US 12,242,964 B2
Method and apparatus for data efficient semantic segmentation
Qingfeng Liu, San Diego, CA (US); Mostafa El-Khamy, San Diego, CA (US); Jungwon Lee, San Diego, CA (US); and Behnam Babagholami Mohamadabadi, San Diego, CA (US)
Assigned to Samsung Electronics Co., Ltd, (KR)
Filed by Samsung Electronics Co., Ltd., Gyeonggi-do (KR)
Filed on Jun. 26, 2023, as Appl. No. 18/341,050.
Application 18/341,050 is a continuation of application No. 17/241,848, filed on Apr. 27, 2021, granted, now 11,687,780.
Claims priority of provisional application 63/047,438, filed on Jul. 2, 2020.
Prior Publication US 2023/0334318 A1, Oct. 19, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06K 9/62 (2022.01); G06F 18/2321 (2023.01); G06F 18/2323 (2023.01); G06F 18/2415 (2023.01); G06F 18/2431 (2023.01); G06F 18/25 (2023.01); G06N 3/04 (2023.01); G06N 3/045 (2023.01); G06N 3/08 (2023.01); G06N 5/022 (2023.01); G06N 5/025 (2023.01); G06V 10/772 (2022.01); G06V 10/774 (2022.01); G06V 10/80 (2022.01); G06V 30/19 (2022.01)
CPC G06N 3/08 (2013.01) [G06F 18/2321 (2023.01); G06F 18/2323 (2023.01); G06F 18/2415 (2023.01); G06F 18/2431 (2023.01); G06F 18/253 (2023.01); G06F 18/254 (2023.01); G06N 3/04 (2013.01); G06N 3/045 (2023.01); G06N 5/022 (2013.01); G06N 5/025 (2013.01); G06V 10/772 (2022.01); G06V 10/774 (2022.01); G06V 10/806 (2022.01); G06V 10/809 (2022.01); G06V 30/19107 (2022.01); G06V 30/1914 (2022.01); G06V 30/19147 (2022.01); G06V 30/1918 (2022.01)] 18 Claims
OG exemplary drawing
 
1. A method for training a neural network, comprising
receiving an input image;
generating an augmented image by applying at least one data augmentation method of a first data augmentation technique to the input image;
generating a mixed image from the input image and the augmented image; and
classifying the input image based on the mixed image and a resulting image from a second data augmentation technique,
wherein the first data augmentation technique and the second data augmentation technique are different data augmentation techniques, one being an Augmix data augmentation technique and another being a Mixup data augmentation technique.