 a processor that executes a process including: dividing an image of a subject into a plurality of blocks
 acquiring, for each block, a distance to an area on the subject corresponding to the block from a reference position based on a threedimensional shape data of the subject
 calculating, for each block, a weight according to the distance
 correcting, for each block, size information of a subject area that is projected in the block based on the weight
 and calculating a statistical amount of the subject based on the each size information corrected for each block.
Image processing apparatus, image processing method, and computerreadable recording medium
Updated Time 12 June 2019
Patent Registration DataPublication Number
US10002441
Application Number
US15/334372
Application Date
26 October 2016
Publication Date
19 June 2018
Current Assignee
FUJITSU LIMITED
Original Assignee (Applicant)
FUJITSU LIMITED
International Classification
G06K9/00,G06K9/32,G06T7/60,G06K9/52,G06T17/30
Cooperative Classification
G06T7/60,G06K9/00912,G06K9/3275,G06K9/52,G06T17/30
Inventor
AOKI, TAKAHIRO
Patent Images
This patent contains figures and images illustrating the invention and its embodiment.
Abstract
An image processing apparatus includes an imaging unit that divides an image of a subject into blocks, a quadricsurface fitting unit that calculates a distance from a reference position to the subject that corresponds to each block, a weight calculating unit that calculates a weight of each block according to the distance, and a barycenter calculating unit that corrects information about a size of the subject projected in each block based on the weight, and calculates barycentric coordinates based on the corrected information about the size of the subject.
Claims
1. An image processing apparatus comprising: a processor that executes a process including: dividing an image of a subject into a plurality of blocks; acquiring, for each block, a distance to an area on the subject corresponding to the block from a reference position based on a threedimensional shape data of the subject; calculating, for each block, a weight according to the distance; correcting, for each block, size information of a subject area that is projected in the block based on the weight; and calculating a statistical amount of the subject based on the each size information corrected for each block.
2. The image processing apparatus according to claim 1, wherein the block is a pixel.
3. The image processing apparatus according to claim 1, wherein the calculating a statistical amount includes calculating a barycenter as the statistical amount.
4. The image processing apparatus according to claim 1, wherein the calculating a statistical amount includes fitting the image of the subject into a quadric surface.
5. The image processing apparatus according to claim 4, wherein the calculating a statistical amount includes calculating a curvature mean on the quadric surface that is obtained by fitting, as the statistical amount.
6. The image processing apparatus according to claim 1, wherein the subject is a palm of a human.
7. A nontransitory computerreadable recording medium having stored therein a program that causes a computer to execute a process comprising: dividing an image of a subject into a plurality of blocks; acquiring, for each block, a distance to an area on the subject corresponding to the block from a reference position based on a threedimensional shape data of the subject; calculating, for each block, a weight according to the distance; correcting, for each block, size information of a subject area that is projected in the block based on the weight; and calculating a statistical amount of the subject based on the each size information corrected for each block.
8. An image processing method to be performed by a computer, the method comprising: dividing an image of a subject into a plurality of blocks using a processor; acquiring, for each block, a distance to an area on the subject corresponding to the block from a reference position based on a threedimensional shape data of the subject using the processor; calculating, for each block, a weight according to the distance using the processor; correcting, for each block, size information of a subject area that is projected in the block based on the weight using the processor; and calculating a statistical amount of the subject based on the each size information corrected for each block using the processor.
Claim Tree

11. An image processing apparatus comprising:

2. The image processing apparatus according to claim 1, wherein
 the block is a pixel.

3. The image processing apparatus according to claim 1, wherein
 the calculating a statistical amount includes calculating a barycenter as the statistical amount.

4. The image processing apparatus according to claim 1, wherein
 the calculating a statistical amount includes fitting the image of the subject into a quadric surface.

6. The image processing apparatus according to claim 1, wherein
 the subject is a palm of a human.


77. A nontransitory computerreadable recording medium having
 stored therein a program that causes a computer to execute a process comprising: dividing an image of a subject into a plurality of blocks
 acquiring, for each block, a distance to an area on the subject corresponding to the block from a reference position based on a threedimensional shape data of the subject
 calculating, for each block, a weight according to the distance
 correcting, for each block, size information of a subject area that is projected in the block based on the weight
 and calculating a statistical amount of the subject based on the each size information corrected for each block.

88. An image processing method to be performed by a computer, the method comprising:
 dividing an image of a subject into a plurality of blocks using a processor
 acquiring, for each block, a distance to an area on the subject corresponding to the block from a reference position based on a threedimensional shape data of the subject using the processor
 calculating, for each block, a weight according to the distance using the processor
 correcting, for each block, size information of a subject area that is projected in the block based on the weight using the processor
 and calculating a statistical amount of the subject based on the each size information corrected for each block using the processor.
Description
CROSSREFERENCE TO RELATED APPLICATION(S)
This application is based upon and claims the benefit of priority of the prior Japanese Patent Application No. 2015220640, filed on Nov. 10, 2015, the entire contents of which are incorporated herein by reference.
FIELD
The embodiments discussed herein are related to an image processing apparatus, and the like.
BACKGROUND
In image processing, statistical calculation, such as average value calculation, is applied in some cases to a subject region when the barycenter of the subject is calculated. For example, in palm vein authentication, image processing apparatuses that execute palm vein authentication calculate a mean average value of coordinates of each point in a palm of a hand being a subject, to identify the barycenter (Japanese Laidopen Patent Publication Nos. 2009238205 and 2007233981).
However, there has been a problem that it is difficult to achieve statistical calculation for a subject region accurately when the subject of image processing is inclined. For example, taking the palm vein authentication as an example, the problem that it is difficult to achieve statistical calculation accurately when an image of a subject region to be an authentication feature is processed is explained.
FIG. 11 explains influence caused when a subject is inclined. As depicted in FIG. 11, a camera projects a subject on an imaging device. When the subject is inclined, imaged areas of the subject to be projected in one pixel of the imaging device are not uniform. That is, a region B of the subject projected in a pixel P_{B }has a larger image area than a region A of the subject projected in a pixel P_{A}. In other words, in the pixel P_{B}, the region B that is longer than the region A is projected. The data amount of the pixel P_{B }is underestimated compared to the data amount of the pixel P_{A}.
In such a case, for example, the image processing apparatus calculates barycentric coordinates X_{G }by using following Equation (1). N refers to the total number of pixels, and x_{i }refers to coordinates of ith pixel. That is, in barycenter calculation, all pixels are calculated with the uniform weight of 1.0.
This results in underestimation of the data amount of the region B compared to the data amount of the region A. Inversely, the data amount of the region A is overestimated compared to the data amount of the region A. As a result, the barycentric coordinates X_{G }are drawn toward the direction of the region A, to be miscalculated.
The above problem is not limited only to the case of the subject being inclined, but also applies to the case of the subject being bent.
SUMMARY
According to an aspect of an embodiment, an image processing apparatus includes a processor that executes a process. The process includes dividing an image of a subject into a plurality of blocks. The process includes measuring a distance from a reference position to the subject corresponding to each of the blocks. The process includes setting a weight of each of the blocks according to the distance. The process includes correcting information about a size of the subject that is projected in each of the blocks based on the weight. The process includes calculating a statistical amount of the subject based on the corrected information about the size of the subject.
The object and advantages of the invention will be realized and attained by means of the elements and combinations particularly pointed out in the claims.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are not restrictive of the invention, as claimed.
BRIEF DESCRIPTION OF DRAWINGS
FIG. 1 is a functional block diagram depicting a configuration of an image processing apparatus according to a first embodiment;
FIG. 2 explains first calculation processing of barycentric coordinates according to the first embodiment;
FIG. 3 explains second calculation processing of barycentric coordinates according to the first embodiment;
FIG. 4 explains weighted quadricsurface fitting according to the first embodiment;
FIG. 5 is a flowchart of barycenter calculation processing according to the first embodiment;
FIG. 6 is a functional block diagram depicting a configuration of an image processing apparatus according to a second embodiment;
FIG. 7 explains calculation processing of barycentric coordinates used in gesture authentication;
FIG. 8 is a functional block diagram depicting a configuration of an image processing apparatus according to a third embodiment;
FIG. 9 is a flowchart of palmvein authentication processing according to a third embodiment;
FIG. 10 depicts one example of a computer that executes an image processing program; and
FIG. 11 explains influence caused when a subject is inclined.
DESCRIPTION OF EMBODIMENT(S)
Preferred embodiments of the present invention will be explained with reference to accompanying drawings. The present invention is not limited to the embodiments.
[a] First Embodiment
FIG. 1 is a functional block diagram depicting a configuration of an image processing apparatus according to a first embodiment. An image processing apparatus 1 depicted in FIG. 1 is designed to improve the accuracy in calculation of a barycenter by applying weighting processing based on an amount of information projected per unit (for example, pixel) of image processing when a barycenter is calculated from an image obtained by imaging a subject.
The image processing apparatus 1 includes an overall control unit 11, an imaging unit 12, a storage unit 13, a threedimensionalshape acquiring unit 14, and an image calculating unit 15.
The overall control unit 11 controls the entire image processing apparatus 1.
The imaging unit 12 performs imaging processing for acquiring an image of a subject. The imaging unit 12 includes a light source 121 and an imaging device 122. The light source 121 emits light, and as a result, the emitted light hit the subject. The imaging device 122 is a device that accumulates light entering through a lens as an electric signal, and is expressed by pixels. That is, the imaging device 122 divides an image of the subject into pixels. The imaging unit 12 is, for example, a camera. The imaging device 122 is, for example, a chargecoupled device (CCD), or a complementary metal oxide semiconductor (CMOS) device.
The storage unit 13 is, for example a semiconductor memory device such as a random access memory (RAM), a flash memory, or a storage device such as a hard disk and an optical disk. The storage unit 13 stores, for example, image data.
The threedimensionalshape acquiring unit 14 acquires threedimensional shape data of the subject. For example, the threedimensionalshape acquiring unit 14 acquires shape from shading (SFS) data, which is threedimensional shape data) by applying an SFS technique. The SFS technique is a technique of estimating a threedimensional shape of an object from brightness values of an image obtained by applying light to the subject. The estimated threedimensional shape data is called SFS data. As one example, in a palm vein sensor, the SFS technique is applied to reflection light of near infrared rays. Although it has been explained that the threedimensionalshape acquiring unit 14 acquires threedimensional shape data by applying the SFS technique, not limited to the SFS technique, and other techniques may be applied as long as threedimensional shape data can be acquired.
The image calculating unit 15 performs calculation processing by using the threedimensional shape data acquired by the threedimensionalshape acquiring unit 14. The image calculating unit 15 includes a quadricsurface fitting unit 151, a weight calculating unit 152, and a barycenter calculating unit 153.
The quadricsurface fitting unit 151 fits the threedimensional shape data of the subject into quadric surface. For example, the quadricsurface fitting unit 151 acquires height of the subject from the threedimensional shape data that is acquired by the threedimensionalshape acquiring unit 14. The quadricsurface fitting unit 151 then fits the subject into a quadric surface based on the height of the subject.
The weight calculating unit 152 calculates an area on the subject corresponding to one pixel on the image, by using the threedimensional shape data of the subject, and calculates a weight according to the calculated area. The weight calculating unit 152 calculates a weight according to an area by using Equation (2) below. N refers to the total number of pixels, S_{i }refers to an area on the subject that corresponds to ith pixel. s_{i }refers to a weight according to an area on the subject that corresponds to the ith pixel, and is a value obtained by normalizing S_{i}.
The barycenter calculating unit 153 calculates barycentric coordinates of the subject by using the weight according to an area on the subject. The barycenter calculating unit 153 calculates barycentric coordinates X_{G }by using Equation (3) below. That is, the barycenter calculating unit 153 corrects information relating to the size of the subject that is projected in each pixel based on the weight. N refers to the total number of pixels, x_{i }refers to coordinates of the ith pixel. s_{i }refers to a weight according to an area on the subject that corresponds to the ith pixel, and is a value acquired by Equation (2).
The calculation method of barycentric coordinates thus using weights according to an area on a subject is referred to as a first calculation method of barycentric coordinates for convenience of explanation.
First calculation processing of the barycentric coordinates according to the first embodiment is explained referring to FIG. 2. FIG. 2 explains the first calculation processing of barycentric coordinates according to the first embodiment. As depicted in FIG. 2, when the subject is inclined, imaged areas of the subject to be projected in one pixel are not uniform. That is, a region B of the subject projected in a pixel P_{B }has a larger area than a region A of the subject projected in a pixel P_{A}. That is, in the pixel P_{B}, the region B that is longer than the region A is projected. In other words, the data amount of the pixel P_{B }is underestimated compared to the pixel P_{A}.
The weight calculating unit 152 calculates an area on the subject that corresponds to one pixel on the image by using the threedimensional shape data of the subject. In this example, the area on the subject that corresponds to the pixel P_{A }is S_{i}. The area that corresponds to the pixel P_{B }is S_{j}. S_{i }is smaller than S_{j}. The weight calculating unit 152 calculates a weight according to the region by using Equation (2). In this example, the weight obtained by normalizing S_{i }is s_{i}.
Subsequently, the barycenter calculating unit 153 calculates barycentric coordinates of the subject by using a value obtained by multiplying coordinates of a pixel by the weight according to the area on the subject that corresponds to the pixel. This is to calculate the barycentric coordinates of the subject after the data amount of the pixel P_{B }that corresponds to the area S_{j }of the region B is equalized to the data amount of the pixel P_{A }that corresponds to the area S_{i }of the region A. Thus, the barycenter calculating unit 153 is enabled to calculate the barycentric coordinates corresponding to a subject region accurately.
Referring back to FIG. 1, although it has been explained that the weight calculating unit 152 calculates a weight according to an imaged area of the subject projected in one pixel, it is not limited thereto. The weight calculating unit 152 may calculate a weight according to a distance to the subject projected in one pixel from a reference position and inclination. That is, the weight calculating unit 152 calculates a distance to the subject corresponding to each pixel from the reference position and inclination based on the threedimensional shape data of the subject. The weight calculating unit 152 calculates a weight according to the distance and the inclination by using Equation (4) and Equation (5) below. The reference position is, for example, a position of the lens included in the imaging unit 12, but not limited thereto. N refers to the total number of pixels, and Z_{i }refers to a height of a point on the subject that corresponds to the ith pixel. n_{i }refers to a unit normal vector at a point on the subject that corresponds to the ith pixel. e_{z }refers to a unit normal vector in a Zaxis direction. W_{i }refers to a value that is determined by a height and inclination of a point on the subject that corresponds to the ith pixel. w_{i }refers to a weight according to a height and inclination on the subject that corresponds to the ith pixel, and is a value obtained by normalizing W_{i}.
When a weight according to a distance and inclination is calculated by the weight calculating unit 152, the barycenter calculating unit 153 calculates barycentric coordinates of the subject according to the weight according to the distance and inclination. The barycenter calculating unit 153 calculates the barycentric coordinates X_{G }by using following Equation (6). That is, the weight calculating unit 152 corrects information relating to a size of the subject that is projected in each pixel based on the weight. N refers to the total number of pixels, and x_{i }refers to coordinates of the ith pixel. w_{i }refers to a weight according to a height and inclination on the subject that corresponds to the ith pixel, and is a value acquired by Equation (5).
The calculation method of barycentric coordinates thus using a weight according to a distance is referred to as a second calculation method of barycentric coordinates for convenience of explanation.
A second calculation processing of barycentric coordinates according to the first embodiment is explained referring to FIG. 3. FIG. 3 explains the second calculation processing of barycentric coordinates according to the first embodiment. As depicted in FIG. 3, when the subject is inclined, heights from the reference position (for example, the lens) to the subject projected in one pixel are not uniform. Therefore, an imaged area to be projected in a pixel increases as the region increases in height. In this example, the region B of the subject is positioned higher than the region A of the subject, and the region B of the subject projected in the pixel P_{B }has a larger image area than the region A of the subject projected in the pixel P_{A}. That is, in the pixel P_{B}, the region B that is longer than the region A is projected. In other words, the data amount of the region B included in the pixel P_{B }is underestimated compared to the data amount of the region A included in the pixel P_{A}. Therefore, the weight calculating unit 152 increases W_{i }as the height Z_{i }of a region projected in one pixel increases to complement it, based on Equation (4).
In addition, the influence of inclination is calculated from an inner product of the unit normal vector e_{z }in the zaxis direction and a unit normal vector at a point of interest (refer to Equation (4)). In this example, the unit normal vector in the region A is n_{i}, and the unit normal vector in the region B is n_{j}. The inner product is 1.0 when the point of interest faces perpendicularly downward, and decreases as the subject inclines. When the subject is inclined, the region is imaged smaller than actual size on the camera side. That is, when the subject is inclined, a predetermined region including a point of interest has a larger area to be imaged in one pixel than that when the subject is not inclined, and is underestimated as a result of calculation in pixel unit. Therefore, the weight calculating unit 152 increases W_{i }as the inclination of the subject increases, to complement it, based on Equation (4).
Subsequently, the weight calculating unit 152 calculates a weight according to a distance and inclination by using W_{i }complemented to Equation (5). The barycenter calculating unit 153 calculates barycentric coordinates of the subject using a value that is obtained by multiplying coordinates of a pixel by a weight corresponding to the pixel. This is to calculate the barycentric coordinates of the subject after equalizing the data amount of the pixel P_{B }corresponding to the distance and inclination of the region B and the data amount of the pixel P_{A }that corresponds to the distance and inclination of the region A. Thus, the barycenter calculating unit 153 is enabled to calculate the barycentric coordinates corresponding to a subject region accurately.
It has been explained that the quadricsurface fitting unit 151 performs fitting to a quadric surface from threedimensional shape data of a subject. However, not limited thereto, the quadricsurface fitting unit 151 may perform, at fitting, weighting according to a physical length of a subject by processing similar to the second calculation processing. At this time, the quadricsurface fitting unit 151 can ignore an item of inclination of the subject, that is, an item n_{i}·e_{z }of inclination in Equation (4). This is applicable when the subject is substantially planar. In this case, the quadricsurface fitting unit 151 calculates a weight of each point x_{i }by using Equation (5), substituting the height Z_{i }of a point on the subject into W_{i }instead of Equation (4). The quadricsurface fitting unit 151 performs quadric surface fitting by using a weight of each point x_{i }as in an example below. Note that it may be configured to use a coefficient W_{i}′=W_{ix }as weighting. While Equation (4) is weight correction according to a physical length, W_{i}′ is weighting according to an area.
As one example, the quadricsurface fitting unit 151 defines an equation of a quadric surface as Equation (7).
{circumflex over (z)}=ax^{2}+by^{2}+cxy+dx+ey+f (7)
The quadricsurface fitting unit 151 acquires a coefficient A that minimizes a squared error between the defined equation of a quadric surface and measured data z of a quadric surface obtained by fitting from threedimensional shape data. The coefficient A to be acquired is (a, b, c, d, e, f)^{T}. That is, the quadricsurface fitting unit 151 acquires the coefficient A that is expressed by Equation (8) below. N refers to the total number of pixels, and w_{k }refers to a weight according to height on the subject that corresponds to a kth pixel.
A vector of the coefficient A is acquired as in Equation (9). A matrix Z is measured data (z_{1}, z_{2}, . . . , z_{N})^{T}. A matrix W is a diagonal matrix in which weights of respective points are set diagonally, and is expressed by Equation (10). A matrix X is indicated by Equation (11), and is expressed by coordinates of N pieces of pixels.
The quadricsurface fitting unit 151 calculates a weighted quadric surface by substituting values of respective items of the acquired coefficient A into corresponding items in Equation (7). Weighted quadricsurface fitting according to the first embodiment is explained referring to FIG. 4. FIG. 4 explains the weighted quadricsurface fitting according to the first embodiment. As depicted in FIG. 4, the quadricsurface fitting unit 151 acquires the height Z_{i }(i refers to a pixel) of the subject from SFS data of the subject. The quadricsurface fitting unit 151 fits the subject into a quadric surface based on the height Z_{i }of the subject. The quadricsurface fitting unit 151 uses a weight according to the height Z_{i }at fitting, to fit the subject into a weighted quadric surface. Thus, the quadricsurface fitting unit 151 is enabled to acquire a quadric surface corresponding to a subject region accurately.
Flowchart of Barycenter Calculation Processing
FIG. 5 is a flowchart of barycenter calculation processing according to the first embodiment. In FIG. 5, the threedimensionalshape acquiring unit 14 acquires SFS data as threedimensional shape data of a subject. Moreover, the weight calculating unit 152 and the barycenter calculating unit 153 calculate barycentric coordinates by the second calculation processing. Furthermore, N refers to the total number of pixels.
As depicted in FIG. 5, the quadricsurface fitting unit 151 acquires the height (distance) Z_{i }(i=1 to N) of the subject from the SFS data acquired by the threedimensionalshape acquiring unit 14 (step S11). For example, the quadricsurface fitting unit 151 calculates the height Z_{i }from the reference position to the subject corresponding to each pixel by using SFS of the subject.
The quadricsurface fitting unit 151 fits the acquired SFS data into a quadric surface (step S12).
The weight calculating unit 152 calculates a unit normal vector N_{i }at a point on the subject corresponding to an ith point from an equation of the quadric surface obtained by fitting (step S13). The unit normal vector e_{z }in the zaxis direction is (0, 0, 1).
The weight calculating unit 152 calculates the weight w_{i }of each point (step S14). For example, the weight calculating unit 152 calculates a weight according to a height and inclination of the subject corresponding to each pixel. The weight calculating unit 152 calculates a weight corresponding to each pixel by using Equation (4) and Equation (5).
Subsequently, the barycenter calculating unit 153 calculates barycentric coordinates using the weight w_{i }(step S15). For example, the barycenter calculating unit 153 calculates barycentric coordinates of the subject by using Equation (6). Thereafter, if the calculated barycentric coordinates significantly deviate from the center of an imaging region, it may be arranged to guide the position of the subject.
As described, according to the first embodiment, the image processing apparatus 1 divides an image of a subject into multiple pixels. The image processing apparatus 1 acquires a distance from a reference position to the subject that corresponds to each pixel. The image processing apparatus 1 sets a weight of each pixel according to the distance. The image processing apparatus 1 corrects information relating to the size of the subject projected in each pixel based on the weight. The image processing apparatus 1 calculates a posture of the subject based on the corrected information of the size of the subject. According to the configuration, the image processing apparatus 1 can grasp the posture of a subject accurately.
Moreover, according to the first embodiment, the image processing apparatus 1 calculates a barycenter of a subject based on the corrected information of the size of the subject. According to the configuration, the image processing apparatus 1 can calculate a barycenter of a subject accurately.
Furthermore, according to the first embodiment, the image processing apparatus 1 fits an image of a subject into a quadric surface based on corrected information of the size of the subject. According to the configuration, the image processing apparatus 1 can fit an image of a subject into a quadric surface accurately.
[b] Second Embodiment
The image processing apparatus 1 of the first embodiment calculates an area on a subject corresponding to one pixel on an image by using threedimensional shape data of the subject, and calculates a weight according to the calculated area. The image processing apparatus 1 further calculates barycentric coordinates of the subject using the weight according to the area. That is, the image processing apparatus 1 has been explained as to calculate barycentric coordinates by using the first calculation method of barycentric coordinates. Moreover, the image processing apparatus 1 calculates a distance from a reference position to the subject projected in one pixel and inclination, and calculates a weight according to the calculated distance and inclination. The image processing apparatus 1 further calculates barycentric coordinates of the subject using the weight according to the distance and inclination. That is, the image processing apparatus 1 has been explained as to calculate barycentric coordinates by using the second calculation method of barycentric coordinates. In a second embodiment, a case in which the image processing apparatus 1 uses the calculation methods of barycentric coordinates in gesture authentication is explained.
Configuration of Image Processing Apparatus According to Second Embodiment
FIG. 6 is a functional block diagram depicting a configuration of an image processing apparatus according to the second embodiment. Identical reference symbols are assigned to identical components to those of the image processing apparatus 1 depicted in FIG. 1, and duplicated explanation of configuration and operation is omitted. The second embodiment differs from the first embodiment in that a gesture authentication unit 21 is added.
The gesture authentication unit 21 performs gesture authentication using barycentric coordinates of a subject that are calculated by the image calculating unit 15.
Use of Calculation Processing of Barycentric Coordinates
The calculation processing of barycentric coordinates that is used in the gesture authentication is explained referring to FIG. 7. FIG. 7 explains the calculation processing of barycentric coordinates used in the gesture authentication. In FIG. 7, a case of calculating barycentric coordinates by using the first calculation method is explained.
As depicted in FIG. 7, an imaging region to be imaged by the imaging unit 12 is depicted. A central portion in the imaging region is the subject. In this example, the subject is a finger.
The image calculating unit 15 calculates barycentric coordinates of a subject as follows. In this example, explanation is given taking a pixel P_{C }and pixel P_{D }of the imaging device 122 as an example. In the pixel P_{C}, a central portion of a finger, that is, a region that faces straight to the camera is projected. In the pixel P_{D}, a side portion of the finger, that is, an inclined region is projected. Therefore, the region projected in the pixel P_{D }has a larger imaged area than the region projected in the pixel P_{C}. That is, the data amount of the pixel P_{D }is underestimated compared to the data amount of the pixel P_{C}. Therefore, the weight calculating unit 152 calculates an area on the finger that corresponds to one pixel on the image by using threedimensional shape data of the finger, and calculates a weight according to the area by using Equation (2). Subsequently, the barycenter calculating unit 153 calculates barycentric coordinates of the subject using Equation (3) by substituting the weight according to the area and coordinates of the pixel into Equation (3).
It has been explained that the weight calculating unit 152 calculates an area on the subject that corresponds to one pixel by using threedimensional shape data of a subject. Alternatively, the threedimensionalshape acquiring unit 14 can measure an area on the subject that corresponds to one pixel from a threedimensional shape. Furthermore, the threedimensionalshape acquiring unit 14 can be a threedimensionalshape measuring device that is integrated in or attached externally to the image processing apparatus 1. When the threedimensionalshape acquiring unit 14 is a threedimensionalshape measuring device, the threedimensionalshape acquiring unit 14 is enabled to calculate an area on the subject that corresponds to one pixel speedily.
[c] Third Embodiment
The image processing apparatus 1 of the first embodiment calculates an area on a subject corresponding to one pixel on an image by using threedimensional shape data of the subject, and calculates a weight according to the calculated area. The image processing apparatus 1 further calculates barycentric coordinates of the subject using the weight according to the area. That is, the image processing apparatus 1 has been explained as to calculate barycentric coordinates by using the first calculation method of barycentric coordinates. Moreover, the image processing apparatus 1 calculates a distance from a reference position to the subject projected in one pixel and inclination, and calculates a weight according to the calculated distance and inclination. The image processing apparatus 1 further calculates barycentric coordinates of the subject using the weight according to the distance and inclination. That is, the image processing apparatus 1 has been explained as to calculate barycentric coordinates by using the second calculation method of barycentric coordinates. In a third embodiment, a case in which the image processing apparatus 1 uses the second calculation method of barycentric coordinates in palm vein authentication is explained.
Configuration of Image Processing Apparatus According to Third Embodiment
FIG. 8 is a functional block diagram depicting a configuration of an image processing apparatus according to the third embodiment. Identical reference symbols are assigned to identical components to those of the image processing apparatus 1 depicted in FIG. 1, and duplicated explanation of configuration and operation is omitted. The third embodiment differs from the first embodiment in that a database unit 31, a feature extracting unit 32, and a collation processing unit 33 are added. Moreover, the third embodiment differs from the first embodiment in that a curvature calculating unit 154 is added to the image calculating unit 15. The third embodiment differs from the first embodiment in that the weight calculating unit 152 is replaced with a weight calculating unit 152A. Note that although the quadricsurface fitting unit 151 can perform either weighted quadricsurface fitting or nonweighted quadricsurface fitting, a case of weighted quadricsurface fitting is explained herein.
The database unit 31 stores data that is used for palm vein authentication. As one example, the database unit 31 stores registration templates. In the registration template, feature data of a palm of each user is registered.
The feature extracting unit 32 extracts feature data of a palm that is used for palm authentication from a captured image.
The collation processing unit 33 collates the feature data extracted by the feature extracting unit 32 and the feature data registered in the registration template, and outputs a similarity degree.
The weight calculating unit 152A calculates a normal vector on the subject from an equation of a quadric surface that is obtained by fitting by the quadricsurface fitting unit 151. That is, the weight calculating unit 152A calculates n_{i }in Equation (4). n_{i }is a unit normal vector at a point on a subject that corresponds to the ith pixel.
As one example, an equation of a quadric surface obtained by fitting is expressed by Equation (12).
Z=aX^{2}+bY^{2}+cXY+dX+eY+f (12)
Expressing a point P on the quadric surface with two variables, u and v, a normal vector n at the point P is acquired as in Equation (13). P_{u }refers to partial differentiation by the variable u, and P_{v }revers to partial differentiation by the variable v.
Suppose the two variables u and v are an X coordinate and a Y coordinate on a curved surface, respectively. As (X, Y, Z) is a point of a quadric surface, and Z is expressed by Equation (12), following Equation (14) is obtained.
P=(X,Y,Z)=(u,v,au^{2}+bv^{2}+cuv+du+ev+f) (14)
P_{u }and P_{v }at the point (X, Y, Z) are acquired as in Equation (15).
The weight calculating unit 152A calculates the normal vector n at each point P(X, Y, Z) on the quadric surface as expressed in Equation (16) by using Equation (13) and Equation (15).
For the acquired normal vector at each point P and the predetermined unit normal vector in the zaxis direction, the weight calculating unit 152A calculates a weight according to a distance and inclination by using Equation (4) and Equation (5).
Referring back to FIG. 8, the curvature calculating unit 154 calculates a curvature on the subject from the equation of the quadric surface obtained by fitting by the quadricsurface fitting unit 151.
As an example, the equation of the quadric surface is expressed by Equation (12). Suppose the two variables u and v are an X coordinate and a Y coordinate on a curved surface, respectively. As (X, Y, Z) is a point of a quadric surface, following Equation (14) is obtained. Furthermore, A distance to a point Q that is close to the point P is expressed by following Equation (17). E, F, G in Equation (17) are expressed by Equation (18) below.
ds^{2}=Edu^{2}+Fdudv+Gdv^{2} (17)
E={right arrow over (P)}_{u}·{right arrow over (P)}_{u }F={right arrow over (P)}_{u}·{right arrow over (P)}_{v }G={right arrow over (P)}_{v}·{right arrow over (P)}_{v} (18)
In approximation of the quadric surface, E, F, G are expressed by Equation (19) below.
E=1+(2au+cv+d)^{2 }
F=(2au+cv+d)(2bv+cu+e)
G=1+(2bv+cu+e)^{2} (19)
With such a precondition, the curvature calculating unit 154 acquires a Gaussian curvature and a mean curvature on the quadric surface. First, the curvature calculating unit 154 acquires values L, M, and N expressed by Equation (20).
L={right arrow over (P)}_{uu}·{right arrow over (n)} M={right arrow over (P)}_{uv}·{right arrow over (n)} N={right arrow over (P)}_{vv}·{right arrow over (n)} (20)
Furthermore, on the quadric surface, P_{uu}, P_{uv}, and P_{vv }are acquired by Equation (21) below.
The curvature calculating unit 154 then substitutes acquired P_{uu}, P_{uv}, and P_{vv }into Equation (20), to acquire L, M, and N by Equation (22) below.
Using L, M, and N in Equation (22), the curvature calculating unit 154 acquires a Gaussian curvature K (=k1·k2) and a mean curvature H (=(k1+k2)/2) as in Equation (23) below. Note that k1 and k2 refer to main curvatures.
That is, the curvature calculating unit 154 acquires a Gaussian curvature K_{i }at each point, and acquires a mean value X_{K }of Gaussian curvatures as in Equation (24) below. Note that a weight w_{i }may be a weight that is acquired by the weight calculating unit 152A.
In addition, the curvature calculating unit 154 performs guidance processing when the mean value X_{K }of Gaussian curvatures is larger than a predetermined threshold. This is performed based on the estimation that the palm is curled. The predetermined threshold may be a value enabling to estimate that a palm is curled. In such a case, the curvature calculating unit 154 guides to open the palm. For example, the curvature calculating unit 154 outputs a message, “Please hold your hand open”. When the mean value X_{K }of Gaussian curvatures is smaller than the predetermined threshold, the curvature calculating unit 154 proceeds to authentication processing. This is because it is estimated that the hand is open.
Flowchart of Palm Vein Authentication Processing
FIG. 9 is a flowchart of the palmvein authentication processing according to the third embodiment. In FIG. 9, the threedimensionalshape acquiring unit 14 acquires SFS data as threedimensional shape data of a subject. N refers to the total number of pixels.
As depicted in FIG. 9, the quadricsurface fitting unit 151 acquires a height (distance) Z_{i }(i=1 to N) of the subject from acquired SFS data (step S31). For example, the quadricsurface fitting unit 151 calculates the height Z_{i }from a reference position to the subject corresponding to each pixel by using the SFS of the subject.
The quadricsurface fitting unit 151 fits the acquired SFS data into a quadric surface (step S32). For example, the quadricsurface fitting unit 151 performs weighted quadricsurface fitting using a height of the subject as a weight.
Subsequently, the curvature calculating unit 154 calculates the curvature K_{i }at each point on the subject that corresponds to an ith point of pixel from an equation of the quadric surface obtained by fitting by the quadricsurface fitting unit 151 (step S33).
The weight calculating unit 152A calculates the weight w_{i }at each point (step S34). For example, the weight calculating unit 152A calculates a normal vector of each point on the quadric surface from the quadric surface obtained by fitting by the quadricsurface fitting unit 151. For the calculated normal vector at each point and a predetermined unit normal vector in the zaxis direction, the weight calculating unit 152A calculates a weight by using Equation (4) and Equation (5).
The barycenter calculating unit 153 calculates barycentric coordinates using the weight w_{i }at each point (step S35). For example, the barycenter calculating unit 153 calculates barycentric coordinates of the subject by using Equation (6).
Subsequently, the curvature calculating unit 154 calculates a curvature mean K from the curvatures K_{i }of the respective points using w_{i }as a weight (step S36). For example, the curvature calculating unit 154 calculates the curvature mean K by using Equation (24) for wi and the curvature K_{i}.
Subsequently, the barycenter calculating unit 153 determines whether the calculated barycentric coordinates are out of a range of a threshold α from the center of the captured image (step S37). When determining that the barycentric coordinates are out of the range of the threshold α from the center of the captured image (step S37: YES), the barycenter calculating unit 153 proceeds to step S39 to perform the guidance processing. This is because it has been determined to be deviated significantly from the center of the imaging region. The threshold α may be a value enabling to estimate that coordinates are approximated to predetermined coordinates.
On the other hand, when determining that the barycentric coordinates are within the range of the threshold α from the center of the captured image (step S37: NO), the curvature calculating unit 154 determines whether the curvature mean K is larger than a threshold Th (step S38). When determining that the curvature mean K is larger than the threshold Th (step S38: YES), the curvature calculating unit 154 proceeds to step S39 to perform the guidance processing. This is because it has been determined that the palm is curled.
At step S39, the barycenter calculating unit 153 or the curvature calculating unit 154 performs the guidance processing (step S39). Thereafter, the palmvein authentication processing is once ended.
On the other hand, when determining that the curvature mean K is not larger than the threshold Th (step S38: NO), the curvature calculating unit 154 causes the feature extracting unit 32 and the collation processing unit 33 to start the authentication processing (step S40). When the authentication processing is finished, the palmvein authentication processing is ended.
As described, according to the third embodiment, an image processing apparatus 1B calculates a distance from a reference position to a palm that corresponds to each pixel and inclination thereof. The image processing apparatus 1B sets a weight of each pixel according to the distance and inclination. Moreover, the image processing apparatus 1B calculates curvatures of the palm corresponding to respective pixels based on threedimensional shape of the palm. The image processing apparatus 1B calculates a mean value of curvatures using a weight. The image processing apparatus 1B calculates barycentric coordinates using the weight. According to the configuration, using the calculated barycentric coordinates, the image processing apparatus 1B can guide a palm to a position suitable for authentication, and can perform palm vein authentication accurately. Using the calculated mean value of curvatures, the image processing apparatus 1B can guide the palm to a posture suitable for authentication, and can perform palm vein authentication accurately. Particularly, in the case of palm vein authentication, as a distance between the imaging unit 12 and a palm is, for example, about 5 centimeters to be short, influence of inclination of the palm is large. That is, depending on the inclination of a palm, the amounts of information of the palm to be imaged into pixels become uneven. Therefore, the image processing apparatus 1B is enabled to perform palm vein authentication accurately by correcting the unevenness.
Although the image processing apparatus 1B has been explained to be used for palm vein authentication as one example, not limited thereto, the image processing apparatus 1B can be used for other kinds of biometric authentication such as iris authentication, besides palm vein authentication.
Others
The respective illustrated components of the image processing apparatus 1 needs not necessarily be physically configured as illustrated. That is, a specific mode of distribution and integration of the image processing apparatus 1 is not limited to the one illustrated, and all or a part thereof can be configured to be functionally or physically distributed or integrated in an arbitrary unit. For example, the weight calculating unit 152 and the barycenter calculating unit 153 may be integrated into one unit. Moreover, the quadricsurface fitting unit 151 may be divided into a functional unit of fitting a subject into a quadric surface and a functional unit of calculating a height from the quadric surface obtained by fitting. Furthermore, the storage unit 13 may be configured to be connected to the image processing apparatus 1 through a network as an external device.
Moreover, the respective processing explained in the above embodiments can be implemented by executing a program prepared in advance by a computer such as a personal computer and a work station. Therefore, in the following, one example of a computer that executes an image processing program that implements similar functions to those of the image processing apparatus 1 depicted in FIG. 1 is explained. FIG. 10 depicts one example of the computer that executes the image processing program.
As depicted in FIG. 10, a computer 200 includes a central processing unit (CPU) 203 that performs various kinds of calculation processing, an input device 215 that accepts an input of data from a user, and a display control unit 207 that controls a display device 209. Furthermore, the computer 200 includes a derive device 213 that reads a program and the like from a recording medium, and a communication control unit 217 that communicates data with other computers through a network. Moreover, the computer 200 includes a memory 201 that temporarily stores various kinds of information, and a hard disk drive (HDD) 205. The memory 201, the CPU 203, the HDD 205, the display control unit 207, the drive device 213, the input device 215, and the communication control unit 217 are connected to each other through a bus 219.
The drive device 213 is a device used for, for example, a removable disk 211. The HDD 205 stores an image processing program 205a and imageprocessing related information 205b.
The CPU 203 reads the image processing program 205a, develops on the memory 201, and executes as processes. The processes correspond to the respective functional units of the image processing apparatus 1. The imageprocessing related information 205b corresponds to the storage unit 13. Furthermore, for example, the removable disk 211 stores various kinds of information such as the image processing program 205a.
The image processing program 205a needs not necessarily be stored in the HDD 205 from the beginning. For example, the program can be stored in a “portable physical medium” such as a flexible disk (FD), a compactdisk readonly memory (CDROM), a digital versatile disk (DVD), a magnetooptical disk, and an integrated circuit (IC) card, and the computer 200 can read the image processing program 205a from such a medium to execute the program.
According to one embodiment, statistical calculation for a subject region can be accurately performed in image processing.
All examples and conditional language recited herein are intended for pedagogical purposes of aiding the reader in understanding the invention and the concepts contributed by the inventor to further the art, and are not to be construed as limitations to such specifically recited examples and conditions, nor does the organization of such examples in the specification relate to a showing of the superiority and inferiority of the invention. Although the embodiments of the present invention have been described in detail, it should be understood that the various changes, substitutions, and alterations could be made hereto without departing from the spirit and scope of the invention.
Great research starts with great data.
Use the most comprehensive innovation intelligence platform to maximise ROI on research.
More Patents & Intellectual Property
 Method for measuring depth of field and image pickup device using same
 Enterprise Patent & IP Solutions
 Improve R&D Innovation
 Intellectual Property (IP) Tools
 IP & Patent Strategies
 Market Intelligence for Innovation
 IP Data API
 Chemical Structure Search
 DNA Sequence Search
 Free Intellectual Property Courses
 IP & Patent Glossary