System and method for liveness verification

Patent metadata

Field	Value
Publication number	US-12573239-B2
Application number	US-202217881930-A
Country	US
Kind code	B2
Filing date	Aug 5, 2022
Priority date	Aug 5, 2022
Publication date	Mar 10, 2026
Grant date	Mar 10, 2026

Abstract

A method for determining authenticity of an object in real-time is disclosed. The method being performed by a processor includes receiving image data of the object. The image data includes both the captured image showing a visual representation of the object as well as depth data describing the relative distance between different objects in the image and the camera. This depth information is analyzed in order to determine a liveness of the object. The analysis may include verifying a sufficient change in depth between the subject and a background as well as comparing an edge of a foreground object to an expected edge or an edge determined from the image. Based on this analysis, a determination can be made as to whether the image is of a live (e.g., real) subject.

First claim

What is claimed is:

1 . A method for user verification in a computing device, comprising:
activating a camera associated with the computing device in response to receiving a request for user verification; prompting the user to provide an image of the user's face using the camera; receiving at least one video frame from a video feed, the video frame including depth data for pixels across the video frame; analyzing the depth data associated with the video frame to distinguish a subject in a foreground from a background; performing a preliminary liveness verification on the video frame, the preliminary liveness verification comprising determining that the video frame includes a live image of the subject based on the analyzing, the preliminary liveness verification including determining that a depth of the subject in the foreground differs from a depth of the background by a predetermined threshold, and including determining that there is a first region of the video frame having a first depth that is similar in shape to an expected shape of the subject; capturing the image that includes the subject from the video frame in response to the preliminary liveness verification; performing edge detection analysis on the depth data to detect a depth edge associated with a foreground object corresponding to the subject; performing edge detection analysis on the captured image to detect an image edge associated with the subject; and performing a subsequent liveness verification of the captured image based on a comparison of the depth edge to the image edge.

2 . The method of claim 1 , wherein a second region of the video frame has a second depth that differs from the first depth of the first region.

3 . The method of claim 1 , further comprising:
prior to the verifying:
determining that the video frame does not include a live subject based on the analyzing; and
prompting the user to take a photo of a live subject in response to the determining.

4 . The method of claim 3 , wherein the determining determines that there is no region of the video frame that has a depth similarity that corresponds to a shape of a face.

5 . The method of claim 1 , further comprising:
prompting the user to provide an image of a printed face on a photographic identification document; and comparing a captured face within the captured image to the printed face.

6 . The method of claim 5 , further comprising:
permitting access to a secure system in response to the captured face substantially matching the printed face; and denying access to the secure system in response to the captured face not substantially matching the printed face.

7 . A computing device for providing access to a secure system, the computing device comprising:
an image capture device; a communication interface configured to communicate with a server; and a user interface configured to interface with a backend server and controlled by one or more processors configured to perform operations comprising:
receiving an access request from a user;
prompting the user, in response to the access request, to provide a live image of a face of the user;
receiving at least one video frame of a video feed from the user, the at least one video frame including depth information for pixels across the at least one video frame;
performing a preliminary liveness verification on the at least one video frame, the preliminary liveness verification comprising determining, based on the depth information, that the at least one video frame includes a live image of a face, the preliminary liveness verification including verifying that a depth of an object in the at least one video frame differs from a depth of a background by a predetermined threshold, and verifying that there is a first region of the at least one video frame having a first depth that is similar in shape to a shape of a face;
capturing a video frame from among the at least one video frame as the live image in response to the preliminary liveness verification;
performing edge detection analysis on the depth information to detect a depth edge associated with a foreground object corresponding to the face of the user;
performing edge detection analysis on the captured video frame to detect an image edge associated with the face of the user; and
performing a subsequent liveness verification of the captured video frame based on a comparison of the depth edge to the image edge.

8 . The computing device of claim 7 , wherein the determining includes identifying a first region of the at least one video frame that has a substantially similar depth compared to a second region of the at least one video frame.

9 . The computing device of claim 7 , wherein the one or more processors is further configured to prompt the user to provide a live image of a printed face on a photographic identification document.

10 . The computing device of claim 9 , wherein the one or more processors is further configured to:
perform facial analysis of each of the printed face and the live face; and compare the printed face to the live face.

11 . The computing device of claim 10 , wherein the one or more processors is further configured to:
grant the access request in response to the printed face substantially matching the live face; and deny the access request in response to the printed face not substantially matching the live face.

12 . A method for verifying a user, comprising:
receiving an access request from the user; prompting, in response to the access request, the user to provide a live image of the user's face; receiving a video frame of a video feed from the user via an image capture device, the video frame including depth information for pixels across the video frame; performing a preliminary liveness verification on the video frame, the preliminary liveness verification including first determining, based on the depth information, that the video frame is a live image and second determining, based on a pixel analysis of the video frame, that the video frame includes a face, the second determining including verifying that a depth of an object in the video frame differs from a depth of a background by a predetermined threshold, and including verifying that there is a first region of the video frame having a first depth that is similar in shape to a shape of a face; capturing the video frame as the live image in response to the preliminary liveness verification; performing edge detection on the depth information to detect a depth edge associated with a foreground object corresponding to the user's face; performing edge detection analysis on the captured video frame to detect an image edge associated with the user's face; and performing a subsequent liveness verification of the captured video frame based on a comparison of the depth edge to the image edge.

13 . The method of claim 12 , further comprising activating the image capture device associated with a user device, wherein the image capture device is capable of detecting depth data.

14 . The method of claim 12 , wherein the first determining includes:
identifying the first region of the depth information that has substantially similar depths.

15 . The method of claim 12 , further comprising:
prior to the first determining:
third determining, based on the depth information, that the video frame is not a live image; and
prompting the user to provide a live image.

16 . The method of claim 15 , wherein the third determining includes failing to identify any region of the depth information that has substantially si

Assignees

• Capital One Services Llc

Inventors

• Patil Swapnil
• Zearfoss Patrick
• Pribble Jason

Classifications

• G06V40/172 Primary

  Classification, e.g. identification · CPC title

• G06T7/50

  Depth or shape recovery · CPC title

• G06T2207/10028

  Range image; Depth image; 3D point clouds · CPC title

• G06V40/161

  Detection; Localisation; Normalisation · CPC title

• G06V40/45 Primary

  Detection of the body part being alive · CPC title