US 12,387,459 B2
System and method for image de-identification to humans while remaining recognizable by machines
Eliran Kuta, Tel Aviv (IL); Sella Blondheim, Tel Aviv (IL); Gil Perry, Tel Aviv (IL); Or Gorodissky, Tel Aviv (IL); Matan Ben Yosef, Korazim (IL); and Yoav Hacohen, Jerusalem (IL)
Assigned to DE-IDENTIFICATION LTD., Tel Aviv (IL)
Appl. No. 17/785,434
Filed by DE-IDENTIFICATION LTD., Tel Aviv (IL)
PCT Filed Dec. 15, 2020, PCT No. PCT/IL2020/051287
§ 371(c)(1), (2) Date Jun. 15, 2022,
PCT Pub. No. WO2021/124321, PCT Pub. Date Jun. 24, 2021.
Claims priority of provisional application 62/948,370, filed on Dec. 16, 2019.
Prior Publication US 2023/0027309 A1, Jan. 26, 2023
Int. Cl. G06K 9/00 (2022.01); G06V 10/74 (2022.01); G06V 10/774 (2022.01); G06V 10/82 (2022.01); G06V 40/16 (2022.01)
CPC G06V 10/761 (2022.01) [G06V 10/774 (2022.01); G06V 10/82 (2022.01); G06V 40/16 (2022.01)] 18 Claims
OG exemplary drawing
 
1. A method of image de-identification, the method comprising:
receiving an input image of a human face;
iteratively modifying the input image to produce an output image of the human face by modifying style parameters while maintaining structure features of the input image until:
a score indicating how unlikely a human is to associate the input image with the output image reaches a first threshold indicating that the human would not be able to associate the input image with the output image, and
a score indicating an ability of a computerized unit to associate the input image with the output image reaches a second threshold indicating that the computerized unit would be able to associate the input image with the output image; and
providing the output image,
wherein modifying the input image is performed by a deep neural network, wherein the style parameters comprise high level features of the neural network, and wherein the neural network modifies the input image by modifying the high level features while maintaining the structure features,
wherein modifying the style parameters while maintaining the structure features is carried out with application of a loss function custom character:
custom character=λ0∥Θ0(Xsrc)−Θ0(XD)∥−λ1∥Θ1(Xsrc)−Θ1(XD)∥ . . . −λk∥Θk(Xsrc)−Θk(XD)∥
where k is a scale index, λ is a weight, Θ is a parameter of the neural network, Xsrc is the input image, XD is the output image, and Θk are the high level features.