Face location tracking method, apparatus, and electronic device
View Patent ↗A first face region within a first image is determined. The first face region includes a location of a face within the first image. Based on the determined first face region within the first image, a predicted face region within a second image is determined. A first region of similarity within the predicted face region is determined. The first region of similarity has at least a predetermined degree of similarity to the first face region within the first image. Whether a second face region is present within the second image is determined. The location of the face within the second image is determined based on the first region of similarity, the determination of whether the second face region is present within the second image, and a face region selection rule.
1. A computer-implemented method for tracking locations of a face across a plurality of images comprising a first image and a second image, the method comprising:
determining a first face region within the first image, the first face region including the location of the face within the first image;
based on the determined first face region within the first image, determining a predicted face region within the second image;
determining a first region of similarity within the predicted face region, the first region of similarity having at least a predetermined degree of similarity to the first face region within the first image;
determining whether a second face region is present within the second image; and
determining the location of the face within the second image based on the first region of similarity, the determination of whether the second face region is present within the second image, and a face region selection rule comprising:
when the second face region is not present within the second image, the location of the face is the first region of similarity;
when the second face region is present within the second image, and a spatial overlap coefficient between the second face region and the first region of similarity is 0, the location of the face is the first region of similarity;
when the second face region is present within the second image, and the spatial overlap coefficient between the second face region and the first region of similarity is less than a preset threshold, the location of the face is the first region of similarity; and
when the second face region is present within the second image, and the spatial overlap coefficient between the second face region and the first region of similarity is greater than or equal to the preset threshold, the location of the face is the second face region.
2. The computer-implemented method of claim 1 , wherein determining the first region of similarity within the predicted face region comprises:
setting a first comparison region within the predicted face region;
determining a first degree of similarity between the first comparison region and the first face region;
determining that the first degree of similarity satisfies the predetermined degree of similarity; and
based on the determination that the first degree of similarity satisfies the predetermined degree of similarity, determining the first comparison region to be the first region of similarity.
3. The computer-implemented method of claim 2 , wherein determining the first degree of similarity between the first comparison region and the first face region comprises performing calculations according to the following equations:
min
X
=
max
(
-
left
ori
,
-
left
des
)
max
X
=
max
(
width
-
left
ori
,
width
-
left
des
)
min
Y
=
max
(
-
top
ori
,
-
top
des
)
max
Y
=
max
(
height
-
top
ori
,
height
-
top
des
)
sumDis
=
{
∑
i
=
max
(
1
,
minX
)
min
(
width
,
maxX
)
∑
j
=
max
(
1
,
minY
)
min
(
height
,
maxX
)
min
{
f
(
i
,
j
)
-
g
(
i
,
j
)
,
x
}
}
effctiveNum
=
[
min
(
width
,
max
X
)
-
max
(
1
,
min
X
)
]
*
[
min
(
height
,
max
X
)
-
max
(
1
,
min
Y
)
]
dis
=
sumDis
*
(
width
*
height
)
/
effctiveNum
wherein left ori , left des , top ori , and top des respectively represent a left boundary location of the first face region, a left boundary location of the first comparison region, an upper boundary location of the first face region, and an upper boundary location of the first comparison region,
wherein width represents a width of the first face region and height represents a height of the first face region,
wherein f(i,j) represents a grayscale value of a pixel of the first image having coordinates (i,j) in the first face region of the first image and g(i,j) represents a grayscale value of a pixel of the second image having coordinates (i,j) in the first comparison region of the second image, and
wherein x represents a preset threshold and dis is the first degree of similarity between the first face region and the first comparison region.
4. The computer-implemented method of claim 2 , wherein determining the first region of similarity within the predicted face region further comprises:
setting a second comparison region within the predicted face region, the second comparison region being different from the first comparison region;
determining a second degree of similarity between the second comparison region and the first face region; and
determining that the second degree of similarity satisfies the predetermined degree of similarity, and
wherein determining the first comparison region to be the first region of similarity comprises:
determining that the first degree of similarity is greater than the second degree of similarity; and
based on the determination that the first degree of similarity is greater than the second degree of similarity, determining the first comparison region to be the first region of similarity.
5. The computer-implemented method of claim 2 , wherein determining the first region of similarity within the predicted face region further comprises:
setting a second comparison region within the predicted face region by translating the first comparison region by a first step, the first step being two or more pixels of the second image.
6. The computer-implemented method of claim 5 , further comprising:
based on the determination of the first region of similarity within the predicted face region, determining a second region of similarity within the predicted face region, comprising:
setting a third comparison region within the predicted face region by translating the first region of similarity by a second step smaller than the first step;
determining a third degree of similarity between the third comparison region and the first face region;
determining that the third degree of similarity is greater than the first degree of similarity; and
based on the determination that the third degree of similarity is greater than the first degree of similarity, determining the third comparison region to be the second region of similarity,
wherein the determination of the location of the face within the second image is further based on the second region of similarity.
7. The computer-implemented method of claim 1 , wherein determining a first face region within the first image comprises:
determining that the first image comprises a plurality of faces;
determining a plurality of face region areas; and
determining a region comprising a face of the plurality of faces having the largest face region area to be the first face region.
8. A non-transitory, computer-readable medium storing one or more instructions executable by a computer system to perform operations comprising:
determining a first face region within a first image, the first face region including a location of a face within the first image;
based on the determined first face region within the first image, determining a predicted face region within a second image;
determining a first region of similarity within the predicted face region, the first region of similarity having at least a predetermined degree of similarity to the first face region within the first image;
determining whether a second face region is present within the second image; and
determining the location of the face within the second image based on the first region of similarity, the determination of whether the second face region is present within the second image, and a face region selection rule comprising:
when the second face region is not present within the second image, the location of the face is the first region of similarity;
when the second face region is present within the second image, and a spatial overlap coefficient between the second face region and the first region of similarity is 0, the location of the face is the first region of similarity;
when the second face region is present within the second image, and the spatial overlap coefficient between the second face region and the first region of similarity is less than a preset threshold, the location of the face is the first region of similarity; and
when the second face region is present within the second image, and the spatial overlap coefficient between the second face region and the first region of similarity is greater than or equal to the preset threshold, the location of the face is the second face region.
9. The computer-readable medium of claim 8 , wherein determining the first region of similarity within the predicted face region comprises:
setting a first comparison region within the predicted face region;
determining a first degree of similarity between the first comparison region and the first face region;
determining that the first degree of similarity satisfies the predetermined degree of similarity; and
based on the determination that the first degree of similarity satisfies the predetermined degree of similarity, determining the first comparison region to be the first region of similarity.
10. The computer-readable medium of claim 9 , wherein determining the first degree of similarity between the first comparison region and the first face region comprises performing calculations according to the following equations:
min
X
=
max
(
-
left
ori
,
-
left
des
)
max
X
=
max
(
width
-
left
ori
,
width
-
left
des
)
min
Y
=
max
(
-
top
ori
,
-
top
des
)
max
Y
=
max
(
height
-
top
ori
,
height
-
top
des
)
sumDis
=
{
∑
i
=
max
(
1
,
minX
)
min
(
width
,
maxX
)
∑
j
=
max
(
1
,
minY
)
min
(
height
,
maxX
)
min
{
f
(
i
,
j
)
-
g
(
i
,
j
)
,
x
}
}
effctiveNum
=
[
min
(
width
,
max
X
)
-
max
(
1
,
min
X
)
]
*
[
min
(
height
,
max
X
)
-
max
(
1
,
min
Y
)
]
dis
=
sumDis
*
(
width
*
height
)
/
effctiveNum
wherein left ori , left des , top ori , and top des respectively represent a left boundary location of the first face region, a left boundary location of the first comparison region, an upper boundary location of the first face region, and an upper boundary location of the first comparison region,
wherein width represents a width of the first face region and height represents a height of the first face region,
wherein f(i,j) represents a grayscale value of a pixel of the first image having coordinates (i,j) in the first face region of the first image and g(i,j) represents a grayscale value of a pixel of the second image having coordinates (i,j) in the first comparison region of the second image, and
wherein x represents a preset threshold and dis is the first degree of similarity between the first face region and the first comparison region.
11. The computer-readable medium of claim 9 , wherein determining the first region of similarity within the predicted face region further comprises:
setting a second comparison region within the predicted face region, the second comparison region being different from the first comparison region;
determining a second degree of similarity between the second comparison region and the first face region; and
determining that the second degree of similarity satisfies the predetermined degree of similarity, and
wherein determining the first comparison region to be the first region of similarity comprises:
determining that the first degree of similarity is greater than the second degree of similarity; and
based on the determination that the first degree of similarity is greater than the second degree of similarity, determining the first comparison region to be the first region of similarity.
12. The computer-readable medium of claim 9 , wherein determining the first region of similarity within the predicted face region further comprises:
setting a second comparison region within the predicted face region by translating the first comparison region by a first step, the first step being two or more pixels of the second image.
13. The computer-readable medium of claim 12 , wherein the operations further comprise:
based on the determination of the first region of similarity within the predicted face region, determining a second region of similarity within the predicted face region, comprising:
setting a third comparison region within the predicted face region by translating the first region of similarity by a second step smaller than the first step;
determining a third degree of similarity between the third comparison region and the first face region;
determining that the third degree of similarity is greater than the first degree of similarity; and
based on the determination that the third degree of similarity is greater than the first degree of similarity, determining the third comparison region to be the second region of similarity,
wherein the determination of the location of the face within the second image is further based on the second region of similarity.
14. The computer-readable medium of claim 8 , wherein determining a first face region within the first image comprises:
determining that the first image comprises a plurality of faces;
determining a plurality of face region areas; and
determining a region comprising a face of the plurality of faces having the largest face region area to be the first face region.
15. 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 one or more operations comprising:
determining a first face region within a first image, the first face region including a location of a face within the first image;
based on the determined first face region within the first image, determining a predicted face region within a second image;
determining a first region of similarity within the predicted face region, the first region of similarity having at least a predetermined degree of similarity to the first face region within the first image;
determining whether a second face region is present within the second image; and
determining the location of the face within the second image based on the first region of similarity, the determination of whether the second face region is present within the second image, and a face region selection rule comprising:
when the second face region is not present within the second image, the location of the face is the first region of similarity;
when the second face region is present within the second image, and a spatial overlap coefficient between the second face region and the first region of similarity is 0, the location of the face is the first region of similarity;
when the second face region is present within the second image, and the spatial overlap coefficient between the second face region and the first region of similarity is less than a preset threshold, the location of the face is the first region of similarity; and
when the second face region is present within the second image, and the spatial overlap coefficient between the second face region and the first region of similarity is greater than or equal to the preset threshold, the location of the face is the second face region.
16. The computer-implemented system of claim 15 , wherein determining the first region of similarity within the predicted face region comprises:
setting a first comparison region within the predicted face region;
determining a first degree of similarity between the first comparison region and the first face region;
determining that the first degree of similarity satisfies the predetermined degree of similarity; and
based on the determination that the first degree of similarity satisfies the predetermined degree of similarity, determining the first comparison region to be the first region of similarity.
17. The computer-implemented system of claim 16 , wherein determining the first degree of similarity between the first comparison region and the first face region comprises performing calculations according to the following equations:
min
X
=
max
(
-
left
ori
,
-
left
des
)
max
X
=
max
(
width
-
left
ori
,
width
-
left
des
)
min
Y
=
max
(
-
top
ori
,
-
top
des
)
max
Y
=
max
(
height
-
top
ori
,
height
-
top
des
)
sumDis
=
{
∑
i
=
max
(
1
,
minX
)
min
(
width
,
maxX
)
∑
j
=
max
(
1
,
minY
)
min
(
height
,
maxX
)
min
{
f
(
i
,
j
)
-
g
(
i
,
j
)
,
x
}
}
effctiveNum
=
[
min
(
width
,
max
X
)
-
max
(
1
,
min
X
)
]
*
[
min
(
height
,
max
X
)
-
max
(
1
,
min
Y
)
]
dis
=
sumDis
*
(
width
*
height
)
/
effctiveNum
wherein left ori , left des , top ori , and top des respectively represent a left boundary location of the first face region, a left boundary location of the first comparison region, an upper boundary location of the first face region, and an upper boundary location of the first comparison region,
wherein width represents a width of the first face region and height represents a height of the first face region,
wherein f(i,j) represents a grayscale value of a pixel of the first image having coordinates (i,j) in the first face region of the first image and g(i,j) represents a grayscale value of a pixel of the second image having coordinates (i,j) in the first comparison region of the second image, and
wherein x represents a preset threshold and dis is the first degree of similarity between the first face region and the first comparison region.
18. The computer-implemented system of claim 16 , wherein determining the first region of similarity within the predicted face region further comprises:
setting a second comparison region within the predicted face region, the second comparison region being different from the first comparison region;
determining a second degree of similarity between the second comparison region and the first face region; and
determining that the second degree of similarity satisfies the predetermined degree of similarity, and
wherein determining the first comparison region to be the first region of similarity comprises:
determining that the first degree of similarity is greater than the second degree of similarity; and
based on the determination that the first degree of similarity is greater than the second degree of similarity, determining the first comparison region to be the first region of similarity.
19. The computer-implemented system of claim 16 , wherein determining the first region of similarity within the predicted face region further comprises:
setting a second comparison region within the predicted face region by translating the first comparison region by a first step, the first step being two or more pixels of the second image.
20. The computer-implemented system of claim 19 , further comprising:
based on the determination of the first region of similarity within the predicted face region, determining a second region of similarity within the predicted face region, comprising:
setting a third comparison region within the predicted face region by translating the first region of similarity by a second step smaller than the first step;
determining a third degree of similarity between the third comparison region and the first face region;
determining that the third degree of similarity is greater than the first degree of similarity; and
based on the determination that the third degree of similarity is greater than the first degree of similarity, determining the third comparison region to be the second region of similarity,
wherein the determination of the location of the face within the second image is further based on the second region of similarity.