US 12,293,487 B2
Electronic device performing image inpainting and method of operating the same
Naejin Kong, Suwon-si (KR); Harshith Goka, Suwon-si (KR); Kiwoong Park, Suwon-si (KR); and Roman Suvorov, Suwon-si (KR)
Assigned to Samsung Electronics Co., Ltd., Suwon-si (KR)
Filed by Samsung Electronics Co., Ltd., Suwon-si (KR)
Filed on May 17, 2022, as Appl. No. 17/746,429.
Application 17/746,429 is a continuation of application No. PCT/KR2022/006289, filed on May 3, 2022.
Claims priority of application No. 10-2021-0064335 (KR), filed on May 18, 2021; and application No. 10-2021-0101523 (KR), filed on Aug. 2, 2021.
Prior Publication US 2022/0375050 A1, Nov. 24, 2022
Int. Cl. G06T 5/50 (2006.01); G06T 7/90 (2017.01); G06V 10/26 (2022.01); G06V 10/774 (2022.01)
CPC G06T 5/50 (2013.01) [G06T 7/90 (2017.01); G06V 10/267 (2022.01); G06V 10/774 (2022.01); G06T 2207/20224 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method of performing image inpainting by an electronic device to reconstruct a region in an input image to be reconstructed, the method comprising:
determining a missing region in an original image;
generating an input image to be reconstructed from the original image, based on the determined missing region, wherein the missing region is a portion of the input image having a smaller size than the input image;
obtaining a mask image indicating the missing region;
determining whether to extract a structural feature of the missing region, based on an attribute of the missing region;
obtaining, based on the determination to extract the structural feature, structure vectors by applying the input image and the mask image to a first model to extract the structural feature of structures in the input image, wherein the structure vectors include structure vectors of the missing region, and represent the structural feature, with each structure vector including one or more lines and one or more junctions; and
obtaining an inpainted image in which the missing region in the input image is reconstructed by applying the input image, the mask image, and a structure vector image converted from the structure vectors to a second model for reconstructing the input image.