TWI693560B - Face recognition method - Google Patents

Abstract

The present invention relates to a face recognition method, in which a face recognition system detects a face of a person to be tested, and then performs a face recognition action; thereby, the face recognition method of the present invention mainly uses depth The image capture unit captures the hardware design of the three-dimensional image of the user's face, and uses the image processing unit to select a plurality of feature points in the face image and the calculation and comparison of the length and depth of the connecting line between the two feature points For the method, it can effectively complete the safe recognition of facial image features, and can further combine the calculation and comparison actions of the triangle area enclosed by the three lines, and indeed achieve the main advantage of safer and more comprehensive face safe recognition.

Description

Face recognition method

The invention relates to a face recognition method, in particular to a safety monitoring method for performing face recognition by using a depth camera to perform facial image capture and compare the connection length and depth of a plurality of feature points.

According to the press, with the rapid development of technology, the formation of a smart city can not only deploy a dense network and a powerful control center, but also effectively fully grasp the distribution line of the crowd to provide rapid and correct social security monitoring of situation prevention. For example, in terms of traffic flow monitoring, a monitoring system is set up on street lights or traffic signs, and even the exchange of information between vehicles. Through a single platform of real-time monitoring and management, it can target scenes such as traffic jams, car accidents, or crowd crimes. Control, which effectively assists in the functions of evacuating car tide or tracking criminal vehicle movements; face recognition system is an emerging part of smart city, especially for people who don’t like to carry a key out, face recognition is a big boon, This is because more and more people's residences are using automatic induction access control systems to replace the original keyhole access control equipment, which does achieve convenience and safety. However, most of the current automatic induction access control systems are The radio frequency identification (RFID) method is adopted, which is mainly carried out by a method equipped with a radio frequency identification reader (RFID reader). The method is to provide the user to carry an object with a radio frequency identification tag (RDID tag) When the user brings the RFID tag close to the RFID reader, the RFID reader can verify whether the RFID tag is frequency-matched to determine whether to let the user enter; however, the main disadvantage of the RFID technology is to recognize the card If you don’t recognize people, there will still be many security concerns.

Face recognition is currently the most secure method of the access control system, which can effectively solve the shortcomings of the above-mentioned radio frequency identification technology; however, the current face recognition is mostly 2D recognition method, which mainly captures the three-dimensional face with a 2D camera into one Planar information, the facial recognition system only looks for identification data in 2D information that can be compared, so that as long as the interested person retrieves photos or plans with relevant face data, they can effectively deceive 2D facial recognition The system also achieves the goal of easy access to the door; therefore, how to effectively combine the plane and three-dimensional information of the face through innovative hardware design to achieve a safer and more comprehensive face safety identification, is still the face Identification and other related industry developers and related researchers need to continuously strive to overcome and solve the problems.

Today, the inventor is considering that the traditional face recognition method still has many deficiencies in the actual implementation, so it is a tireless spirit, and it is improved by its rich professional knowledge and many years of practical experience. , And researched and created the invention accordingly.

The main purpose of the present invention is to provide a face recognition method, in particular to a security monitoring method for face recognition by using a depth camera to perform facial image capture and compare the connection length and depth of a plurality of feature points, It is mainly designed by the depth image capturing unit to capture the three-dimensional image of the user's face, and the image processing unit is used to select a plurality of feature points in the facial image and the length and depth of the connecting line segment between the two feature points The calculation and comparison method can effectively complete the safe identification of facial image features, and can further combine the calculation and comparison actions of the triangle area enclosed by the three lines, and indeed achieve the main advantages of safer and more comprehensive face safe identification By.

In order to achieve the above-mentioned implementation objectives, the present inventors have devised the following implementation techniques, in which the face recognition method of the present invention performs a face recognition action after a face recognition system detects the face of a person to be tested ; First, use a depth image capture unit to capture an image of the original face; then, use a built-in selection module of the image processing unit to select a plurality of first feature points from the original face image, and separately Calculate the first distance between the first feature point and the depth image capture unit; then, use the built-in length calculation module of the image processing unit to connect two or two first feature points to obtain a plurality of first line segments, and respectively Calculate the length of the first line segment to obtain a plurality of first length values, and calculate the first depth difference between the first distance values of the first feature points of the two links; then, the first feature points, The original face image of the first line segment, the first length value, and the first depth difference value is stored in a storage unit; then, repeat the above steps to use the depth image capture unit to capture an image of the face to be measured And obtain a plurality of second feature points corresponding to the position of the first feature point on the face to be tested, a plurality of second line segments connected by the second feature point and a second length value, and a plurality of second distance values and The second depth difference value is stored in the storage unit; thereafter, the first length value and the corresponding second length value are compared using a comparison unit, and the first depth difference value and the corresponding second depth difference value; finally , When the difference between the first length value and the corresponding second length value, and the difference between the first depth difference and the corresponding second depth difference are within an error range, the face to be tested is The original face is the same.

The face recognition method as described above, wherein the depth image capturing unit is one of a device such as a TOF camera or a dual-lens camera.

The face recognition method as described above, wherein when the depth image capturing unit is a TOF camera, the depth image capturing unit at least includes a light emitting module, a receiving module, and an image capturing module.

The face recognition method as described above, wherein the light emitting module emits a plurality of light spots on the original face or the face to be tested.

The face recognition method as described above, wherein the receiving module receives a plurality of reflected light spots reflected by the original face or the face to be measured, and transmits them to the image processing unit.

The face recognition method as described above, wherein the image capturing module includes a first optical lens, and a first photosensitive element connected to the first optical lens, the first optical lens captures the original face or to-be-measured The image of the face is formed on the first photosensitive element.

In the face recognition method described above, the first optical lens is a device such as a fisheye lens, a wide-angle lens, or a standard lens.

In the face recognition method as described above, the first photosensitive element is one of a photocoupler (Charge Coupled Device, CCD) or a complementary metal oxide semiconductor (Complementary Metal-Oxide-Semiconductor, CMOS) and other devices.

The face recognition method as described above, wherein when the depth image capturing unit is a dual-lens camera, the depth image capturing unit includes two second optical lenses and a second light sensor respectively connected to the second optical lenses Component, the second optical lens simultaneously captures the image of the original face or the face to be tested and forms an image on the second photosensitive element.

In the face recognition method as described above, the second optical lens is a device such as a fisheye lens, a wide-angle lens, or a standard lens.

In the face recognition method as described above, the second photosensitive element is one of photoelectric coupling element (CCD) or complementary metal oxide semiconductor (CMOS).

The face recognition method as described above, wherein the error range is between 98% and 100%.

The face recognition method as described above, wherein an area calculation module built in the image processing unit can be further used to calculate the first triangle of the first triangle enclosed by the three first line segments before storing the original face image The area value, and the first area value plus the first total area value.

The face recognition method as described above, wherein after comparing the first length value with the corresponding second length value, the area calculation module can be further used to calculate the second area of the second triangle enclosed by the three second line segments Value, and the second total area value plus the second total area value.

The face recognition method as described above, wherein the area calculation module uses a Hailong formula to calculate the first area value or the second area value.

The face recognition method as described above, wherein after the last step is completed, the first total area value and the second total area value can be further compared using the comparison unit, if the first total area value and the second total area value are When the difference between is within the error range, the face to be tested is the same as the original face.

In this way, the face recognition method of the present invention is mainly designed by the depth image capturing unit to capture the three-dimensional image of the user's face, combined with the image processing unit to select a plurality of feature points and two two features in the face image The calculation and comparison method of the length and depth of the connecting line between points effectively completes the safe identification of facial image features, and can further combine the calculation and comparison of the area of the triangle enclosed by the three lines, which is indeed more secure and The main advantage of a more comprehensive face safety recognition; in addition, the face recognition method of the present invention mainly uses the TOF camera to capture the depth difference between a plurality of feature points such as facial features of the face and the feature points as the face One of the important basis of recognition can effectively solve the shortcomings of the two-dimensional face recognition system that intentional people deceive the two-dimensional face recognition system with photos or floor plans of the same face, and indeed achieve the main advantages of a safer and more difficult face recognition system and method; Furthermore, the face recognition method of the present invention is mainly to compare the depth value of the feature points on the face, the length of the connecting line, and the triangle area enclosed by the connecting line to increase the face recognition The accuracy of the method really achieves the main purpose of accurate recognition to pay attention to safety considerations. Finally, the face recognition method of the present invention mainly uses the depth image capture unit to capture the three-dimensional features of the human face, allowing the user to face at different angles. When capturing the face image of the depth image capturing unit, the face recognition system can accurately give the depth difference value of the facial feature points, which effectively solves the shortcomings of the traditional face recognition system because of the incorrect angle. It does achieve the main advantage of significantly reducing the time of facial recognition.

(1)‧‧‧Face recognition system

(11)‧‧‧Depth image capture unit

(111)‧‧‧Lighting module

(112)‧‧‧Receiving module

(113)‧‧‧Image capture module

(1131)‧‧‧First optical lens

(12)‧‧‧Image processing unit

(121)‧‧‧Select module

(122)‧‧‧Length calculation module

(123)‧‧‧ Area calculation module

(13)‧‧‧Storage unit

(14)‧‧‧Comparison unit

(2)‧‧‧ROI range

(3)‧‧‧ subject

(A)‧‧‧ original face

(A1)‧‧‧First characteristic point

(A2)‧‧‧The first line

(Ae1)‧‧‧First area value

(Ae2)‧‧‧Second area value

(B) ‧‧‧ face to be tested

(B1)‧‧‧Second characteristic point

(B2)‧‧‧Second line

(L1)‧‧‧First length value

(L2)‧‧‧Second length value

(T1)‧‧‧The first triangle

(T2)‧‧‧Second Triangle

(S1)‧‧‧Step 1

(S2)‧‧‧Step 2

(S3)‧‧‧Step 3

(S4)‧‧‧Step 4

(S5)‧‧‧Step 5

(S6)‧‧‧Step 6

(S7)‧‧‧Step 7

Figure 1: Flow chart of the steps of the face recognition method of the present invention

Figure 2: Schematic diagram of the hardware architecture of the face recognition system used in the face recognition method of the present invention

Figure 3: Schematic diagram of feature point selection in a preferred embodiment of the face recognition method of the present invention

Figure 4: Schematic diagram of line connection of a preferred embodiment of the face recognition method of the present invention

Figure 5: Schematic diagram of facial recognition to be tested according to a preferred embodiment of the facial recognition method of the present invention

In order to facilitate your examination committee to understand the technical features, content, advantages and achievable effects of the present invention, the present invention is described in detail in conjunction with the drawings and in the form of expressions of the embodiments, and the drawings used therein, Its purpose is only for illustration and supplementary description, and may not be the true scale and precise configuration after the implementation of the present invention, so it should not be interpreted and limited to the relationship between the scale and the configuration of the attached drawings, and the scope of the present invention in the actual implementation is limited , He Xianming.

First of all, please refer to Figures 1 and 2, which are flowcharts of the steps of the face recognition method of the present invention, and a schematic diagram of the hardware architecture of the face recognition system used, in which the field of the face recognition method of the present invention is applied It can be used, for example, but not limited to, face recognition for entrance and exit. The approach is to first use a face recognition system (1) to create a recognition model for face recognition in a storage unit (13); then, when After the face recognition system (1) detects the face of a person under test (3) in an ROI range (2), the ratio between the face of the person under test (3) and the pre-stored recognition model is activated For the action, if the success rate of the comparison is within an error range, the entrance is opened for the person to be tested (3) to enter; the face recognition method of the present invention mainly includes the following steps: Step 1 (S1): Use a The depth image capture unit (11) captures an image of the original face (A); in addition, the depth image capture unit (11) is a device such as a TOF camera or a dual-lens camera. When the fetching unit (11) is a TOF camera, the depth image capturing unit (11) at least includes a light emitting module (111), a receiving module (112), and an image capturing module (113), wherein The light emitting module (111) emits a plurality of light spots on the original face (A), and the receiving module (112) receives a plurality of reflected light spots reflected by the original face (A) and transmits them to the image processing unit ( 12), and the image capturing module (113) includes a first optical lens (1131), and a first photosensitive element (1132) connected to the first optical lens (1131), and the first optical lens (1131) ) Is to capture the image of the original face (A) and image it on the first photosensitive element (1132), where the first optical lens (1131) is one of fisheye lens, wide-angle lens or standard lens and other devices, and the first The photosensitive element (1132) is one of photoelectric coupling element (CCD) or complementary metal oxide semiconductor (CMOS) and other devices. In a preferred embodiment of the present invention, since face recognition must be established for recognition Therefore, the invention in this case uses the depth image capturing unit (11) presented in the form of a TOF camera to capture the face image of the person under test (3). This face image is called the original face (A) image. , Mainly by the first optical lens (1131) built in the TOF camera and presented in the form of a fisheye lens to capture the image of the original face (A) in the first presented in the form of a photoelectric coupling device (CCD) Imaging on the photosensitive element (1132); step two (S2): using a built-in selection module (121) of the image processing unit (12) to select a plurality of first feature points (A1) from the image of the original face (A) ), and calculate the first distance between the first feature point (A1) and the depth image capture unit (11); please refer to the face recognition method of the present invention shown in Figure 2 and Figure 3 Features of a preferred embodiment Selection schematic diagram, in which the image processing unit (12) is electrically connected to the depth image capturing unit (11), and the selection module (121) built in the image processing unit (12) is the facial features of the original face (A) Set the first feature point (A1) for each of 11 positions, and calculate the first distance between the 11 first feature points (A1) and the depth image capturing unit (11) respectively. This first distance value is The depth value of the first feature point (A1); Step three (S3): Use the built-in length calculation module (122) of the image processing unit (12) to connect two or two first feature points (A1) to obtain a plurality of first Line segment (A2), and calculate the length of the first line segment (A2) to obtain a plurality of first length values (L1), and calculate the first length value of the first feature point (A1) of the two lines ( L1) the first depth difference (De1); please also refer to the line segment connection diagram of a preferred embodiment of the face recognition method of the present invention shown in FIGS. 2 and 4 together, in which image processing is used The built-in length calculation module (122) of the unit (12) connects the two first feature points (A1) with the first line segment (A2), and 11 first feature points (A1) can be connected to each other 20 first line segments (A2), and the length calculation module (122) respectively calculates the first length value (L1) of the 20 first line segments (A2), and then connects the first feature points according to each other (A1) The depth value of each is used to calculate the difference between the two depth values, which is the first depth difference; however, it must be noted that the number of first feature points (A1) and the first feature point (A1) ) The number of the first line segments (A2) connected to each other is for convenience of description, and is not limited to the present invention, and those skilled in the art should know the number of different first feature points (A1) The number of first line segments (A2) connected to the first feature point (A1) as long as the facial recognition system (1) can perform the length difference between the first line segment (A2) and the first depth (De1) The calculations can be regarded as an extension of the technical features of the present invention, and will not reflect the actual implementation of the present invention; Step 4 (S4): will have a first feature point (A1), a first line segment (A2), the first The length value (L1) and the image of the original face (A) of the first depth difference are stored in a storage unit (13); in a preferred embodiment of the invention, the storage unit (13) is electrically Connect the image processing unit (12), the original face (A) image obtained by the above steps, and the first feature point (A1), the first line segment (A2), the first included in the original face (A) The length value (L1) and the first depth difference information are all stored in the storage unit (13), which can be used as a recognition model for face recognition at the entrance and exit; step five (S5): repeat step one (S1) ) To step 3 (S3), use the depth image capturing unit (11) to capture an image of the face (B) to be tested, and obtain a plurality of corresponding first feature points (A1) on the face (B) to be tested Of The second feature point (B1) of the position, the plurality of second line segments (B2) and the second length value (L2) connected by the second feature point (B1), and the plurality of second distance values and the second depth difference Value, and stored in the storage unit (13); please also refer to the face recognition schematic diagram of a preferred embodiment of the face recognition method of the present invention shown in FIG. 2 and FIG. 5 (S1) to step three (S3) after the creation of the original face (A) image recognition model is completed, when the person to be tested (3) wants to enter the entrance and exit and must perform face recognition, the depth image acquisition unit (11) is Capture its facial image, this facial image is called the face (B) image to be tested; then, select the module (121) to the facial features of the face (B) to be tested to correspond to the first feature point ( Set the second feature points (B1) for each of the 11 positions of A1), and calculate the second distance between the 11 second feature points (B1) and the depth image capturing unit (11) respectively. It is the depth value of the second feature point (B1); then, the length calculation module (122) connects the two second feature points (B1) with the second line segment (B2) to each other with 11 second feature points ( The interconnection between B1) can obtain 20 second line segments (B2), and the length calculation module (122) separately calculates the second length values (L2) of the 20 second line segments (B2); then, Calculate the difference between the two depth values according to the depth values of the second feature points (B1) connected to each other, this is the second depth difference; finally, there will be a second feature point (B1), a second line segment (B2), the second length value (L2), and the second depth difference value of the image of the face (B) to be measured are stored in the storage unit (13); Step 6 (S6): using a comparison unit (14 ) Compare the first length value (L1) with the corresponding second length value (L2), and the first depth difference and the corresponding second depth difference; in a preferred embodiment of the present invention, and The image processing unit (12) electrically connected comparison unit (14) performs a comparison operation of the first length value (L1) and the corresponding second length value (L2), and the first depth difference value and the corresponding second For the comparison operation of the depth difference, please refer to Figure 4 and Figure 5 again, between the first characteristic point (A1) at point C and the first characteristic point (A1) at point D in Figure 4 The first length value (L1) of the connected first line segment (A2) can be the same as the second feature point (B1) from point C to the second feature point (B1) of point D in Figure 5 The second length value (L2) of the connected second line segment (B2) is compared with each other; and Step 7 (S7): When the difference between the first length value (L1) and the corresponding second length value (L2), And when the difference between the first depth difference and the corresponding second depth difference is within an error range, the face to be measured (B) is the same as the original face (A); in addition, the error range is between 98%~100%; in a preferred embodiment of the present invention, when the comparison unit (14) compares the first The difference between a length value (L1) and the corresponding second length value (L2), and the difference between the first depth difference and the corresponding second depth difference is between 98% and 100% The face (B) and the original face (A) are the same person's face, then the face recognition is successful and the subject (3) is allowed to enter and exit.

In addition, before storing the original face (A) image in step 4 (S4), an area calculation module (123) built in the image processing unit (12) may be used to calculate the three first line segments (A2). The first area value (Ae1) of the enclosed first triangle (T1), and the first total area value added by the first area value (Ae1); in addition, the area can be further used after step 6 (S6) is completed The calculation module (123) calculates the second area value (Ae2) of the second triangle (T2) enclosed by the three second line segments (B2), and the second total area value added by the second area value (Ae2); In addition, the area calculation module (123) uses a sea dragon formula to calculate the first area value (Ae1) and the second area value (Ae2); furthermore, after the completion of step seven (S7), the comparison unit (123) can be further used ( 14) Compare the first total area value with the second total area value. If the difference between the first total area value and the second total area value is within the error range, the face to be measured (B) is the original face Part (A) is the same; please refer to Figures 4 and 5 again. The first triangle (T1) enclosed by three first line segments (A2) is calculated by the Hailong formula. (Ae1), and the sum of the first area values (Ae1) of the 10 first triangles (T1) enclosed by the 20 first line segments (A2) is the first total area value; the same is applied to the face to be measured Part (B) images can also calculate the second area value (Ae2) and the second total area value, and then continue to compare the first total area value and the second total area value after step 7 (S7). The error is within the range of 98%~100%, which means that the face to be tested (B) and the original face (A) are the same human face, allowing the person to be tested (3) to enter and exit.

Furthermore, when the depth image capturing unit (11) is a dual-lens camera, the depth image capturing unit (11) includes two second optical lenses (not shown in the figure), and one separate from the second optical lens The connected second photosensitive element (not shown in the figure), the second optical lens simultaneously captures the image of the original face (A) or the face to be tested (B) and forms an image on the second photosensitive element; and a preferred implementation thereof In the same example, the dual-lens camera can also obtain the depth values of the first feature point (A1) and the second feature point (B1), which can also achieve the function of face recognition.

It can be seen from the above implementation description that the face recognition method of the present invention has the following advantages compared with the prior art:

1. The face recognition method of the present invention is mainly designed by the depth image capturing unit to capture the three-dimensional image of the user's face, and the image processing unit is used to select a plurality of feature points and two or two feature points in the face image The calculation and comparison method of the length and depth of the connecting line segment can effectively complete the safe identification of facial image features, and can further combine the calculation and comparison of the triangle area enclosed by the three line segments, which is indeed safer and more The main advantage of comprehensive face recognition.

2. The face recognition method of the present invention mainly uses the TOF camera to capture the depth difference between a plurality of feature points and feature points of the facial features of the face, as one of the important basis of face recognition, which can effectively solve the intentional person The use of photos or plans with the same human face to deceive the shortcomings of the two-dimensional face recognition system does indeed achieve the main advantages of a safer and more difficult-to-decipher face recognition system and method.

3. The face recognition method of the present invention is mainly to compare the depth value of the feature points on the face, the length of the connecting line, and the triangle area enclosed by the connecting line, etc. to increase the face recognition. Accuracy does indeed achieve the main purpose of accurate identification to attach importance to safety considerations.

4. The face recognition method of the present invention mainly uses the depth image capturing unit to capture the three-dimensional features of the human face, allowing the user to face the face image of the depth image capturing unit at different angles. Recognition systems can accurately give the depth difference of facial feature points, effectively solving the shortcomings of traditional facial recognition systems due to incorrect angles and misjudgments, and indeed achieving the main advantage of greatly reducing facial recognition time.

In summary, the face recognition method of the present invention can indeed achieve the expected use effect through the disclosed embodiments, and the present invention has not been disclosed before the application. It has fully complied with the provisions of the Patent Law and Claim. I filed an application for a patent for invention in accordance with the law, pleaded for the review, and granted the patent.

However, the illustrations and descriptions disclosed above are only preferred embodiments of the present invention, and are not intended to limit the scope of protection of the present invention; those who are familiar with the art, according to the characteristic scope of the present invention, do other things Equivalent changes or modifications should be regarded as not departing from the design scope of the present invention.

(S1)‧‧‧Step 1

(S2)‧‧‧Step 2

(S3)‧‧‧Step 3

(S4)‧‧‧Step 4

(S5)‧‧‧Step 5

(S6)‧‧‧Step 6

(S7)‧‧‧Step 7

Claims (13)

A face recognition method is to perform a face recognition action after a face recognition system detects the face of a person to be tested. The face recognition method includes the following steps: Step 1: Use a depth image The capture unit captures an image of the original face; Step 2: Use a built-in selection module of the image processing unit to select a plurality of first feature points from the image of the original face and calculate the first feature separately The first distance between the point and the depth image acquisition unit; step 3: use the built-in length calculation module of the image processing unit to connect two or two of these first feature points to obtain a plurality of first line segments, and Calculate the lengths of the first line segments to obtain a plurality of first length values, and calculate the first depth difference between the first distance values of the first feature points of the two connecting lines, and use a built-in The area calculation module of the image processing unit calculates the first area value of the first triangle surrounded by the three first line segments, and the first total area value added up by the first area values; Step 4: will have The images of the original faces of the first feature points, the first line segments, the first length values, and the first depth difference values are stored in a storage unit; Step 5: Repeat step 1 to step Step three, use the depth image capturing unit to capture an image of a face to be tested, and obtain a plurality of corresponding faces on the face to be tested The second feature point at the position of the first feature point, a plurality of second line segments connected by the second feature points and a second length value, and a plurality of second distance values and a second depth difference value, and stored in In the storage unit; Step 6: Use a comparison unit to compare the first length value and the corresponding second length value, and the first depth difference value and the corresponding second depth difference value, while using the area The calculation module calculates the second area value of the second triangle enclosed by the three second line segments, and the second total area value added up by the second area values; and Step 7: When the first length value corresponds to The difference between the second length value and the difference between the first depth difference and the corresponding second depth difference are within an error range, and the first total area value and The second total area value, if the difference between the first total area value and the second total area value is within the error range, the face to be tested is the same as the original face. The face recognition method as described in item 1 of the patent application, wherein the depth image capturing unit is one of a TOF camera or a dual-lens camera. The face recognition method as described in item 2 of the patent application scope, wherein when the depth image capturing unit is a TOF camera, the depth image capturing unit at least includes a light emitting module, a receiving module, and a Image capture module. The face recognition method as described in item 3 of the patent application scope, wherein the light emitting module emits a plurality of light spots on the original face or the face to be tested. The face recognition method as described in item 3 of the patent application scope, wherein the receiving module receives a plurality of reflected light spots reflected by the original face or the face to be measured, and transmits them to the image processing unit. A face recognition method as described in item 3 of the patent application, wherein the image capture module includes a first optical lens, and a first photosensitive element connected to the first optical lens, the first optical lens The image of the original face or the face to be measured is captured and imaged on the first photosensitive element. The face recognition method as described in item 6 of the patent application scope, wherein the first optical lens is one of a fisheye lens, a wide-angle lens, or a standard lens. The face recognition method as described in item 6 of the patent application range, wherein the first photosensitive element is one of a photoelectric coupling element (CCD) or a complementary metal oxide semiconductor (CMOS). The face recognition method as described in item 2 of the patent application scope, wherein when the depth image capturing unit is a dual-lens camera, the depth image capturing unit includes two second optical lenses, and a The second photosensitive element is connected to the second optical lens, and the second optical lens simultaneously captures images of the original face or the face to be tested and forms an image on the second photosensitive element. The face recognition method as described in item 9 of the patent application scope, wherein the second optical lens is one of a fisheye lens, a wide-angle lens, or a standard lens. The face recognition method as described in item 9 of the patent application scope, wherein the second photosensitive element is one of a photoelectric coupling element (CCD) or a complementary metal oxide semiconductor (CMOS). The face recognition method as described in item 1 of the patent application scope, wherein the error range is between 98% and 100%. The face recognition method as described in item 1 of the patent application scope, wherein the area calculation module calculates the first area value or the second area value using a Hailong formula.

Images