公开(公告)号：US20220343470A1

公开(公告)日：2022-10-27

申请号：US17859435

申请日：2022-07-07

Applicant: Adobe Inc.

Inventor： Ionut Mironica ,  Oscar Bolaños ,  Andreea Birhala

IPC: G06T5/00 ,  G06T7/00 ,  G06K9/62 ,  G06T7/33 ,  G06T5/50

Abstract: In implementations of correcting dust and scratch artifacts in digital images, an artifact correction system receives a digital image that depicts a scene and includes a dust or scratch artifact. The artifact correction system generates, with a generator of a generative adversarial neural network (GAN), a feature map from the digital image that represents features of the dust or scratch artifact and features of the scene. A training system can train the generator adversarially to reduce visibility of dust and scratch artifacts in digital images against a discriminator, and train the discriminator to distinguish between reconstructed digital images generated by the generator and real-world digital images. The artifact correction system generates, from the feature map and with the generator, a reconstructed digital image that depicts the scene of the digital image and reduces visibility of the dust or scratch artifact of the digital image.

公开(公告)号：US11763430B2

公开(公告)日：2023-09-19

申请号：US17859435

申请日：2022-07-07

Applicant: Adobe Inc.

Inventor： Ionut Mironica ,  Oscar Bolaños ,  Andreea Birhala

IPC: G06T7/33 ,  G06T5/50 ,  G06F18/213 ,  G06T5/00 ,  G06T7/00 ,  G06V10/82 ,  G06V10/30

CPC classification number: G06T5/005 ,  G06F18/213 ,  G06T5/50 ,  G06T7/0002 ,  G06T7/337 ,  G06V10/30 ,  G06V10/82 ,  G06T2200/24 ,  G06T2207/20081 ,  G06T2207/20084 ,  G06T2207/30168

Abstract: In implementations of correcting dust and scratch artifacts in digital images, an artifact correction system receives a digital image that depicts a scene and includes a dust or scratch artifact. The artifact correction system generates, with a generator of a generative adversarial neural network (GAN), a feature map from the digital image that represents features of the dust or scratch artifact and features of the scene. A training system can train the generator adversarially to reduce visibility of dust and scratch artifacts in digital images against a discriminator, and train the discriminator to distinguish between reconstructed digital images generated by the generator and real-world digital images. The artifact correction system generates, from the feature map and with the generator, a reconstructed digital image that depicts the scene of the digital image and reduces visibility of the dust or scratch artifact of the digital image.