# US 8,787,625 B2 · Use of relatively permanent pigmented or vascular skin mark patterns in images for personal identification

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 8,787,625 B2
- Document ID: US8787625B2
- Status: Granted
- Kind: B2
- Application number: 13/457,340
- Filed: 2012-04-26
- Published: 2012-11-01
- Granted: 2014-07-22
- Earliest claimed: 2011-04-27
- Assignees: Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center
- Applicants: Noah Ames Craft; Wai Kin Adams Kong; Arfika Nurhudatiana
- Inventors: Noah Ames Craft; Wai Kin Adams Kong; Arfika Nurhudatiana
- Web page: https://patentgrep.com/patent/US8787625B2

## CPC

- G06V 40/10: Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06T 3/02: Affine transformations (for image registration G06T3/147; for image mosaicing G06T3/4038)
- G06T 7/337: {involving reference images or patches}
- G06T 2207/30088: Skin; Dermal
- G06V 40/14: Vascular patterns
- G06V 2201/03: Recognition of patterns in medical or anatomical images
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

Embodiments include methods, devices, software, and systems for identifying a person based on relatively permanent pigmented or vascular skin mark (RPPVSM) patterns in images. Locations of RPPVSMs in different images of people are point matched, and a correspondence probability that the point matched RPPVSMs are from different people is calculated. Other embodiments are also described. Other embodiments are also described and claimed.

## Claims

### Claim 1

A method comprising:
point matching a first plurality of RPPVSMs in a first image with a second plurality of RPPVSMs in a second image; and
calculating a correspondence probability that the first plurality of RPPVSMs and the second plurality of RPPVSMs are from different persons, wherein one of the point matching or the calculating is performed by a processor, and wherein calculating a correspondence probability includes using:
p
⁡
(
A
,
C
,
n
,
m
,
p
)
=
C
p
n
⁢
∏
i
=
0
p
-
1
⁢
⁢
(
(
m
-
i
)
⁢
C
A
-
iC
)
⁢
∏
i
=
p
n
-
1
⁢
⁢
(
A
-
(
m
+
p
-
i
)
⁢
C
A
-
iC
)
where C=πr2, A is a size of overlapped total skin area of two images containing RPPVSMs, r is a distance threshold between two RPPVSMs to be considered matching, and p is a number of matching RPPVSMs of the first and second images.

### Claim 2 (depends on 1)

The method of claim 1, wherein the first plurality of RPPVSMs are on a first area skin of a person in the first image, and the second plurality RPPVSMs are on a second area of skin of a person in the second image.

### Claim 3 (depends on 1)

The method of claim 1, wherein the second plurality of RPPVSMs includes at least a third plurality of RPPVSMs in at least a third image.

### Claim 4 (depends on 1)

The method of claim 1, further comprising a medical expert identifying the first plurality of RPPVSMs in the first image; and a medical expert identifying the second plurality of RPPVSMs in the second image.

### Claim 5 (depends on 1)

The method of claim 1, wherein point matching includes image registration and parameter estimation comprising:
selecting a base image;
selecting registration points for the images; and
transforming the images and translating the images to align the registration points.

### Claim 6 (depends on 5)

The method of claim 5, wherein selecting a base image comprises selecting an image with the lowest resolution; and
wherein selecting the registration points comprises one of selecting skin marks that are not a RPPVSMs, selecting points at a boundary of an area of skin that the RPPVSMs are within, selecting points with a larger distance between the points than between any two adjacent RPPVSMs of the first or second image, and selecting three registration points.

### Claim 7 (depends on 5)

The method of claim 5, wherein transforming and translating comprises one of a global transformation, and Affine transformation, a combined transformation for all images, a transformation using a parameter matrix (A) and a vector (b) where x→Ax+b.

### Claim 8 (depends on 5)

The method of claim 5, wherein transforming and translating comprises:
preserving co-linearity relation between the registration points; and
preserving ratios of distances along lines between the registration points.

### Claim 9 (depends on 5)

The method of claim 5, wherein transforming and translating the images comprises a transformation using a parameter matrix (A) and a vector (b) where x→Ax+b, and calculating parameter matrix A and vector b using:
[
u
v
]
=
[
a
0
a
1
a
2
b
0
b
1
b
2
]
⁡
[
1
x
y
]
where (u, v) is the transformed coordinate of (x, y) and
[
a
0
a
1
a
2
b
0
b
1
b
2
]
is a parameter matrix that can be obtained by solving the three equations,
[
x
ei
y
ei
]
=
[
a
0
a
1
a
2
b
0
b
1
b
2
]
⁡
[
1
x
si
y
si
]
,
where i=1, 2 and 3.

### Claim 10 (depends on 1)

The method of claim 1, wherein calculating the correspondence probability comprises:
estimating parameters A—the size of overlapped total skin area of two images containing RPPVSMs, r—the distance threshold between two RPPVSMs to be considered matching, and p—the number of matching RPPVSMs of the first and second images.

### Claim 11 (depends on 1)

The method of claim 1, wherein calculating the correspondence probability comprises identifying clusters of possible matching RPPVSMs of the images;
wherein the clusters of possible matching RPPVSMs are identified using a combined plot of the first and second plurality of RPPVSMs after a linear transformation and translation of the registration points preserving co-linearity relation between the registration points; and preserving ratios of distances along lines between the registration points.

### Claim 12 (depends on 10)

The method of claim 10, wherein estimating A comprises counting the area of overlap of the first and second area of skin;
wherein estimating r comprises selecting a value of r for which Pr the probability of genuine correspondence of any two RPPVSMs is a selected threshold; and
wherein P is calculated based on r.

### Claim 13 (depends on 12)

The method of claim 12, wherein estimating A and r comprises using:
Pr(√{square root over ((x1−x2)2+(y1−y2)2)}{square root over ((x1−x2)2+(y1−y2)2)}≦r)≧0.975,
where (x1, y1) and (x2, y2) are two genuine corresponding nevi from the same person.

### Claim 14 (depends on 13)

The method of claim 13 wherein calculating the correspondence probability comprises calculating the probabilities of RPPVSMs correspondence of the first plurality of RPPVSMs corresponding to the second plurality of RPPVSMs using:
p
⁡
(
A
,
C
,
n
,
m
,
p
)
=
C
p
n
⁢
∏
i
=
0
p
-
1
⁢
⁢
(
(
m
-
i
)
⁢
C
A
-
iC
)
⁢
∏
i
=
p
n
-
1
⁢
⁢
(
A
-
(
m
+
p
-
i
)
⁢
C
A
-
iC
)
,
where C=πr2, p is the number of matched points, and Cpn means n!/p!(n−p)!;
[
u
v
]
=
[
a
0
a
1
a
2
b
0
b
1
b
2
]
⁡
[
1
x
y
]
where (u, v) is the transformed coordinate of (x, y) and
[
a
0
a
1
a
2
b
0
b
1
b
2
]
is a parameter matrix that can be obtained by solving the three equations,
[
x
ei
y
ei
]
=
[
a
0
a
1
a
2
b
0
b
1
b
2
]
⁡
[
1
x
si
y
si
]
,
where i=1, 2 and 3; and
Pr(√{square root over ((x1−x2)2+(y1−y2)2)}{square root over ((x1−x2)2+(y1−y2)2)}≦r)≧0.975,
where (x1, y1) and (x2, y2) are two corresponding nevi from one person in a given training database.

### Claim 15 (depends on 1)

The method of claim 1, wherein calculating the correspondence probability comprises:
p
⁡
(
A
,
C
,
n
,
m
,
p
)
=
C
p
n
⁢
∏
i
=
0
p
-
1
⁢
⁢
(
(
m
-
i
)
⁢
C
A
-
iC
)
⁢
∏
i
=
p
n
-
1
⁢
⁢
(
A
-
(
m
+
p
-
i
)
⁢
C
A
-
iC
)
,
where C=πr2, p is the number of matched points, and Cpn means n!/p!(n−p)!.

### Claim 16 (depends on 1)

The method of claim 1, calculating whether the correspondence probability is sufficient for the point matching to be considered a traditional Hard Biometric Trait.

### Claim 17

A non-transitory machine-readable medium having data therein which when accessed by a processor causes the processor to:
point match a first plurality of RPPVSMs in a first image with a second plurality of RPPVSMs in a second image; and
calculate a correspondence probability that the first plurality of RPPVSMs and the second plurality of RPPVSMs are from different persons, wherein calculating a correspondence probability includes using:
p
⁡
(
A
,
C
,
n
,
m
,
p
)
=
C
p
n
⁢
∏
i
=
0
p
-
1
⁢
⁢
(
(
m
-
i
)
⁢
C
A
-
iC
)
⁢
∏
i
=
p
n
-
1
⁢
⁢
(
A
-
(
m
+
p
-
i
)
⁢
C
A
-
iC
)
,
where C=πr2, A is a size of overlapped total skin area of two images containing RPPVSMs, r is a distance threshold between two RPPVSMs to be considered matching, and p is a number of matching RPPVSMs of the first and second images.

### Claim 18 (depends on 17)

The non-transitory machine-readable medium of claim 17 wherein point matching includes image registration and parameter estimation comprising:
selecting a base image;
selecting registration points for the images; and
transforming the images and translating the images to align the registration points.

### Claim 19 (depends on 18)

The non-transitory machine-readable medium of claim 18 wherein calculating the correspondence probability comprises;
estimating parameters A—the size of overlapped total skin area of two images containing RPPVSMs, r—the distance threshold between two RPPVSMs to be considered matching, and p—the number of matching RPPVSMs of the first and second images.

### Claim 20 (depends on 19)

The non-transitory machine-readable medium of claim 19 wherein transforming and translating the images comprises calculating parameter matrix A and vector b using:
[
u
v
]
=
[
a
0
a
1
a
2
b
0
b
1
b
2
]
⁡
[
1
x
y
]
where (u, v) is the transformed coordinate of (x, y) and
[
a
0
a
1
a
2
b
0
b
1
b
2
]
is a parameter matrix that can be obtained by solving the three equations,
[
x
ei
y
ei
]
=
[
a
0
a
1
a
2
b
0
b
1
b
2
]
⁡
[
1
x
si
y
si
]
,
where i=1, 2 and 3; and wherein estimating A and r comprises using:
Pr(√{square root over ((x1−x2)2+(y1−y2)2)}{square root over ((x1−x2)2+(y1−y2)2)}≦r)≧0.975,
where (x1, y1) and (x2, y2) are two corresponding nevi in the first and second images.

### Claim 21

A system comprising:
a processor;
a digital image input;
a memory coupled to said processor and containing code which when executed:
point matches a first plurality of RPPVSMs in a first image with a second plurality of RPPVSMs in a second image; and
calculates a correspondence probability that the first plurality of RPPVSMs and the second plurality of RPPVSMs are from different persons, wherein calculating a correspondence probability includes using:
p
⁡
(
A
,
C
,
n
,
m
,
p
)
=
C
p
n
⁢
∏
i
=
0
p
-
1
⁢
⁢
(
(
m
-
i
)
⁢
C
A
-
iC
)
⁢
∏
i
=
p
n
-
1
⁢
⁢
(
A
-
(
m
+
p
-
i
)
⁢
C
A
-
iC
)
,
where C=πr2, A is a size of overlapped total skin area of two images containing RPPVSMs, r is a distance threshold between two RPPVSMs to be considered matching, and p is a number of matching RPPVSMs of the first and second images.

### Claim 22 (depends on 21)

The system of claim 21, wherein point matching includes image registration and parameter estimation comprising:
selecting a base image;
selecting registration points for the images; and
transforming the images and translating the images to align the registration points.

### Claim 23 (depends on 22)

The system of claim 22 wherein calculating the correspondence probability comprises identifying clusters of possible matching RPPVSMs of the images;
estimating parameters A—the size of overlapped total skin area of two images containing RPPVSMs, r—the distance threshold between two RPPVSMs to be considered matching, and p—the number of matching RPPVSMs of the first and second images.

### Claim 24 (depends on 23)

The system of claim 23 wherein transforming and translating the images comprises calculating parameter matrix A and vector b using:
[
u
v
]
=
[
a
0
a
1
a
2
b
0
b
1
b
2
]
⁡
[
1
x
y
]
where (u, v) is the transformed coordinate of (x, y) and
[
a
0
a
1
a
2
b
0
b
1
b
2
]
is a parameter matrix that can be obtained by solving the three equations,
[
x
ei
y
ei
]
=
[
a
0
a
1
a
2
b
0
b
1
b
2
]
⁡
[
1
x
si
y
si
]
,
where i=1, 2 and 3; and wherein estimating A and r comprises using:
Pr(√{square root over ((x1−x2)2+(y1−y2)2)}{square root over ((x1−x2)2+(y1−y2)2)}≦r)≧0.975,
where (x1, y1) and (x2, y2) are two corresponding nevi in the first and second images.