Spoof detection using illumination sequence randomization
Technology described herein includes a method that includes obtaining a set of color-coded sequences, each of which includes a sequence of colors. Each color-coded sequence has auto-correlation properties characterized by a merit factor larger than a first predetermined threshold, and cross-correlation properties among the color-coded sequences characterized by a demerit factor lower than a second predetermined threshold. A color-coded sequence is randomly selected from the set of color-coded sequences. A subject is illuminated in accordance with the sequence of colors in the selected color-coded sequence. A sequence of images of the subject are captured, and are temporally synchronized with illumination by the color-coded sequence. A filtered response image is generated from the sequence of images by a matched filtering process. Based on the filtered response image, it is determined that the subject is an alternative representation of a live person. In response, access to a secure system is prevented.
1. A computer-implemented method comprising:
obtaining a set of multiple color-coded sequences, wherein each color-coded sequence comprises a sequence of colors, wherein each color-coded sequence has auto-correlation properties characterized by a merit factor larger than a first predetermined threshold, and cross-correlation properties among the color-coded sequences characterized by a demerit factor lower than a second predetermined threshold;
randomly selecting a color-coded sequence from the set of color-coded sequences;
illuminating a subject in accordance with the sequence of colors in the selected color-coded sequence;
capturing a sequence of images of the subject, wherein the sequence of images are temporally synchronized with illumination by the color-coded sequence;
generating, by a matched filtering process using the selected color-coded sequence, a filtered response image from the sequence of images of the subject;
determining, based on the filtered response image, that the subject is an alternative representation of a live person; and
responsive to determining that the subject is the alternative representation of the live person, preventing access to a secure system.
2. The method of claim 1 , wherein the alternative representation of the live person comprises a previously captured photograph of the live person printed on paper, or an image presented on a display device.
3. The method of claim 1 , wherein the alternative representation of the live person comprises video replay of the live person on a display device.
4. The method of claim 1 , wherein determining, based on the filtered response image, that the subject is the alternative representation of the live person comprises processing the filtered response image using a machine learning process trained to discriminate between filtered response images of live persons and filtered response images of alternative representations of the live persons.
5. The method of claim 1 , wherein generating, by the matched filtering process using the selected color-coded sequence, the filtered response image from the sequence of images of the subject comprises:
generating, from the sequence of images, a sequence of spatially aligned images;
calculating a sequence of filtered response images by applying a matched filter to the sequence of spatially aligned images, wherein the matched filter corresponds to the selected color-coded sequence; and
selecting the filtered response image from the sequence of filtered response images.
6. The method of claim 1 , wherein the merit factor is defined as:
∑
d
=
1
M
C
f
,
f
(
0
,
d
)
2
∑
s
!=
0
2
N
-
1
C
f
,
f
(
s
,
d
)
2
,
wherein M is a total number of colors in the color-coded sequence, N is length of the color-coded sequence, d is an index for the colors, C f,f is a one-dimensional discrete auto-correlation function of a color-coded sequence f, s is a correlation index in the function C f,f , and
wherein the demerit factor is defined as:
∑
d
=
1
M
∑
s
!=
0
2
N
-
1
C
f
,
g
(
s
,
d
)
2
C
f
,
f
(
0
,
d
)
·
C
g
,
g
(
0
,
d
)
,
wherein C g,g is a one-dimensional discrete auto-correlation of a color-coded sequence g, C f,g is a one-dimensional discrete cross-correlation of the color-coded sequence f and the color-coded sequence g.
7. The method of claim 1 , further comprising:
randomly selecting a second color-coded sequence from the set of color-coded sequences;
illuminating a second subject in accordance with the sequence of colors in the second color-coded sequence;
capturing a second sequence of images of the second subject, wherein the second sequence of images are temporally synchronized with illumination by the second color-coded sequence;
generating, by a matched filtering process using the second selected color-coded sequence, a second filtered response image from the second sequence of images of the second subject;
determining, based on the second filtered response image, that the second subject is a live person; and
responsive to determining that the second subject is the live person, initiating an authentication process for determining if the live person is authorized to access the secure system.
8. A computer-implemented system, comprising:
one or more computers; and
one or more computer memory devices interoperably coupled with the one or more computers and having tangible, non-transitory, machine-readable media storing one or more instructions that, when executed by the one or more computers, perform operations comprising:
obtaining a set of multiple color-coded sequences, wherein each color-coded sequence comprises a sequence of colors, wherein each color-coded sequence has auto-correlation properties characterized by a merit factor larger than a first predetermined threshold, and cross-correlation properties among the color-coded sequences characterized by a demerit factor lower than a second predetermined threshold;
randomly selecting a color-coded sequence from the set of color-coded sequences;
illuminating a subject in accordance with the sequence of colors in the selected color-coded sequence;
capturing a sequence of images of the subject, wherein the sequence of images are temporally synchronized with illumination by the color-coded sequence;
generating, by a matched filtering process using the selected color-coded sequence, a filtered response image from the sequence of images of the subject;
determining, based on the filtered response image, that the subject is an alternative representation of a live person; and
responsive to determining that the subject is the alternative representation of the live person, preventing access to a secure system.
9. The computer-implemented system of claim 8 , wherein the alternative representation of the live person comprises a previously captured photograph of the live person printed on paper, or an image presented on a display device.
10. The computer-implemented system of claim 8 , wherein the alternative representation of the live person comprises video replay of the live person on a display device.
11. The computer-implemented system of claim 8 , wherein determining, based on the filtered response image, that the subject is the alternative representation of the live person comprises processing the filtered response image using a machine learning process trained to discriminate between filtered response images of live persons and filtered response images of alternative representations of the live persons.
12. The computer-implemented system of claim 8 , wherein generating, by the matched filtering process using the selected color-coded sequence, the filtered response image from the sequence of images of the subject comprises:
generating, from the sequence of images, a sequence of spatially aligned images;
calculating a sequence of filtered response images by applying a matched filter to the sequence of spatially aligned images, wherein the matched filter corresponds to the selected color-coded sequence; and
selecting the filtered response image from the sequence of filtered response images.
13. The computer-implemented system of claim 8 , wherein the merit factor is defined as:
∑
d
=
1
M
C
f
,
f
(
0
,
d
)
2
∑
s
!=
0
2
N
-
1
C
f
,
f
(
s
,
d
)
2
,
wherein M is a total number of colors in the color-coded sequence, N is length of the color-coded sequence, d is an index for the colors, C f,f is a one-dimensional discrete auto-correlation function of a color-coded sequence f, s is a correlation index in the function C f,f , and
wherein the demerit factor is defined as:
∑
d
=
1
M
∑
s
!=
0
2
N
-
1
C
f
,
g
(
s
,
d
)
2
C
f
,
f
(
0
,
d
)
·
C
g
,
g
(
0
,
d
)
,
wherein C g,g is a one-dimensional discrete auto-correlation of a color-coded sequence g, C f,g is a one-dimensional discrete cross-correlation of the color-coded sequence f and the color-coded sequence g.
14. The computer-implemented system of claim 8 , the operations further comprise:
randomly selecting a second color-coded sequence from the set of color-coded sequences;
illuminating a second subject in accordance with the sequence of colors in the second color-coded sequence;
capturing a second sequence of images of the second subject, wherein the second sequence of images are temporally synchronized with illumination by the second color-coded sequence;
generating, by a matched filtering process using the second selected color-coded sequence, a second filtered response image from the second sequence of images of the second subject;
determining, based on the second filtered response image, that the second subject is a live person; and
responsive to determining that the second subject is the live person, initiating an authentication process for determining if the live person is authorized to access the secure system.
15. A non-transitory, computer-readable medium storing one or more instructions executable by a computer system to perform operations comprising:
obtaining a set of multiple color-coded sequences, wherein each color-coded sequence comprises a sequence of colors, wherein each color-coded sequence has auto-correlation properties characterized by a merit factor larger than a first predetermined threshold, and cross-correlation properties among the color-coded sequences characterized by a demerit factor lower than a second predetermined threshold;
randomly selecting a color-coded sequence from the set of color-coded sequences;
illuminating a subject in accordance with the sequence of colors in the selected color-coded sequence;
capturing a sequence of images of the subject, wherein the sequence of images are temporally synchronized with illumination by the color-coded sequence;
generating, by a matched filtering process using the selected color-coded sequence, a filtered response image from the sequence of images of the subject;
determining, based on the filtered response image, that the subject is an alternative representation of a live person; and
responsive to determining that the subject is the alternative representation of the live person, preventing access to a secure system.
16. The non-transitory, computer-readable medium of claim 15 , wherein the alternative representation of the live person comprises a previously captured photograph of the live person printed on paper, or an image presented on a display device.
17. The non-transitory, computer-readable medium of claim 15 , wherein the alternative representation of the live person comprises video replay of the live person on a display device.
18. The non-transitory, computer-readable medium of claim 15 , wherein determining, based on the filtered response image, that the subject is the alternative representation of the live person comprises processing the filtered response image using a machine learning process trained to discriminate between filtered response images of live persons and filtered response images of alternative representations of the live persons.
19. The non-transitory, computer-readable medium of claim 15 , wherein generating, by the matched filtering process using the selected color-coded sequence, the filtered response image from the sequence of images of the subject comprises:
generating, from the sequence of images, a sequence of spatially aligned images;
calculating a sequence of filtered response images by applying a matched filter to the sequence of spatially aligned images, wherein the matched filter corresponds to the selected color-coded sequence; and
selecting the filtered response image from the sequence of filtered response images.
20. The non-transitory, computer-readable medium of claim 15 , wherein the merit factor is defined as:
∑
d
=
1
M
C
f
,
f
(
0
,
d
)
2
∑
s
!=
0
2
N
-
1
C
f
,
f
(
s
,
d
)
2
,
wherein M is a total number of colors in the color-coded sequence, N is length of the color-coded sequence, d is an index for the colors, C f,f is a one-dimensional discrete auto-correlation function of a color-coded sequence f, s is a correlation index in the function C f,f , and
wherein the demerit factor is defined as:
∑
d
=
1
M
∑
s
!=
0
2
N
-
1
C
f
,
g
(
s
,
d
)
2
C
f
,
f
(
0
,
d
)
·
C
g
,
g
(
0
,
d
)
,
wherein C g,g is a one-dimensional discrete auto-correlation of a color-coded sequence g, C f,g is a one-dimensional discrete cross-correlation of the color-coded sequence f and the color-coded sequence g.