US 12,141,952 B2
Exposure defects classification of images using a neural network
Akhilesh Kumar, San Jose, CA (US); Zhe Lin, Fremont, WA (US); and William Lawrence Marino, Hockessin, DE (US)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by Adobe Inc., San Jose, CA (US)
Filed on Sep. 30, 2022, as Appl. No. 17/957,639.
Application 17/957,639 is a continuation of application No. 16/888,473, filed on May 29, 2020, granted, now 11,494,886.
Prior Publication US 2023/0024955 A1, Jan. 26, 2023
Int. Cl. G06N 3/08 (2023.01); G06N 5/04 (2023.01); G06T 7/00 (2017.01)
CPC G06T 7/0002 (2013.01) [G06N 3/08 (2013.01); G06N 5/04 (2013.01); G06T 2207/20081 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
executing a first classification model on each image of a collection of images, the first classification model outputting a binary classification indicating whether each image includes an exposure defect;
executing a second classification model on a subset of defective images from the collection of images classified by the first classification model to have the exposure defects, the second classification model outputting a level of exposure for each image of the subset; and
executing an action using the subset of the collection of images.
 
10. One or more computer storage media storing computer-useable instructions that, when executed by one or more computing devices, cause the one or more computing devices to perform operations comprising:
classifying, using a first classification model, each image of a collection of images with a binary classification indicating whether the image includes an exposure defect;
classifying, using a second classification model, each image of a subset of defective images having the exposure defect based on a level of exposure; and
executing an action based on the subset of the collection of images.
 
16. A system comprising one or more processors and memory configured to provide computer program instructions to the one or more processors, the computer program instructions comprising:
a defect detector configured to:
execute a first classification model configured to classify each image of a collection of images with a binary classification that indicates whether the image includes an exposure defect;
execute a second classification model configured to classify defective images of a subset of images indicated as including the exposure defect according to a level of exposure for each image of the subset; and
executing an action based on the subset of the collection of images.