IP Library Granted Patent US 12676024
Granted Patent B1
US 12676024 · App. 17/849,196 · Granted Jul 7, 2026

Systems and methods for passive-subject liveness verification in digital media

Inventor: Gareth Genner (Atlanta, GA)
Assignee: T Stamp Inc.
G06V40/176G06T7/251G06V40/103G06V40/165G06V40/45G06V40/179
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 12676024
App. No.
17/849,196
Granted
Jul 7, 2026
Kind
B1
Abstract

The present system may be deployed in various scenarios to provide proof of liveness (also referred to herein as “liveness verification”) of an image without interaction of the subject of the image. The liveness verification process generally comprises imperative analysis and dynamic analysis of the image, after which liveness of the image may be determined.

Claims (76)

1 . A method of verifying liveness of a subject shown in an image file, comprising the steps of:

selecting one or more particular liveness analyses to be performed in relation to at least a portion of an image file;

preprocessing the image file for the one or more particular liveness analyses, thereby producing at least one preprocessed image, wherein the preprocessed image purports to show a subject's face, wherein preprocessing the image file comprises:

selecting one or more particular preprocessing methodologies based at least in part on the one or more particular liveness analyses, each of the one or more particular preprocessing methodologies corresponding to at least one particular liveness analysis of the one or more particular liveness analyses; and

performing the one or more particular preprocessing methodologies to thereby produce the at least one preprocessed image;

performing the one or more particular liveness analyses on the at least one preprocessed image by at least:

performing an imperative analysis on the at least one preprocessed image; and

performing a dynamic analysis on the at least one preprocessed image; and

determining liveness of the subject based on a result of the imperative analysis and a result of the dynamic analysis.

2 . The method of claim 1 , wherein performing the dynamic analysis on the preprocessed image comprises analyzing the preprocessed image to determine that at least one of the following regarding the preprocessed image is within a predetermined range: diffusion speed, contrast, blurriness, specularity, texture, frequency, or gamut differences.

3 . The method of claim 2 , wherein the subject is a user of an electronic computing device attempting to access a third party system via the electronic computing device, and wherein the method further comprises the step of, in response to verifying liveness of the subject based on the result of the imperative analysis and/or the result of the dynamic analysis, permitting the user to access the third party system via the electronic computing device.

4 . The method of claim 3 , wherein the third party system comprises an email client, a banking system, a software system, a SaaS platform, a website, a smartphone, a computer, a watch, an ATM, a vehicle, or a building.

5 . The method of claim 3 , wherein the electronic computing device comprises a camera with a flash configured to capture the image file, and wherein prior to determining specularity, the electronic computing device receives a signal for engaging the flash for illuminating the subject's face.

6 . The method of claim 2 , wherein performing the dynamic analysis on the preprocessed image comprises performing a texture analysis on the preprocessed image by at least:

converting the preprocessed image to generate a grayscale version of the preprocessed image;

cropping the grayscale version of the preprocessed image to thereby provide a cropped grayscale version of the preprocessed image comprising only the subject's face; and

generating a resized image from the cropped grayscale version of the preprocessed image, wherein the resized image is larger than the cropped grayscale version of the preprocessed image.

7 . The method of claim 6 , wherein performing the dynamic analysis on the preprocessed image comprises performing a frequency analysis by at least computing a fast fourier transform on the resized image.

8 . The method of claim 1 , wherein performing the imperative analysis comprises performing contour detection on the preprocessed image for determining that the preprocessed image does not comprise a frame surrounding the subject's face.

9 . The method of claim 8 , wherein performing contour detection comprises:

extracting one or more contours from the preprocessed image, wherein the one or more contours generally surround the subject's face; and

determining that each of the one or more contours does not comprise at least four corners with angles approximately 90 degrees.

10 . The method of claim 1 , wherein preprocessing the image file comprises converting the image file into a different format, a different resolution, or a different image size.

11 . The method of claim 1 , wherein:

the image file comprises a video; and

preprocessing the image file comprises:

extracting a plurality of extracted images from the video; and

selecting at least one image from the plurality of extracted images.

12 . The method of claim 11 , wherein selecting the at least one image from the plurality of extracted images comprises randomly selecting one or more extracted images from the plurality of extracted images to thereby provide the at least one image.

13 . The method of claim 11 , wherein selecting the at least one image from the plurality of extracted images comprises applying a sampling algorithm and/or one or more predefined rules.

14 . A system for verifying liveness of a subject shown in an image, comprising:

a server operatively configured to:

receive an image that purports to show a subject's face;

select one or more particular liveness analyses to be performed in relation to at least a portion of the image;

preprocess the image for the one or more particular liveness analyses, thereby producing at least one preprocessed image, wherein preprocessing the image comprises:

selecting one or more particular preprocessing methodologies based at least in part on the one or more particular liveness analyses, each of the one or more particular preprocessing methodologies corresponding to at least one particular liveness analysis of the one or more particular liveness analyses; and

performing the one or more particular preprocessing methodologies to thereby produce the at least one preprocessed image;

perform an imperative analysis on the at least one preprocessed image;

perform a dynamic analysis on the at least one preprocessed image; and

determine liveness of the subject based on a result of the imperative analysis and a result of the dynamic analysis.

15 . The system of claim 14 , wherein performing the dynamic analysis on the at least one preprocessed image comprises analyzing the at least one preprocessed image to determine that at least one of the following regarding the at least one preprocessed image is within a predetermined range: diffusion speed, contrast, blurriness, specularity, texture, frequency, or gamut differences.

16 . The system of claim 15 , wherein the subject is a user of an electronic computing device attempting to access a third party system via the electronic computing device, and wherein the server is further operatively configured to, in response to verifying liveness of the subject based on the result of the imperative analysis and/or the result of the dynamic analysis, permitting the user to access the third party system via the electronic computing device.

17 . The system of claim 16 , wherein the third party system comprises an email client, a banking system, a software system, a SaaS platform, a website, a smartphone, a computer, a watch, an ATM, a vehicle, or a building.

18 . The system of claim 16 , wherein the electronic computing device comprises a camera with a flash configured to capture the image, and wherein prior to determining specularity, the electronic computing device receives a signal for engaging the flash for illuminating the subject's face.

19 . The system of claim 15 , wherein performing the dynamic analysis on the at least one preprocessed image comprises performing a texture analysis on the at least one preprocessed image by at least:

converting the at least one preprocessed image to generate a grayscale version of the at least one preprocessed image;

cropping the grayscale version of the at least one preprocessed image to thereby provide a cropped grayscale version of the at least one preprocessed image comprising only the subject's face; and

generating a resized image from the cropped grayscale version of the at least one preprocessed image, wherein the resized image is larger than the cropped grayscale version of the at least one preprocessed image.

20 . The system of claim 19 , wherein performing the dynamic analysis on the at least one preprocessed image comprises performing a frequency analysis by at least computing a fast fourier transform on the resized image.

21 . The system of claim 14 , wherein performing the imperative analysis comprises performing contour detection on the at least one preprocessed image for determining that the at least one preprocessed image does not comprise a frame surrounding the subject's face.

22 . The system of claim 21 , wherein performing contour detection comprises:

extracting one or more contours from the at least one preprocessed image, wherein the one or more contours generally surround the subject's face; and

determining that each of the one or more contours does not comprise at least four corners with angles approximately 90 degrees.

23 . The system of claim 14 , wherein performing the dynamic analysis on the at least one preprocessed image comprises performing one or more machine learning-based image processing methodologies on the at least one preprocessed image, the one or more machine learning-based image processing methodologies being operatively configured to identify patterns indicative of liveness in the preprocessed image based on a processing of spoofed images and non-spoofed images in a training dataset.

24 . A tangible, non-transitory, computer-readable medium comprising instructions encoded therein, wherein the instructions, when executed by one or more processors, cause the one or more processors to:

select one or more liveness analyses to be performed in relation to at least one image of an image file based on one or more considerations, wherein the image file purports to show a subject's face;

preprocess the at least one image to provide at least one preprocessed image, wherein preprocessing the at least one image comprises:

selecting one or more particular preprocessing methodologies based at least in part on the one or more liveness analyses to be performed, each of the one or more particular preprocessing methodologies corresponding to at least one liveness analysis of the one or more liveness analyses; and

performing the one or more particular preprocessing methodologies to thereby provide the at least one preprocessed image;

perform an imperative analysis on the at least one preprocessed image;

perform a dynamic analysis on the at least one preprocessed image; and

determine liveness of the subject based on a result of the imperative analysis and a result of the dynamic analysis.

25 . The tangible, non-transitory, computer-readable medium of claim 24 , wherein:

performing the dynamic analysis on the at least one preprocessed image comprises analyzing the at least one preprocessed image to determine that at least one of the following regarding the at least one preprocessed image is within a predetermined range: diffusion speed, contrast, blurriness, specularity, texture, frequency, or gamut differences; and

performing the imperative analysis comprises performing contour detection on the at least one preprocessed image for determining that the at least one preprocessed image does not comprise a frame surrounding the subject's face.

26 . A method comprising:

selecting one or more liveness analyses to be performed in relation to at least one image of an image file based on one or more considerations, wherein the image file purports to show a subject's face;

preprocessing the at least one image to provide at least one preprocessed image, wherein preprocessing the at least one image comprises:

selecting one or more particular preprocessing methodologies based at least in part on the one or more liveness analyses to be performed, each of the one or more particular preprocessing methodologies corresponding to at least one liveness analysis of the one or more liveness analyses; and

performing the one or more particular preprocessing methodologies to thereby provide the at least one preprocessed image;

performing the one or more liveness analyses by performing at least one of:

an imperative analysis on the at least one preprocessed image; or

a dynamic analysis on the at least one preprocessed image; and

determining liveness of the subject based on the one or more liveness analyses.

27 . The method of claim 26 , wherein the one or more considerations comprises an analysis of the image file.

28 . The method of claim 26 , wherein the one or more considerations comprises a required level of security regarding liveness or a required level of certainty regarding liveness.