WO2017202191A1

WO2017202191A1 - Facial data measurement method and system

Info

Publication number: WO2017202191A1
Application number: PCT/CN2017/083358
Authority: WO
Inventors: 谭军
Original assignee: 广州帕克西软件开发有限公司
Priority date: 2016-05-27
Filing date: 2017-05-05
Publication date: 2017-11-30
Also published as: CN106022281A

Abstract

Provided in the present invention is a facial data measurement method. The method comprises the following steps: generating spatial coordinates of N facial feature point groups for a facial image using a preset mapping matrix, N being a positive integer; calculating an arithmetic mean of the spatial coordinates of the N facial feature point groups; and outputting the calculated spatial coordinates of the facial feature point groups. Also provided is a system for use with the method. By generating spatial coordinates of multiple facial feature point groups for a facial image using a preset mapping matrix, the present invention achieves automatic measurement of facial data, with a high accuracy in measured facial data. In addition, facial data can be measured only using a facial image obtained by a camera, so that the costs are low, and the method can be readily adopted.

Description

Facial data measuring method and system

Technical field

The present invention relates to the field of information processing technologies, and more particularly to a method and system for measuring facial data.

Background technique

Facial recognition is a biometric recognition technology based on human facial feature information. It is widely used in community security, network video surveillance, entry and exit management testing, employee attendance and home entertainment, etc. Fast and accurate measurement of facial data in an image or video stream is an important aspect of facial recognition technology.

At present, the traditional measurement method of human head data (ie, facial data) is roughly divided into the following two types: the first is to directly measure the head data through the ruler, and the second is to obtain the three-dimensional model of the human head through the depth sensor, and then indirectly measure The data of the human head model. Among them, the first measurement scheme can measure less data, and has strong professionalism, requires professional personnel to operate, the operation process is complicated, and the measurement process takes a long time; the second measurement scheme can measure more data, and has high precision and can be fully automatic. Measurement, but the measurement takes a long time, and the hardware cost is high, it is difficult to popularize, and it is widely used.

Summary of the invention

The technical problem to be solved by the present invention is to provide a face data measuring method and system for the above-mentioned defects of the human head data measurement in the prior art.

The technical solution adopted by the present invention to solve the above problems is to provide a method for measuring facial data, the method comprising the following steps:

Generating, by using a preset mapping matrix, a spatial coordinate of the N sets of facial feature point groups, and N is a positive integer;

Calculating an arithmetic mean of spatial coordinates of the N sets of facial feature point groups;

The spatial coordinates of the calculated facial feature point group are output.

In the above facial data measuring method, the step of generating a spatial coordinate of a set of facial feature point groups by using the face image through the mapping matrix comprises:

Obtaining the facial image;

Acquiring the facial image by the SURF algorithm to obtain a facial feature point group of the facial image;

Positioning the facial feature point group by an ASM algorithm to obtain pixel coordinates of the facial feature point group;

Mapping the pixel coordinates of the facial feature point group through the mapping matrix to obtain spatial coordinates of the facial feature point group.

In the above-described face data measuring method, before the step of generating the spatial coordinates of the N sets of facial feature point groups by the face image through the mapping matrix, the method further includes:

The mapping matrix is generated according to the acquired calibration image.

In the above facial data measurement method, the step of generating the mapping matrix according to the acquired calibration image includes:

Obtaining the calibration image, the calibration image including a face and a credit card;

Identifying an actual distance of the outer corner of the eye according to an outer corner pixel distance in the face and a credit card pixel width of the credit card;

The face in the calibration image is linearly scaled according to the actual distance of the outer corner of the eye to obtain a mapping matrix between the pixel coordinates of the face and the world coordinates.

In the above face data measuring method, N is greater than or equal to 6.

The invention also provides a facial data measuring system, the system comprising:

a generating module, configured to generate a spatial coordinate of the N sets of facial feature point groups by using a preset mapping matrix, and N is a positive integer;

a calculation module, configured to calculate an arithmetic mean value of spatial coordinates of the N sets of facial feature point groups;

And an output module, configured to output the calculated spatial coordinates of the facial feature point group.

In the above facial data measurement system, the generating module includes:

a first acquiring unit, configured to acquire the facial image;

a feature point acquiring unit, configured to acquire the facial image by using a SURF algorithm Facial feature point group;

a positioning unit, configured to locate the facial feature point group by using an ASM algorithm to obtain pixel coordinates of the facial feature point group;

a mapping unit, configured to map pixel coordinates of the facial feature point group by using the mapping matrix to obtain spatial coordinates of the facial feature point group.

In the above facial data measurement system, the system further includes:

And a preset module, configured to generate the mapping matrix according to the acquired calibration image.

In the above facial data measurement system, the preset module includes:

a second acquiring unit, configured to acquire the calibration image, where the calibration image includes a face and a credit card;

An identification unit, configured to identify an actual distance of an outer corner of the eye according to an outer corner pixel distance in the face and a credit card pixel width of the credit card;

And a calibration unit, configured to linearly calibrate the face in the calibration image according to an actual distance of an outer corner of the eye to obtain a mapping matrix between pixel coordinates and world coordinates of the face.

In the above facial data measuring system, N is greater than or equal to 6.

The beneficial effects of the facial data measuring method and system of the present invention are as follows: the present invention automatically generates facial data by generating a spatial coordinate of a plurality of sets of facial feature point groups through a preset mapping matrix, and accurate measurement of the facial data. High sex. At the same time, the facial data can be measured after obtaining the facial image through the camera, the cost is low, and it is easy to popularize, and can be widely applied to home entertainment and the like.

DRAWINGS

The present invention will be further described below in conjunction with the accompanying drawings and embodiments, in which:

1 is a flow chart of a first embodiment of a method for measuring facial data according to the present invention;

2 is a flow chart of generating spatial coordinates of a set of facial feature point groups by a preset mapping matrix according to a preferred embodiment of the present invention;

3 is a flow chart of a second embodiment of a method for measuring facial data according to the present invention;

4 is a flowchart of generating a preset mapping matrix according to the acquired calibration image according to a preferred embodiment of the present invention;

Figure 5 is a schematic structural view of Embodiment 1 of the face data measuring system of the present invention;

6 is a schematic structural diagram of a generation module in FIG. 5;

Fig. 7 is a schematic structural view of a second embodiment of the face data measuring system of the present invention.

detailed description

The invention realizes automatic measurement of facial data by generating spatial coordinates of a plurality of sets of facial feature point groups by using a preset image through a preset mapping matrix, and the accuracy of the measured facial data is high. At the same time, the facial data can be measured after the facial image is acquired by the camera, which is low in cost and easy to popularize.

The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

As shown in FIG. 1, it is a flowchart of Embodiment 1 of the face data measuring method of the present invention. In this embodiment, referring to FIG. 1, the measurement method includes the following steps:

In step S102, the facial image is generated by using a preset mapping matrix to generate spatial coordinates of the N sets of facial feature point groups;

In this step, N is a positive integer. Specifically, as shown in FIG. 2, the step of generating spatial coordinates of a set of facial feature point groups by using a preset mapping matrix includes: acquiring a facial image in step S1021; the facial image may be a photo file directly captured by the camera Also, it is an image file that can be captured in a video file. In step S1022, the facial image is obtained by the SURF (Speed-up robust features) algorithm, and the facial feature point group includes 86 feature points. In step S1023, the facial feature point group is positioned by the ASM (Active Shape Model) algorithm to the pixel coordinates of the facial feature point group. In step S1024, the pixel coordinates of the facial feature point group are mapped by the mapping matrix to obtain the spatial coordinates of the facial feature point group.

In step S104, an arithmetic mean value of the spatial coordinates of the N sets of facial feature point groups is calculated; further, in the embodiment, N is greater than or equal to 6, by generating the spatial coordinates of the set of facial feature point groups as described above. In step, spatial coordinates of 6 sets of facial feature point groups are generated, so that the measured facial data is more accurate.

In step S106, the calculated facial feature point space coordinate group is output. In this embodiment, the pass Through the above steps, the face data is automatically measured, and the processing speed of the measurement is high, and the measurement of the face data can be completed in 1 second.

As shown in FIG. 3, it is a flowchart of Embodiment 2 of the face data measuring method of the present invention. In this embodiment, the difference from the face data measurement method of the embodiment is that, before the step S102, the method further includes: in step S100, generating a preset mapping matrix according to the acquired calibration image, specifically, as shown in FIG. 4, comprising: in step S1001, acquiring a calibration image, wherein the calibration image includes a face and a credit card; and in step S1002, identifying an actual distance of the outer corner of the eye based on the outer corner pixel distance in the face and the credit card pixel width of the credit card. Wherein, the distance between the outer corner pixel and the credit card pixel width of the credit card can be obtained by an image detection algorithm, and the calculation formula of the outer corner pixel distance is: the actual distance of the outer corner of the eye = (outer corner pixel distance / credit card pixel width) * the actual width of the credit card The actual standard credit card has an actual width of 54mm. In step S1003, the face in the calibration image is linearly scaled according to the actual distance of the outer corner of the eye to obtain a mapping matrix between the pixel coordinates of the face and the world coordinates.

FIG. 5 is a schematic structural view of Embodiment 1 of the face data measuring system of the present invention. In the present embodiment, referring to FIG. 5, the measurement system 100 includes a generation module 102, a calculation module 104, and an output module 106.

The generating module 102 is configured to generate a spatial coordinate of the N sets of facial feature point groups by using a preset mapping matrix. Where N is a positive integer. Specifically, as shown in FIG. 6 , the generating module 102 includes a first acquiring unit 1021 , a feature point acquiring unit 1022 , a positioning unit 1023 , and a mapping unit 1024 , where the first acquiring unit 1021 is configured to acquire a facial image; The image file that is taken directly by the camera is also the image file that can be captured in the video file. The feature point obtaining unit 1022 is configured to acquire the facial feature point group of the facial image by using the SURF algorithm; the facial feature point group includes 86 feature points. The positioning unit 1023 is configured to locate the facial feature point group by the ASM algorithm to the pixel coordinates of the facial feature point group. The mapping unit is configured to map the pixel coordinates of the facial feature point group through the mapping matrix to obtain the spatial coordinates of the facial feature point group.

The calculation module 104 is configured to calculate an arithmetic mean value of the spatial coordinates of the N sets of facial feature point groups; further, in the embodiment, N is greater than or equal to 6, and the spatial coordinates of the 6 sets of facial feature point groups are generated by the generating module 102. In this way, the measured facial data is more accurate.

The output module 106 is configured to output the calculated facial feature point space coordinate group. In the present embodiment, the face data is automatically measured by the above steps, and the processing speed of the measurement is high, and the measurement of the face data can be completed in 1 second.

As shown in FIG. 7, it is a flowchart of Embodiment 2 of the face data measuring system of the present invention. In this embodiment, the difference from the facial data measurement system of the embodiment is that the measurement system 100 further includes a preset module 101, and the preset module 101 is configured to generate a preset mapping according to the acquired calibration image. The matrix, specifically, the preset module 101 includes a second obtaining unit 1011, an identifying unit 1012, and a calibrating unit 1013, wherein the second acquiring unit 1011 is configured to acquire a calibration image, wherein the calibration image includes a face and a credit card; and the identifying unit 1012 For identifying the actual distance of the outer corner of the eye based on the outer corner pixel distance in the face and the credit card pixel width of the credit card. Wherein, the distance between the outer corner pixel and the credit card pixel width of the credit card can be obtained by an image detection algorithm, and the calculation formula of the outer corner pixel distance is: the actual distance of the outer corner of the eye = (outer corner pixel distance / credit card pixel width) * the actual width of the credit card The actual standard credit card has an actual width of 54mm. The calibration unit 1013 is configured to linearly calibrate the face in the calibration image according to the actual distance of the outer corner of the eye to obtain a mapping matrix between the pixel coordinates of the face and the world coordinates.

The above is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily think of changes or within the technical scope disclosed by the present invention. Alternatives are intended to be covered by the scope of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

A method for measuring facial data, characterized in that the method comprises the following steps:

Generating, by using a preset mapping matrix, a spatial coordinate of the N sets of facial feature point groups, and N is a positive integer;

Calculating an arithmetic mean of spatial coordinates of the N sets of facial feature point groups;

The spatial coordinates of the calculated facial feature point group are output.
The method of measuring facial data according to claim 1, wherein the step of generating a spatial coordinate of the set of facial feature point groups by using the mapping image through the mapping matrix comprises:

Obtaining the facial image;

Acquiring the facial image by the SURF algorithm to obtain a facial feature point group of the facial image;

Positioning the facial feature point group by an ASM algorithm to obtain pixel coordinates of the facial feature point group;

Mapping the pixel coordinates of the facial feature point group through the mapping matrix to obtain spatial coordinates of the facial feature point group.
The method of measuring facial data according to claim 1, wherein before the step of generating the spatial coordinates of the N sets of facial feature point groups by using the mapping image, the method further comprises:

The mapping matrix is generated according to the acquired calibration image.
The method of measuring a face data according to claim 3, wherein the step of generating the mapping matrix according to the acquired calibration image comprises:

Obtaining the calibration image, the calibration image including a face and a credit card;

Identifying an actual distance of the outer corner of the eye according to an outer corner pixel distance in the face and a credit card pixel width of the credit card;

The face in the calibration image is linearly scaled according to the actual distance of the outer corner of the eye to obtain a mapping matrix between the pixel coordinates of the face and the world coordinates.
The face data measuring method according to any one of claims 1 to 4, wherein N is greater than or equal to 6.
A facial data measuring system, characterized in that the system comprises:

a generating module, configured to generate a spatial coordinate of the N sets of facial feature point groups by using a preset mapping matrix, and N is a positive integer;

a calculation module, configured to calculate an arithmetic mean value of spatial coordinates of the N sets of facial feature point groups;

And an output module, configured to output the calculated spatial coordinates of the facial feature point group.
The facial data measuring system according to claim 6, wherein the generating module comprises:

a first acquiring unit, configured to acquire the facial image;

a feature point acquiring unit, configured to acquire the facial feature image of the facial image by using the SURF algorithm;

a positioning unit, configured to locate the facial feature point group by using an ASM algorithm to obtain pixel coordinates of the facial feature point group;

a mapping unit, configured to map pixel coordinates of the facial feature point group by using the mapping matrix to obtain spatial coordinates of the facial feature point group.
The facial data measurement system of claim 6 wherein the system further comprises:

And a preset module, configured to generate the mapping matrix according to the acquired calibration image.
The facial data measuring system according to claim 8, wherein the preset module comprises:

a second acquiring unit, configured to acquire the calibration image, where the calibration image includes a face and a credit card;

An identification unit, configured to identify an actual distance of an outer corner of the eye according to an outer corner pixel distance in the face and a credit card pixel width of the credit card;

And a calibration unit, configured to linearly calibrate the face in the calibration image according to an actual distance of an outer corner of the eye to obtain a mapping matrix between pixel coordinates and world coordinates of the face.
A facial data measuring system according to any one of claims 6-9, wherein N is greater than or equal to 6.