Apparatus and method for generating image feature data
View Patent ↗The present invention relates to an apparatus and method for efficiently generating feature data which properly determines a feature point indicating features of images and describes the feature point. The apparatus for generating image feature data comprises: a feature point determining unit which determines a feature point from an image and extracts information on the determined feature point; a feature point filtering unit which determines, as a final feature point, at least one feature point from among the feature points determined by the feature point determining unit; and a feature data generating unit which generates image feature data based on the final feature points determined by the feature point filtering unit and feature point information on the final feature points.
1. An apparatus for generating feature data of an image, comprising:
a feature point determination unit for determining feature points from an image and extracting feature point information for the determined feature points;
a feature point filtering unit for determining one or more of the feature points determined by the feature point determination unit to be final feature points; and
a feature data generation unit for generating feature data of the image based on the final feature points determined by the feature point filtering unit and feature point information for the final feature points,
wherein the feature point information extracted by the feature point determination unit includes intensities of the feature points,
wherein the feature point filtering unit determines a point having a larger intensity than other points located in a surrounding area of a corresponding feature point to be a final feature point, based on the intensities of the feature points, and
wherein the feature point filtering unit determines important points R 2 (c i ) satisfying the expression
R
2
(
c
i
)
=
{
c
i
❘
f
(
c
i
)
>
max
R
1
(
c
i
)
[
f
(
c
i
)
]
×
T
1
}
among the points located in the surrounding area of the feature point, where c i denotes an i-th feature point, f(c i ) denotes intensity of the i-th feature point, R 1 (c i ) denotes a set of points in the surrounding area of the feature point,
max
R
1
(
c
i
)
[
f
(
c
i
)
]
denotes a maximum value of intensities of R 1 (c i ), and T 1 denotes a threshold, and determines a feature point satisfying the expression
f
(
c
i
)
>
∑
c
i
∈
R
2
(
c
i
)
f
(
c
i
)
#
(
R
2
(
c
i
)
)
×
T
2
to be the final feature point, where # denotes an operator for obtaining a size of the set and T 2 denotes a threshold.
2. A method of generating feature data of an image, comprising:
determining feature points from an image and extracting feature point information for the determined feature points;
determining one or more of the determined feature points to be final feature points; and
generating feature data of the image based on the determined final feature points and feature point information for the final feature points,
wherein at the extracting feature point information for the determined feature points, the feature point information includes intensities of the feature points,
wherein the step of determining one or more of the determined feature points to be final feature points is configured to determine a point having a larger intensity than other points located in a surrounding area of a corresponding feature point to be a final feature point, based on the intensities of the feature points, and to determine important points R 2 (c i ) satisfying the expression
R
2
(
c
i
)
=
{
c
i
❘
f
(
c
i
)
>
max
R
1
(
c
i
)
[
f
(
c
i
)
]
×
T
1
}
among the points located in the surrounding area of the feature point, where c i denotes an i-th feature point, f(c i ) denotes intensity of the i-th feature point, R 1 (c i ) denotes a set of points in the surrounding area of the feature point,
max
R
1
(
c
i
)
[
f
(
c
i
)
]
denotes a maximum value of intensities of R 1 (c i ), and T 1 denotes a threshold, and determine a feature point satisfying the expression
f
(
c
i
)
>
∑
c
i
∈
R
2
(
c
i
)
f
(
c
i
)
#
(
R
2
(
c
i
)
)
×
T
2
to be the final feature point, where # denotes an operator for obtaining a size of the set and T 2 denotes a threshold.
3. An apparatus for generating feature data of an image, comprising:
a feature point determination unit for determining feature points from an image and extracting feature point information for the determined feature points;
a feature point orientation estimation unit for estimating pieces of orientation information for the respective feature points determined by the feature point determination unit; and
a feature data generation unit for generating binary feature vectors based on the feature point information and the orientation information, for the respective feature points determined by the feature point determination unit, and generating feature data of the image including the generated binary feature vectors,
wherein the feature data generation unit generates for the feature points, surrounding image areas including the respective feature points, aligns the generated surrounding images areas in an identical orientation, divides each of the aligned surrounding image areas into sub-regions, and generates the binary feature vectors based on averages of brightness values of the subs-regions.
4. The apparatus of claim 3 , wherein the feature point orientation estimation unit calculates, for all points in a predetermined surrounding area of each feature point, surrounding gradients of the points, and obtains an average of orientations of the gradients, thus estimating orientation of the feature point.
5. The apparatus of claim 3 , wherein each binary feature vector is generated by at least one selected from among difference vectors and double-difference vectors obtained from the averages of brightness values of the sub-regions.
6. The apparatus of claim 5 , wherein selection of at least one from among the difference vectors and the double-difference vectors obtained from averages of brightness values of the sub-regions is performed in correspondence with respective bits of the binary feature vector.
7. The apparatus of claim 6 , wherein a linear combination or a nonlinear combination is calculated for the difference vectors and the double-difference vectors selected in correspondence with the respective bits, and resulting values of the calculation are compared with a threshold, thus enabling values of corresponding bits of the binary feature vector to be determined.
8. The apparatus of claim 7 , wherein alignment is performed based on a criterion preset for the respective bits of the binary feature vector.
9. The apparatus of claim 3 , wherein the feature data of the image further includes one or more of location information, size information, and orientation information of each feature point.
10. The apparatus of claim 3 , wherein the feature point determination unit further comprises a feature point filtering unit for determining one or more of the determined feature points to be final feature points.
11. The apparatus of claim 3 , wherein:
the feature point information extracted by the feature point determination unit includes intensities of the feature points, and
the feature point filtering unit determines a point having a larger intensity than other points located in a surrounding area of a corresponding feature point to be a final feature point, based on the intensities of the feature points.
12. The apparatus of claim 11 , wherein the feature point filtering unit determines important points R 2 (c i ) satisfying the expression
R
2
(
c
i
)
=
{
c
i
❘
f
(
c
i
)
>
max
R
1
(
c
i
)
[
f
(
c
i
)
]
×
T
1
}
among the points located in the surrounding area of the feature point where c i denotes an i-th feature point, f(c i ) denotes intensity of the i-th feature point, R 1 (c i ) denotes a set of points in the surrounding area of the feature point,
max
R
1
(
c
i
)
[
f
(
c
i
)
]
denotes a maximum value of intensities of R 1 (c i ), and T 1 denotes a threshold and determines a feature point satisfying the expression
f
(
c
i
)
>
∑
c
i
∈
R
2
(
c
i
)
f
(
c
i
)
#
(
R
2
(
c
i
)
)
×
T
2
to be the final feature point where # denotes an operator for obtaining a size of the set and T 2 denotes a threshold.
13. A method of generating feature data of an image, comprising:
determining feature points from an image and extracting feature point information for the determined feature points;
estimating pieces of orientation information for the respective determined feature points; and
generating binary feature vectors based on the feature point information and the orientation information, for the respective determined feature points, and generating feature data of the image including the generated binary feature vectors,
wherein the step of generating feature data of the image including the generated binary feature vectors is configured to generate, for the feature points, surrounding image areas including the respective feature points, aligns the generated surrounding images areas in an identical orientation, divides each of the aligned surrounding image areas into sub-regions, and generates the binary feature vectors based on averages of brightness values of the subs-regions.