Passive image depth sensing for object verification based on chromatic differentiation

Patent metadata

Field	Value
Publication number	US-11741748-B2
Application number	US-202117373794-A
Country	US
Kind code	B2
Filing date	Jul 13, 2021
Priority date	Oct 14, 2020
Publication date	Aug 29, 2023
Grant date	Aug 29, 2023

Abstract

Techniques are described for passive three-dimensional (3D) image sensing based on chromatic differentiation for use in object verification. For example, multiple sub-images correspond to 3D feature regions of an object. The sub-images can be analyzed to obtain respective sets of feature depth measurements (e.g., depth, textural signatures, etc.) based on multiple differentiated chromatic components of raw image sensor data captured from the object. A verification signal can be output as a function of comparing the respective sets of feature depth measurements from the plurality of characteristic sub-images to previously stored feature depth expectations, such that the verification signal indicates whether an identity of the object is verified and/or whether the object is a spoof.

First claim

What is claimed is:

1. A method for passive image depth sensing for object verification, the method comprising:
capturing, using an image sensor, raw image data of an object characterized by three-dimensional (3D) feature regions, each 3D feature region associated with a respective expected depth and/or a respective expected textural signature;
identifying a plurality of characteristic sub-images from the raw image data, each identified at least by mapping a corresponding one of the 3D feature regions to the raw image data;
obtaining, from the raw image data for each characteristic sub-image, a respective first chromatic response from imaging the object with the image sensor, and a respective second chromatic response from imaging the object with the image sensor;
computing, based on the respective first chromatic response and the respective second chromatic response for each characteristic sub-image, a respective computed depth and/or a respective computed textural signature; and
outputting a verification signal for the object based on comparing the respective computed depth of each characteristic sub-image with the respective expected depth of the corresponding 3D feature region, and/or on comparing the respective computed textural signature of each characteristic sub-image with the respective expected textural signature of the corresponding 3D feature region.

2. The method of claim 1 , wherein:
each respective first chromatic response corresponds to a first wavelength of light reflected off of the object and detected by the image sensor;
each respective second chromatic response corresponds to a second wavelength of the light reflected off of the object and detected by the image sensor; and
the first wavelength is distinguishable from the second wavelength by the image sensor.

3. The method of claim 1 , wherein the computing comprises, for each characteristic sub-image:
computing a ratio between a first magnitude of chromatic response (MCR) derived from the respective first chromatic response and a second MCR derived from the respective second chromatic response; and
determining at least the respective computed depth based on the ratio.

4. The method of claim 1 , wherein:
the object is associated with a stored generalized object model that defines, for each of the plurality of 3D feature regions, the respective expected depth and/or the respective expected textural signature; and
the verification signal indicates whether the object is a spoof.

5. The method of claim 1 , wherein:
each 3D feature region is pre-associated with the respective expected depth and/or the respective expected textural signature obtained from the object and stored during a registration phase of operation prior to the capturing; and
the verification signal indicates whether the object is a spoof.

6. The method of claim 5 , wherein the verification signal further indicates whether an identity of the user is verified.

7. The method of claim 1 , wherein:
the object comprises a plurality of anatomical features of a user, each anatomical feature having respective stored feature data indicating at least a location, a shape, and/or a size of the anatomical feature, the stored feature data obtained and stored during a registration phase of operation prior to the capturing;
the computing comprises extracting feature measurements corresponding to at least a portion of the stored feature data for at least some of the plurality of anatomical features; and
the outputting comprises computing a biometric verification of the user based at least on comparing the feature measurements to the stored feature data, such that the verification signal further indicates whether an identity of the user is verified based on the biometric verification.

8. The method of claim 7 , wherein:
the extracting the feature measurements comprises tracing brightness distribution characteristics of the respective first chromatic response and/or the respective second chromatic response for each characteristic sub-image; and
the computing the biometric verification comprises computing the respective computed textural signature based on the brightness distribution characteristics.

9. The method of claim 7 , further comprising:
computing estimated distance data for the object based at least on comparing the feature measurements to the stored feature data,
wherein the computing the respective computed depth and/or the respective computed textural signature is based at least partially on the computing the estimated distance data.

10. The method of claim 7 , wherein the computing the biometric verification is performed at least partially in parallel with the computing the respective computed depth and/or the respective computed textural signature.

11. The method of claim 1 , wherein:
the computing the respective computed depth and/or the respective computed textural signature is based on image brightness statistic parameters, and the computing comprises normalizing the image brightness statistic parameters across the respective first chromatic response and the respective second chromatic response for at least one of the characteristic sub-images.

12. The method of claim 1 , wherein:
each 3D feature region is further associated with a respective expected location on the object; and
the identifying comprises processing the raw image data to identify respective estimated locations for at least two of the 3D feature regions based on the respective expected locations, and mapping each corresponding one of the at least two of the 3D feature regions to the raw image data based on the respective estimated locations.

13. The method of claim 1 , wherein:
each 3D feature region associated with the respective expected depth and the respective expected textural signature;
the computing comprises computing, based on the respective first chromatic response and the respective second chromatic response for each characteristic sub-image, the respective computed depth and the respective computed textural signature; and
the outputting is based on comparing the respective computed depth of each characteristic sub-image with the respective expected depth of the corresponding 3D feature region, and on comparing the respective computed textural signature of each characteristic sub-image with the respective expected textural signature of the corresponding 3D feature region.

14. The method of claim 1 , wherein the object comprises at least a portion of a human face.

15. A passive image depth sensing system for object verification, the system comprising:
a lens assembly to receive light reflected off of an object and to focus chromatic components of the received light in accordance with respective focal lengths, the object characterized by three-dimensional (3D) feature regions, each associated with a respective expected depth and/or a respective expected textural signature;
an image sensor in optical communication with the lens assembly and comprising a plurality of photodetector elements comprising first photodetector elements to produce first chromatic responses to a first chromatic component of the received light, and second photodetector elements to produce second chromatic responses to a second chromatic component of the received light; and
a processor configured to:
identify, from raw image data of the object captured by the image sensor, a plurality of characteristic sub-images by mapping corresponding ones of the 3D feature regions to the raw image data;
obtain, from the raw image data for each characteristic sub-image, a respective first chromatic response and a r

Assignees

• Shenzhen Goodix Tech Co Ltd

Inventors

• He Yi
• Pi Bo

Classifications

• G06V40/166 Primary

  using acquisition arrangements · CPC title

• G06T7/40

  Analysis of texture (depth or shape recovery from texture G06T7/529) · CPC title

• G06T7/55

  from multiple images · CPC title

• G06V10/751

  Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching · CPC title

• G06V20/64

  Three-dimensional [3D] objects · CPC title