Methods and systems for detecting user liveness

Patent metadata

Field	Value
Publication number	US-10685251-B2
Application number	US-201815919312-A
Country	US
Kind code	B2
Filing date	Mar 13, 2018
Priority date	Mar 13, 2018
Publication date	Jun 16, 2020
Grant date	Jun 16, 2020

Abstract

A method for detecting user liveness is provided that includes selecting first and second images from a sequence of images. The first and second images are captured under different illumination conditions. The method further includes locating feature calculation windows in corresponding positions on the first and second images. Each window includes a first area and a second area. Moreover, the method includes calculating, by a computing device, a feature value for each window position based on pixels, within the windows located at the position, from the first and second images. Furthermore, the method includes calculating a feature vector from the feature values, calculating a confidence score from the feature vector, and determining the sequence of images includes images of a live user when the confidence score is equal to or greater than the threshold score.

First claim

What is claimed is:

1. A method for detecting user liveness comprising the steps of:
selecting first and second images from a sequence of images, the first and second images being captured under different illumination conditions;
selecting feature calculation window types from a table;
locating each selected window type in corresponding positions on the first and second images, each window including a first area and a second area;
calculating, for each window type on the first image, an average pixel value from the pixels included in the first area;
calculating, for each window type on the second image, an average pixel value from the pixels included in the second area;
calculating a feature value for each window type as the difference between the average pixel values for the respective window type;
calculating a feature vector from the feature values;
calculating a confidence score from the feature vector; and
determining the sequence of images includes images of a live user when the confidence score is equal to or greater than the threshold score.

2. A method for detecting user liveness in accordance with claim 1 , further comprising the step of randomly illuminating, by the computing device, a biometric modality while the computing device captures the sequence of images.

3. A method for detecting user liveness in accordance with claim 2 , said illuminating step comprising illuminating the biometric modality in accordance with a signal having on and off-states of random duration.

4. A method for detecting user liveness in accordance with claim 1 , said calculating a feature value step further comprising calculating the feature values based on the content of a color channel of the first and second images.

5. A method for detecting user liveness in accordance with claim 1 , said selecting first and second images step comprising:
selecting a first group of images captured under a first illumination condition and a second group of images captured under a second illumination condition; and
combining the image data of the first group of images to create the first image and combining the image data of the second group of images to create the second image.

6. A computing device for detecting user liveness comprising:
a processor; and
a memory configured to store data, said computing device being associated with a network and said memory being in communication with said processor and having instructions stored thereon which, when read and executed by said processor, cause said computing device to:
select first and second images from a sequence of images, the first and second images being captured under different illumination conditions;
select feature calculation window types from a table;
locate each selected window type in corresponding positions on the first and second images, each window including a first area and a second area;
calculate, for each window type on the first image, an average pixel value from the pixels included in the first area;
calculate, for each window type on the second image, an average pixel value from the pixels included in the second area;
calculate a feature value for each window type as the difference between the average pixel values for the respective window type;
calculate a feature vector from the feature values;
calculate a confidence score from the feature vector; and
determine the sequence of images includes images of a live user when the confidence score is equal to or greater than the threshold score.

7. A computing device for detecting user liveness in accordance with claim 6 , wherein the instructions when read and executed by said processor, further cause said computing device to randomly illuminate a biometric modality while capturing the sequence of images.

8. A computing device for detecting user liveness in accordance with claim 7 , wherein the instructions when read and executed by said processor, further cause said computing device to illuminate the biometric modality in accordance with a signal having on and off-states of random duration.

9. A computing device for detecting user liveness in accordance with claim 6 , wherein the instructions when read and executed by said processor, further cause said computing device to calculate the feature values based on the content of a color channel of the first and second images.

10. A computing device for detecting user liveness in accordance with claim 6 , wherein the instructions when read and executed by said processor, further cause said computing device to:
select a first group of images captured under a first illumination condition and a second group of images captured under a second illumination condition; and
combine the image data of the first group of images to create the first image and combine the image data of the second group of images to create the second image.

11. A non-transitory computer-readable recording medium included in a computing device having a computer program recorded thereon for detecting user liveness, the computer program being comprised of instructions, which when read and executed by the computing device, cause the computing device to:
select first and second images from a sequence of images, the first and second images being captured under different illumination conditions;
select feature calculation window types from a table;
locate each selected window type in corresponding positions on the first and second images, each window including a first area and a second area;
calculate, for each window type on the first image, an average pixel value from the pixels included in the first area;
calculate, for each window type on the second image, an average pixel value from the pixels included in the second area;
calculate a feature value for each window type as the difference between the average pixel values for the respective window type;
calculate a feature vector from the feature values;
calculate a confidence score from the feature vector; and
determine the sequence of images includes images of a live user when the confidence score is equal to or greater than the threshold score.

12. A non-transitory computer-readable recording medium in accordance with claim 11 wherein the instructions when read and executed by said computing device, further cause said computing device to randomly illuminate a biometric modality while capturing the sequence of images.

13. A non-transitory computer-readable recording medium in accordance with claim 12 wherein the instructions when read and executed by said computing device, further cause said computing device to illuminate the biometric modality in accordance with a signal having on and off-states of random duration.

14. A non-transitory computer-readable recording medium in accordance with claim 11 wherein the instructions when read and executed by said computing device, further cause said computing device to calculate the feature values based on the content of a color channel of the first and second images.

15. A non-transitory computer-readable recording medium in accordance with claim 11 wherein the instructions when read and executed by said computing device, further cause said computing device to:
select a first group of images captured under a first illumination condition and a second group of images captured under a second illumination condition; and
combine the image data of the first group of images to create the first image and combine the image data of the second group of images to create the second image.

Assignees

• Daon Holdings Ltd

Inventors

• Ionita Mircea

Classifications

• G06V40/45

  Detection of the body part being alive · CPC title

• G06V40/168 Primary

  Feature extraction; Face representation · CPC title

• G06V10/141

  Control of illumination · CPC title

• G06V40/167

  using comparisons between temporally consecutive images · CPC title

• G06V40/172

  Classification, e.g. identification · CPC title