: Samsung Display Co., Ltd.

Inventor： Shuhui Qu ,  Janghwan Lee ,  Yan Kang ,  Jinghua Yao ,  Sai MarapaReddy

IPC: G06K9/62 ,  G06N3/08 ,  G06N3/04

Abstract: A method and a system are presented for controlling a performance of a fusion model. The method includes obtaining a first set and a second set of candidate models for a first and second neural networks, respectively. Each of the first and second set of candidate models is pre-trained with a first source and a second source, respectively. For each possible pairing of one candidate model from the first neural network and one candidate model from the second neural network, a model distance Dm is determined. A subset of possible pairings of one first candidate model and one second candidate model is selected based on the model distance Dm between them. Using the subset of possible parings, the first neural network and the second neural network are combined to generate two branches for a fusion model neural network.

公开(公告)号：US20230316493A1

公开(公告)日：2023-10-05

申请号：US18320866

申请日：2023-05-19

Applicant: Samsung Display Co., Ltd.

Inventor： Yan Kang ,  Janghwan Lee ,  Shuhui Qu ,  Jinghua Yao ,  Sai MarapaReddy

IPC: G06T7/00 ,  G06N3/08 ,  G06N20/20 ,  G06V10/82 ,  G06V10/70 ,  G06F18/241 ,  G06F18/25 ,  G06N3/045 ,  G06V20/69

CPC classification number: G06T7/0004 ,  G06N3/08 ,  G06N20/20 ,  G06V10/82 ,  G06V10/70 ,  G06F18/241 ,  G06F18/251 ,  G06N3/045 ,  G06V20/69 ,  G06T2207/10056 ,  G06T2207/20084 ,  G06T2207/20081 ,  G06T2207/10116

Abstract: A system for manufacturing defect classification is presented. The system includes a first neural network receiving a first data as input and generating a first output, a second neural network receiving a second data as input and generating a second output, wherein first neural network and the second neural network are trained independently from each other, and a fusion neural network receiving the first output and the second output and generating a classification. The first data and the second data do not have to be aligned. Hence, the system and method of this disclosure allows various type of data that are collected during manufacturing to be used in defect classification.

公开(公告)号：US20210096530A1

公开(公告)日：2021-04-01

申请号：US16693101

申请日：2019-11-22

Applicant: Samsung Display Co., Ltd.

Inventor： Sai MarapaReddy ,  Shuhui Qu ,  Janghwan Lee

IPC: G05B19/406 ,  G06N3/08

Abstract: A system and method for classifying products manufactured via a manufacturing process. A processor receives an input dataset, and extracts features of the input dataset at two or more levels of abstraction. The processor combines the extracted features and provides the combined extracted features to a classifier. The classifier is trained based on the combined extracted features for learning a pattern of not-faulty products. The trained classifier is configured to receive data for a product to be classified, to output a prediction for the product based on the received data.

公开(公告)号：US20240242494A1

公开(公告)日：2024-07-18

申请号：US18623923

申请日：2024-04-01

Applicant: SAMSUNG DISPLAY CO., LTD.

Inventor： Shuhui Qu ,  Janghwan Lee ,  Yan Kang ,  Jinghua Yao ,  Sai MarapaReddy

IPC: G06V10/82 ,  G06F18/21 ,  G06F18/2113 ,  G06F18/22 ,  G06F18/25 ,  G06N3/045 ,  G06N3/08 ,  G06V10/764 ,  G06V10/80

CPC classification number: G06V10/82 ,  G06F18/2113 ,  G06F18/217 ,  G06F18/22 ,  G06F18/251 ,  G06N3/045 ,  G06N3/08 ,  G06V10/764 ,  G06V10/809 ,  G06V10/811

Abstract: A method and a system are presented for controlling a performance of a fusion model. The method includes obtaining a first set and a second set of candidate models for a first and second neural networks, respectively. Each of the first and second set of candidate models is pre-trained with a first source and a second source, respectively. For each possible pairing of one candidate model from the first neural network and one candidate model from the second neural network, a model distance Dm is determined. A subset of possible pairings of one first candidate model and one second candidate model is selected based on the model distance Dm between them. Using the subset of possible parings, the first neural network and the second neural network are combined to generate two branches for a fusion model neural network.

公开(公告)号：US11948347B2

公开(公告)日：2024-04-02

申请号：US16938857

申请日：2020-07-24

Applicant: Samsung Display Co., Ltd.

Inventor： Shuhui Qu ,  Janghwan Lee ,  Yan Kang ,  Jinghua Yao ,  Sai MarapaReddy

IPC: G06N3/08 ,  G06F18/21 ,  G06F18/2113 ,  G06F18/22 ,  G06F18/25 ,  G06N3/045 ,  G06V10/764 ,  G06V10/80 ,  G06V10/82

CPC classification number: G06V10/82 ,  G06F18/2113 ,  G06F18/217 ,  G06F18/22 ,  G06F18/251 ,  G06N3/045 ,  G06N3/08 ,  G06V10/764 ,  G06V10/809 ,  G06V10/811

Abstract: A method and a system are presented for controlling a performance of a fusion model. The method includes obtaining a first set and a second set of candidate models for a first and second neural networks, respectively. Each of the first and second set of candidate models is pre-trained with a first source and a second source, respectively. For each possible pairing of one candidate model from the first neural network and one candidate model from the second neural network, a model distance Dm is determined. A subset of possible pairings of one first candidate model and one second candidate model is selected based on the model distance Dm between them. Using the subset of possible parings, the first neural network and the second neural network are combined to generate two branches for a fusion model neural network.

公开(公告)号：US11435719B2

公开(公告)日：2022-09-06

申请号：US16693101

申请日：2019-11-22

Applicant: Samsung Display Co., Ltd.

Inventor： Sai MarapaReddy ,  Shuhui Qu ,  Janghwan Lee

IPC: G06F9/44 ,  G06F9/445 ,  G06F9/455 ,  H04L29/08 ,  G05B19/418 ,  G05B19/406 ,  G06N3/08

Abstract: A system and method for classifying products manufactured via a manufacturing process. A processor receives an input dataset, and extracts features of the input dataset at two or more levels of abstraction. The processor combines the extracted features and provides the combined extracted features to a classifier. The classifier is trained based on the combined extracted features for learning a pattern of not-faulty products. The trained classifier is configured to receive data for a product to be classified, to output a prediction for the product based on the received data.