US 12,272,024 B2
Training method of image processing model, image processing method, apparatus, and device
Xiaozhong Ji, Shenzhen (CN); Yun Cao, Shenzhen (CN); Ying Tai, Shenzhen (CN); Chengjie Wang, Shenzhen (CN); and Jilin Li, Shenzhen (CN)
Assigned to TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED, Shenzhen (CN)
Filed by TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED, Shenzhen (CN)
Filed on May 6, 2022, as Appl. No. 17/739,053.
Application 17/739,053 is a continuation of application No. PCT/CN2021/086576, filed on Apr. 12, 2021.
Claims priority of application No. 202010419181.7 (CN), filed on May 18, 2020.
Prior Publication US 2022/0261965 A1, Aug. 18, 2022
Int. Cl. G06T 5/00 (2024.01); G06N 20/20 (2019.01); G06T 5/50 (2006.01); G06V 10/774 (2022.01)
CPC G06T 5/00 (2013.01) [G06N 20/20 (2019.01); G06T 5/50 (2013.01); G06V 10/774 (2022.01); G06T 2207/20081 (2013.01)] 20 Claims
OG exemplary drawing
 
1. An image processing method, performed by an image processing device, comprising:
obtaining a sample image set, the sample image set comprising a first number of sample images;
constructing an image feature set based on the sample image set, the image feature set comprising an image feature extracted from each of the sample images in the sample image set, and the image feature comprises a blurriness feature and a noise feature;
obtaining a training image set, the training image set comprising a second number of training images;
constructing multiple training image pairs based on the training image set and the image feature set, each of the training image pairs comprising a training target image and a training input image, comprising: obtaining the training input image corresponding to the training target image based on the training target image and the image feature set, a resolution of the training target image being higher than a resolution of the training input image; and
training an image processing model based on the multiple training image pairs.