US 12,136,205 B2
Image-based defects identification and semi-supervised localization
Yan Kang, Sunnyvale, CA (US); Janghwan Lee, Pleasanton, CA (US); Shuhui Qu, Fremont, CA (US); Jinghua Yao, San Jose, CA (US); and Sai MarapaReddy, Newark, CA (US)
Assigned to Samsung Display Co., Ltd., Yongin-si (KR)
Filed by Samsung Display Co., Ltd., Yongin-si (KR)
Filed on May 19, 2023, as Appl. No. 18/320,866.
Application 18/320,866 is a continuation of application No. 16/938,812, filed on Jul. 24, 2020, granted, now 11,694,319.
Claims priority of provisional application 63/008,480, filed on Apr. 10, 2020.
Prior Publication US 2023/0316493 A1, Oct. 5, 2023
Int. Cl. G06T 7/00 (2017.01); G06F 18/241 (2023.01); G06F 18/25 (2023.01); G06N 3/045 (2023.01); G06N 3/08 (2023.01); G06N 20/20 (2019.01); G06V 10/70 (2022.01); G06V 10/82 (2022.01); G06V 20/69 (2022.01)
CPC G06T 7/0004 (2013.01) [G06F 18/241 (2023.01); G06F 18/251 (2023.01); G06N 3/045 (2023.01); G06N 3/08 (2013.01); G06N 20/20 (2019.01); G06V 10/70 (2022.01); G06V 10/82 (2022.01); G06V 20/69 (2022.01); G06T 2207/10056 (2013.01); G06T 2207/10116 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A classification system comprising:
one or more processors; and
a memory storing instructions which, when executed by the one or more processors, cause performance of:
receiving at a first neural network branch a first data as input and generating a first output, the first neural network branch being configured to apply a first attention map to the first data for identifying a first feature associated with the first output;
receiving at a second neural network branch a second data as input and generating a second output, the second neural network branch being configured to apply a second attention map to the second data for identifying a second feature associated with the second output;
receiving at a fusion neural network the first output and the second output; and
generating by the fusion neural network a classification, the fusion neural network comprising a convolutional block attention module having a spatial attention module and a channel attention module.