US 12,169,778 B2
Data object classification using an optimized neural network
Hayko Jochen Wilhelm Riemenschneider, Zurich (CH); Leonhard Markus Helminger, Zurich (CH); Christopher Richard Schroers, Uster (CH); and Abdelaziz Djelouah, Zurich (CH)
Assigned to Disney Enterprises, Inc., Burbank, CA (US); and ETH Zürich (EIDGENÖSSISCHE TECHNISCHE HOCHSCHULE ZÜRICH), Zurich (CH)
Filed by Disney Enterprises, Inc., Burbank, CA (US); and ETH Zürich (EIDGENÖSSISCHE TECHNISCHE HOCHSCHULE ZÜRICH), Zürich (CH)
Filed on May 4, 2023, as Appl. No. 18/143,297.
Application 18/143,297 is a continuation of application No. 17/946,907, filed on Sep. 16, 2022, granted, now 11,669,723.
Application 17/946,907 is a continuation of application No. 16/808,069, filed on Mar. 3, 2020, granted, now 11,475,280, issued on Oct. 18, 2022.
Claims priority of provisional application 62/936,125, filed on Nov. 15, 2019.
Prior Publication US 2023/0274138 A1, Aug. 31, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06N 3/08 (2023.01); G06F 18/214 (2023.01); G06F 18/24 (2023.01); G06N 3/048 (2023.01); G06V 10/75 (2022.01); G06V 10/764 (2022.01); G06V 10/774 (2022.01); G06V 10/776 (2022.01); G06V 10/82 (2022.01)
CPC G06N 3/08 (2013.01) [G06F 18/2155 (2023.01); G06F 18/24 (2023.01); G06N 3/048 (2023.01); G06V 10/75 (2022.01); G06V 10/764 (2022.01); G06V 10/7753 (2022.01); G06V 10/776 (2022.01); G06V 10/82 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
a hardware processor; and
a system memory storing a software code;
the hardware processor configured to execute the software code to:
obtain a plurality of real images;
composite the plurality of real images to form a montage of the plurality of real images;
identify a plurality of labels for association with the montage;
label the montage using one or more of the plurality of identified labels to generate the plurality of images in a training dataset, wherein noise is parametrically introduced into the training dataset, resulting in a subset of the plurality of images being purposely mislabeled;
determine, using a neural network (NN), a classification of the plurality of images in the training dataset; and
configure the NN based on a performance of the NN in determining the classification.