IP Library › Granted Patent US 12,548,294
Granted Patent B2
US 12,548,294 · App. 17/986,422 · Granted Feb 10, 2026

Determining a degree of realism of an artificially generated visual content

Inventors: Yair Adato (Kfar Ben Nun, IL); Bar Fingerman (Hertsliya, IL); Shahar Gad-Shriki (Tel Aviv, IL); Eyal Gutflaish (Beer Sheva, IL)
Assignee: BRIA ARTIFICIAL INTELLIGENCE LTD.
G06V10/764G06F18/214G06F18/24G06T3/40G06T5/50G06T7/0002G06T11/001G06T19/006G06V10/772G06V10/774G06T2207/20081
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,548,294
App. No.
17/986,422
Granted
Feb 10, 2026
Kind
B2
Abstract

Systems, methods and non-transitory computer readable media for determining a degree of realism of an artificially generated visual content are provided. Artificially generated visual contents including a particular artificially generated visual content may be accessed. Captured visual contents may be accessed. For each person of a plurality of persons, a mix of visual contents including at least one artificially generated visual content and at least one captured visual content may be presented to the person, where the mix includes the particular artificially generated visual content. A reaction to the presentation indicative of whether the person believes that visual contents are artificially generated may be received. A degree of realism of the particular artificially generated visual content may be determined based on the reactions.

Claims (56)

1 . A non-transitory computer readable medium containing instructions for causing at least one processor to perform operations for determining a degree of realism of an artificially generated visual content, the operations comprising:

accessing one or more artificially generated visual contents, the one or more artificially generated visual contents include a particular artificially generated visual content;

accessing one or more captured visual contents, each visual content of the one or more captured visual contents was captured using an image sensor from a real physical environment;

for each person of a plurality of persons, presenting to the person a mix of visual contents, the mix of visual contents includes at least one visual content of the one or more artificially generated visual contents and at least one visual content of the one or more captured visual contents, the mix of visual contents includes the particular artificially generated visual content;

for each person of the plurality of persons, receiving from the person a reaction to the presentation, wherein for each visual content of a group of at least one of the mix of visual contents, the reaction is indicative of whether the person believes that the visual content is an artificially generated visual content; and

determining a degree of realism of the particular artificially generated visual content based on the reactions,

wherein the operations further comprise analyzing the particular artificially generated visual content to determine at least one reason for the determined degree of realism of the particular artificially generated visual content,

wherein the operations further comprise presenting the determined at least one reason to an individual.

2 . The non-transitory computer readable medium of claim 1 , wherein the degree of realism of the particular artificially generated visual content is a relative degree with respect to at least one other artificially generated visual content.

3 . The non-transitory computer readable medium of claim 1 , wherein the degree of realism of the particular artificially generated visual content is based on a comparison of the reactions to the particular artificially generated visual content and at least one reaction to another artificially generated visual content of the one or more artificially generated visual contents.

4 . The non-transitory computer readable medium of claim 1 , wherein the degree of realism of the particular artificially generated visual content is based on a comparison of the reactions to the particular artificially generated visual content and at least one reaction to a visual content of the one or more captured visual contents.

5 . The non-transitory computer readable medium of claim 1 , wherein the mix of visual contents presented to one person of the plurality of persons differs from the mix of visual contents presented to another person of the plurality of persons.

6 . The non-transitory computer readable medium of claim 1 , wherein each artificially generated visual content of the one or more artificially generated visual contents is generated using a conditional generative adversarial network with a different input condition.

7 . The non-transitory computer readable medium of claim 1 , wherein the operations further comprise determining whether to delete the particular artificially generated visual content based on the determined degree of realism of the particular artificially generated visual content.

8 . The non-transitory computer readable medium of claim 1 , wherein the operations further comprise determining whether to use the particular artificially generated visual content for a particular usage based on the determined degree of realism of the particular artificially generated visual content.

9 . The non-transitory computer readable medium of claim 1 , wherein the operations further comprise using the at least one reason to generate a new artificially generated visual content.

10 . The non-transitory computer readable medium of claim 1 , wherein the operations further comprise:

calculating a convolution of at least part of the particular artificially generated visual content to thereby obtain a result value of the calculated convolution of the at least part of the particular artificially generated visual content; and

using the result value of the calculated convolution of the at least part of the particular artificially generated visual content to determine the at least one reason for the determined degree of realism of the particular artificially generated visual content.

11 . The non-transitory computer readable medium of claim 1 , wherein the one or more artificially generated visual contents includes a specific artificially generated visual content, the specific artificially generated visual content differs from the particular artificially generated visual content, and wherein the operations further comprise:

determining a degree of realism of the specific artificially generated visual content based on at least one reaction to the specific artificially generated visual content;

comparing data associated with the specific artificially generated visual content to data associated with the particular artificially generated visual content; and

based on a result of the comparison, determining the at least one reason for the determined degree of realism of the particular artificially generated visual content.

12 . The non-transitory computer readable medium of claim 11 , wherein each one of the one or more artificially generated visual contents is a result of applying a function to a different input value, the data associated with the particular artificially generated visual content is the input value associated with the particular artificially generated visual content, and the data associated with the specific artificially generated visual content is the input value associated with the specific artificially generated visual content.

13 . The non-transitory computer readable medium of claim 11 , wherein the operations further comprise:

analyzing the particular artificially generated visual content to determine the data associated with the particular artificially generated visual content; and

analyzing the specific artificially generated visual content to determine the data associated with the specific artificially generated visual content.

14 . The non-transitory computer readable medium of claim 11 , wherein the operations further comprise:

calculating a convolution of at least part of the particular artificially generated visual content to thereby determine the data associated with the particular artificially generated visual content; and

calculating a convolution of at least part of the specific artificially generated visual content to thereby determine the data associated with the specific artificially generated visual content.

15 . The non-transitory computer readable medium of claim 1 , wherein the one or more artificially generated visual contents includes a specific artificially generated visual content, the specific artificially generated visual content differs from the particular artificially generated visual content, and wherein the operations further comprise:

determining a degree of realism of the specific artificially generated visual content based on at least one reaction to the specific artificially generated visual content;

determining a particular visual content, in an area of the visual content, associated with the particular artificially generated visual content;

determining a specific visual content, in an area of the visual content, associated with the specific artificially generated visual content;

using the determined degree of realism of the particular artificially generated visual content, the determined degree of realism of the specific artificially generated visual content, the particular visual content and the specific visual content to determine a new visual content in the area of the new visual content, the new visual content differs from the particular visual content and the specific visual content; and

using the new visual content in the area of the visual content to generate a new artificially generated visual content.

16 . The non-transitory computer readable medium of claim 1 , wherein the operations further comprise:

for each visual content of the one or more artificially generated visual contents, analyzing the visual content using a machine learning model to determine a level of certainty that the visual content is realistic; and

selecting the particular artificially generated visual content from the one or more artificially generated visual contents based on the levels of certainties.

17 . A system for determining a degree of realism of an artificially generated visual content, the system comprising:

at least one processor configured to perform the operations of:

accessing one or more artificially generated visual contents, the one or more artificially generated visual contents include a particular artificially generated visual content;

accessing one or more captured visual contents, each visual content of the one or more captured visual contents was captured using an image sensor from a real physical environment;

for each person of a plurality of persons, presenting to the person a mix of visual contents, the mix of visual contents includes at least one visual content of the one or more artificially generated visual contents and at least one visual content of the one or more captured visual contents, the mix of visual contents includes the particular artificially generated visual content;

for each person of the plurality of persons, receiving from the person a reaction to the presentation, wherein for each visual content of a group of at least one of the mix of visual contents, the reaction is indicative of whether the person believes that the visual content is an artificially generated visual content; and

determining a degree of realism of the particular artificially generated visual content based on the reactions,

wherein the operations further comprise analyzing the particular artificially generated visual content to determine at least one reason for the determined degree of realism of the particular artificially generated visual content,

wherein the operations further comprise presenting the determined at least one reason to an individual.

18 . A method for determining a degree of realism of an artificially generated visual content, the method comprising:

accessing one or more artificially generated visual contents, the one or more artificially generated visual contents include a particular artificially generated visual content;

accessing one or more captured visual contents, each visual content of the one or more captured visual contents was captured using an image sensor from a real physical environment;

for each person of a plurality of persons, presenting to the person a mix of visual contents, the mix of visual contents includes at least one visual content of the one or more artificially generated visual contents and at least one visual content of the one or more captured visual contents, the mix of visual contents includes the particular artificially generated visual content;

for each person of the plurality of persons, receiving from the person a reaction to the presentation, wherein for each visual content of a group of at least one of the mix of visual contents, the reaction is indicative of whether the person believes that the visual content is an artificially generated visual content; and

determining a degree of realism of the particular artificially generated visual content based on the reactions,

wherein the operations further comprise analyzing the particular artificially generated visual content to determine at least one reason for the determined degree of realism of the particular artificially generated visual content,

wherein the operations further comprise presenting the determined at least one reason to an individual.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Nov 26, 2022
From: ADATO, YAIR; FINGERMAN, BAR; GAD-SHRIKI, SHAHAR; GUTFLAISH, EYAL
To: BRIA ARTIFICIAL INTELLIGENCE LTD
Reel/Frame 061883/0156 →
Continuity (3)
Continuation PCTIL2022051189 · Nov 9, 2022
Provisional Application 63279111 · Nov 14, 2021
Related Publication 20230153973A1 · May 18, 2023
References Cited (17)
US 20130195361A1 · Deng · 2013 [cited by examiner]
US 20190080205A1 · Kaufhold et al. · 2019 [cited by applicant]
US 20200066025A1 · Peebler · 2020 [cited by examiner]
US 20200320322A1 · Liang et al. · 2020 [cited by applicant]
US 20200356591A1 · Yada · 2020 [cited by examiner]
US 20210019541A1 · Wang · 2021 [cited by examiner]
US 20210073958A1 · Masuda · 2021 [cited by applicant]
US 20210165932A1 · Mohan · 2021 [cited by examiner]
US 20210350604A1 · Pejsa et al. · 2021 [cited by applicant]
US 20220180602A1 · Hao et al. · 2022 [cited by applicant]
Groh M, Epstein Z, Firestone C, Picard R. Deepfake detection by human crowds, machines, and machine-informed crowds. Proc Natl Acad Sci U S A. Oct. 27, 2021, (Year: 2021). [cited by examiner]
Groh M, Epstein Z, Firestone C, Picard R. Deepfake detection by human crowds, machines, and machine-informed crowds. Proc Natl Acad Sci U S A. Oct. 2, 20217, (Year: 2021). [cited by examiner]
PCT International Search Report for International Application No. PCT/IL2022/051189, mailed Feb. 9, 2023, 5pp. [cited by applicant]
PCT Written Opinion for International Application No. PCT/IL2022/051189, mailed Feb. 9, 2023, 7pp. [cited by applicant]
Chen, Z., Jiang, R., Duke, B., Zhao, H., Aarabi, P. (2022). Exploring Gradient-Based Multi-directional Controls in GANs. In: Avidan, S., Brostow, G., Cisse, M., Farinella, G. M., Hassner, T. (eds) Computer Vision—ECCV 2… [cited by applicant]
Lee, Seunghun, Sunghyun Cho, and Sunghoon Im. “Dranet: Disentangling representation and adaptation networks for unsupervised cross-domain adaptation.” Proceedings of the IEEE/CVF conference on computer vision and patter… [cited by applicant]
S. Jiang, Z. Tao and Y. Fu, “Geometrically Editable Face Image Translation With Adversarial Networks,” in IEEE Transactions on Image Processing, vol. 30, pp. 2771-2783, 2021. doi: 10.1109/TIP.2021.3052084. [cited by applicant]