Method and System

Abstract

Provided are a face recognition method and system that can perform face recognition with high accuracy regardless of the installation conditions of a face recognition device and the surrounding environment. 【Solution means】In a system where a face recognition device and a computer communicate via a communication network, the face recognition method includes: a step of installing a face recognition device having a camera at a predetermined location; a step of selecting an ND filter having a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters having different transmittances; a step of attaching the ND filter to the lens of the camera; a step of transmitting the captured face image from the face recognition device to the computer; a step of extracting a hash value from the face image data using a predetermined hash function; and a step of performing face recognition on the face image captured by the camera by the face recognition device or the computer based on the extracted hash value and the hash value corresponding to a pre-registered face image.

Description

Technical Field

[0001] The present disclosure relates to a face authentication method and system.

Background Art

[0002] Conventionally, as a face authentication system that performs authentication based on face data, a face authentication system that performs personal authentication by collating input face data with pre-registered registered face data is known. For example, in the face authentication system described in Patent Document 1, a target person is authenticated using an image of a face portion in a captured image captured by a face authentication device such as a camera.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] When installing a face authentication device at the entrance door of an apartment or the like, the exposure of the face authentication device to natural light varies depending on the installation location. In addition, the light irradiation conditions may vary due to daily time zones, weather, and changes in the building structure. In particular, in a high-intensity environment, there is a problem that the authentication accuracy of the face image system deteriorates, and conventionally, there has been a problem that equipment for shielding natural light has to be installed around the face authentication device.

[0005] The present disclosure has been made in consideration of such points, and an object thereof is to provide a face authentication method and system capable of performing face authentication with high accuracy regardless of the installation conditions of the face authentication device and the surrounding environment.

Means for Solving the Problems

[0006] The face authentication method of the present disclosure is A step of installing a face recognition device having a camera at a predetermined location; A step of selecting an ND filter with a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters having different transmittances; A step of attaching the selected ND filter to the lens of the camera; A step of transmitting a face image captured by the camera from the face recognition device to a computer; A step of extracting a hash value from face image data by a predetermined hash function by the face recognition device or the computer; A step of performing face recognition on the face image captured by the camera by the face recognition device or the computer based on the extracted hash value and the hash value corresponding to a pre-registered face image; Comprising.

[0007] In the face recognition method, the illuminance limit of the face recognition device is preset, In the step of selecting an ND filter with a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters having different transmittances, an ND filter with a transmittance such that the luminous flux divergence does not exceed the illuminance limit of the face recognition device may be selected.

[0008] Also, in the face recognition method, in the step of selecting an ND filter with a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters having different transmittances, an ND filter with a transmittance of 5% or more and 50% or less may be selected.

[0009] Also, in the face recognition method, in the step of selecting an ND filter with a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters having different transmittances, an ND filter with a transmittance of 10% or more and 40% or less may be selected.

[0010] Also, in the face authentication method, in the step of selecting an ND filter with a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters with different transmittances, an ND filter with a transmittance of 15% or more and 35% or less may be selected.

[0011] In the system of the present disclosure, a face authentication device having a camera, installed at a predetermined location, a computer, A system comprising: an ND filter is attached to the lens of the camera, and the transmittance of the ND filter is a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value, a face image captured by the camera is transmitted from the face authentication device to the computer, a hash value is extracted from the face image data by a predetermined hash function by the face authentication device or the computer, Based on the extracted hash value and the hash value corresponding to a pre-registered face image, face authentication is performed on the face image captured by the camera by the face authentication device or the computer.

[0012] Also, in the system, the ND filter attached to the lens of the camera may be an ND filter with a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed the illuminance limit of the face authentication device.

[0013] Also, in the system, the transmittance of the ND filter attached to the lens of the camera may be 5% or more and 50% or less.

[0014] Also, in the system, the transmittance of the ND filter attached to the lens of the camera may be 10% or more and 40% or less.

[0015] Also, in the system, the transmittance of the ND filter attached to the lens of the camera may be 15% or more and 35% or less.

Advantages of the Invention

[0016] According to the face authentication method and system of the present disclosure, it is possible to provide a face authentication method and system capable of performing face authentication with high accuracy regardless of the installation conditions of the face authentication device and the surrounding environment.

Brief Description of the Drawings

[0017]

Figure 1

Figure 2

Figure 3

Figure 4

Modes for Carrying Out the Invention

[0018] Hereinafter, the present embodiment will be illustrated and described with reference to the drawings.

[0019] <System 1> As shown in FIG. 1, the system 1 according to the present embodiment includes a face authentication device 10 and a computer 20.

[0020] The face authentication device 10 is, for example, a device such as a smartphone, a portable tablet computer, a personal computer, or a mobile phone installed in a property or the like, and the face authentication device 10 includes a camera. A neutral density filter is attached to the lens unit of the camera. In the present embodiment, an ND filter is exemplified and described as the neutral density filter.

[0021] In this embodiment, the ND filter is a member that functions as a light quantity reduction member and is also a member that can reduce the risk of thermal damage to the lens barrel. The ND filter has a transmittance such that the illuminance divergence, which is the illuminance after transmission through the ND filter, does not exceed the illuminance limit as a predetermined value of the face authentication device 10. The face authentication device 10 further includes a control unit 11, a display unit 12, an operation unit 13, an imaging unit 14, a storage unit 15, and a communication unit 16, and these components are electrically connected.

[0022] The control unit 11 includes a CPU, a RAM, a ROM, etc., and is configured to execute various information processes based on programs and data.

[0023] The display unit 12 is, for example, a display and is used to display inputs, outputs, etc. A known liquid crystal display or the like can be used as the display. The operation unit 13 is a keyboard, an operation key, etc., and is for performing the above-described input. Also, a touch panel display that combines the display unit 12 and the operation unit 13 can be used. Hereinafter, as an example, the input by the operation unit 13 is assumed to be performed by touch input. The imaging unit 14 is a camera or the like that can image the face image of a visitor, an identity document, etc.

[0024] The storage unit 15 is composed of an auxiliary storage device such as an HDD or an SSD, for example, and stores various data and programs. The programs are for performing various processes described later. The control unit 11 is configured to execute various processes by interpreting and executing this program.

[0025] The communication unit 16 is, for example, a wired LAN (Local Area Network) module, a wireless LAN module, etc., and includes an interface for performing wired or wireless communication. That is, the communication unit 16 is composed of a communication interface configured to communicate with other devices via a communication network such as an Internet line.

[0026] Note that the specific hardware configuration of the face authentication device 10 can be appropriately omitted, replaced, and added according to the embodiment. For example, the control unit 11 may include a plurality of processors. The control unit 11 may be configured by an FPGA. The storage unit 15 may be configured by the RAM and ROM included in the control unit 11. Further, the face authentication device 10 may be a dedicated terminal for this system, or the above-described program can be installed and used in a general-purpose smartphone or the like.

[0027] In the face authentication device 10, the following processing is performed. That is, when the control unit 11 of the face authentication device 10 expands the program stored in the storage unit 15 into the RAM, the program is interpreted and executed by the CPU.

[0028] Further, the face authentication device 10 transmits the registration application information to the hash value registration means 212 via the communication unit 16. Further, the face authentication device 10 is installed at a predetermined position of the property, and in order to efficiently perform imaging and authentication of the visitor by the camera, the installation height of the face authentication device 10 is set to the eye level height of a general adult.

[0029] From the viewpoint of appropriately performing authentication on a sunny day, the transmittance of the ND filter is preferably 5% or more and 65% or less, more preferably 5% or more and 50% or less, even more preferably 10% or more and 50% or less, even more preferably 10% or more and 40% or less, even more preferably 10% or more and 35% or less, and most preferably 15% or more and 35% or less.

[0030] <Computer 20> The computer 20 is, for example, a management server of a system company or the like. The computer 20 is communicably connected to the face authentication device 10 via a network such as an Internet line. Further, the computer 20 has a control unit 21, a storage unit 22, a communication unit 23, a display unit 24, and an operation unit 25. The control unit 21 is, for example, a CPU or the like, and functions as a face image data receiving means 211, a hash value registering means 212, a hash value extracting means 213, an authentication means 214, and an admission restriction releasing means 215 by executing a program stored in the storage unit 22.

[0031] The face image data receiving means 211 receives face image data captured by a camera and sent to the computer 20.

[0032] The hash value extracting means 213 extracts a hash value from the feature amount data of the face image received by the face image data receiving means 211 using a predetermined hash function.

[0033] The hash value registering means 212 pre-registers in the storage unit 22 the hash value extracted from the feature amount data of the face image of the visitor scheduled to visit the property where the face image is captured on site.

[0034] Here, the hash value registering means 212 forms the feature amount data of the face image of the visitor scheduled to visit triggered by receiving registration application information from the face authentication device 10 or an external device, and registers the identification information of the visitor scheduled to visit and information related to the face image. The information related to the face image of the visitor scheduled to visit includes at least one of the hash value, checksum, and binarized compressed data of the feature amount data of the face image of the visitor scheduled to visit. Here, the feature amount data of the face image is data of a face image indicating the features of a person's face, and is, for example, a set of pixel values of pixels corresponding to the face image. Further, the identification information of the visitor scheduled to visit means an identifier such as an ID for identifying the visitor scheduled to visit.

[0035] The authentication means 214 performs face authentication of the visitor by comparing information on the face image of the visitor captured by the face authentication device 10 with information on the face image of the visitor-to-be-visited registered by the hash value registration means 212.

[0036] Also, the authentication means 214 detects the contour lines for each facial part included in the captured face image of the visitor and the registered face image of the visitor-to-be-visited based on the pixel values included in the information on the face image of the visitor captured by the face authentication device 10 and the information on the face image of the visitor-to-be-visited registered by the hash value registration means 212. The authentication means 214 collates the information on the face image of the visitor captured by the face authentication device 10 with each piece of information using the pairing ID as a landmark. Here, the facial parts included in the face image include, for example, eyebrows, eyes, nose, mouth, ears, moles, etc.

[0037] The authentication means 214 compares the detected contour lines of the captured face image of the visitor and the registered face image of the visitor-to-be-visited.

[0038] Also, based on the degree of approximation of the contour lines of the detected captured face image of the visitor and the registered face image of the visitor-to-be-visited, which are the comparison targets, the authentication means 214 determines the identity of the persons in both face images as a result of the comparison.

[0039] In addition, the authentication means 214 calculates the degree of approximation of each image. The degree of approximation is calculated based on, for example, the number of closed curves in the images, the number of closed curves adjacent to the adjacent closed curves, the number of closed curves included, and the like. For example, when the closed curves in the image are appropriately one-to-one associated within the image, if there are no closed curves that are not associated, the authentication means 214 uses the number of closed curves as a score. In addition, when the number of adjacent closed curves (adjacent lines) and the number of included closed curves (nested parent) between the associated closed curves match, the authentication means 214 further adds one point to the score for each. Such calculations are performed for all combinations of closed curves, and the maximum score is set as the final score between the images. The higher this score, the more qualitatively the authentication means 214 determines that the degree of approximation of the two images is large according to the height of the score.

[0040] In addition, when the calculated score is equal to or higher than a preset threshold value, the authentication means 214 determines that the person represented by the captured face image of the visitor and the registered face image of the visitor scheduled to visit is the same person. In addition, when the calculated score is lower than the preset threshold value, the authentication means 214 determines that the person represented by the captured face image of the visitor and the registered face image of the visitor scheduled to visit is not the same person, while leaving the possibility that they may be the same person.

[0041] In addition, when the authentication means 214 determines that the person represented by the captured face image of the visitor and the registered face image of the visitor scheduled to visit is the same person, it outputs a signal (permission signal) corresponding to permission as authentication. On the other hand, when the authentication means 214 determines that the person represented by the captured face image of the visitor and the registered face image of the visitor scheduled to visit is not the same person, it outputs a signal (denial signal) corresponding to denial as authentication.

[0042] When it is determined that the person represented by the captured face image of the visitor and the registered face image of the visitor scheduled to visit is the same person, the authentication means 214 outputs restriction release information for releasing the restriction on entry into a predetermined area based on the permission signal.

[0043] On the other hand, when it is determined that the person represented by the captured face image of the visitor and the face image of the registered visitor-to-be is not the same person, the authentication means 214 performs a process of not outputting the access restriction release information for releasing the restriction on entry into a predetermined area based on the negative signal.

[0044] When the access restriction release means 215 receives the access restriction release information output by the authentication means 214, in order to release the restriction on entry into a predetermined area by an access restriction device such as a door lock mechanism, the access restriction release means 215 sends information for causing the access restriction device to release the restriction on entry into the predetermined area.

[0045] Note that the program executed by the control unit 21 is not limited to the one stored in the storage unit 22. As the program executed by the control unit 21, a program transmitted from an external device to the control unit 21 via the communication unit 23, a program stored in a storage medium such as a USB memory attached to the computer 20, a program stored in a server (for example, a cloud server) different from the computer 20, etc. may be used.

[0046] The storage unit 22 is composed of, for example, an HDD (Hard Disk Drive), a RAM (Random Access Memory), a ROM (Read Only Memory), an SSD (Solid State Drive), etc. Various programs including the program executed by the control unit 21 are stored in the storage unit 22. Also, information regarding the identification information and face image of the registered visitor is stored in the storage unit 22.

[0047] The communication unit 23 is, for example, a wired LAN (Local Area Network) module, a wireless LAN module, etc., and includes an interface for performing wired or wireless communication. That is, the communication unit 23 is composed of a communication interface configured to communicate with other devices via a communication network such as an Internet line.

[0048] Next, the flow of each process of the system 1 will be described.

[0049] First, a face recognition device 10 having a camera is installed at a predetermined location of the property (step S10).

[0050] Next, from a plurality of ND filters with different transmittances, an ND filter with a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value is selected (step S20).

[0051] Next, the selected ND filter is attached to the lens of the camera (step S30). With the camera attached with the ND filter, a face image of the visitor is captured at the property.

[0052] When a face image of the visitor is captured at the property, the face image of the visitor captured by the camera is transmitted from the face recognition device 10 to the computer 20 (step S40).

[0053] The computer 20 extracts a hash value from the face image data using a predetermined hash function (step S50).

[0054] Next, based on the extracted hash value and the hash value corresponding to the pre-registered face image, the computer 20 performs face authentication on the face image captured by the camera (step S60).

[0055] The authentication means 214 compares the information regarding the face image of the visitor captured by the camera with the information regarding the face image of the visitor-to-be-registered registered by the hash value registration means 212, thereby performing authentication of the visitor (determining whether the person represented by the captured face image of the visitor and the registered face image of the visitor-to-be-registered is the same person) (step S70).

[0056] Here, when it is determined that the person represented by the captured face image of the visitor and the face image of the registered visitor scheduled to visit is the same person, the authentication means 214 outputs admission restriction release information for releasing the restriction on entry into a predetermined area by the entry restriction device based on the permission signal (step S80).

[0057] On the other hand, when it is determined that the person represented by the captured face image of the visitor and the face image of the registered visitor scheduled to visit is not the same person, the authentication means 214 outputs a denial signal and does not output admission restriction release information for releasing the restriction on entry into a predetermined area by the entry restriction device based on the denial signal (step S90).

[0058] Next, when the admission restriction release means 215 receives the admission restriction release information output by the authentication means 214, in order to release the restriction on entry into a predetermined area by the entry restriction device, it sends information to the entry restriction device to release the restriction on entry into the predetermined area. In this way, the admission restriction process for the visitor who visited the property is performed.

[0059] (Embodiment) As shown in FIG. 3, in a natural light environment, verification of authentication by an authentication device using a plurality of ND filters was performed. Note that it was confirmed that the authentication could be performed from all 17 directions, namely, 8 azimuths in the front, back, left, right, and diagonal directions from the horizontal (0°) and the upper diagonal 45°, and 1 direction from directly above, when irradiating the authentication device with light sources. As the base materials used for the verification, 2 silver reflectors, 2 constant light lights, 3 location lights, 1 authentication device (ProFaceX), and 1 illuminometer were used. After adjusting the light so that the vicinity of the camera of the authentication device had the appropriate illuminance with the lights and reflectors, measurements were taken. The authentication process was performed 5 times, and when all of these performances were successful, it was defined as "authentication possible (○)". Also, when success was not obtained in all performances, it was defined as "authentication impossible (×)".

[0060] As shown in Fig. 4, without a filter, for ND-0.2 and ND-0.3, authentication was not possible when the illuminance was 100,000 Lux, whereas authentication was possible with ND-0.5 or higher.

[0061] According to the face authentication method and system 1 according to the present embodiment having the above-described configuration, the face authentication method includes a step of installing a face authentication device 10 having a camera at a predetermined location, a step of selecting an ND filter having a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters having different transmittances, a step of attaching the selected ND filter to the lens of the camera, a step of transmitting a face image captured by the camera from the face authentication device 10 to a computer 20, a step of extracting a hash value from the face image data by a predetermined hash function by the face authentication device 10 or the computer 20, and a step of performing face authentication on the face image captured by the camera by the face authentication device 10 or the computer 20 based on the extracted hash value and the hash value corresponding to a pre-registered face image. Conventionally, as a face authentication system that performs authentication based on face data, a face authentication system that collates input face data with pre-registered registered face data to perform personal authentication is known. Here, when installing a face authentication device at the entrance door of an apartment or the like, the exposure of the face authentication device to natural light varies depending on the installation location. In addition, the light irradiation conditions may vary due to daily time zones, weather, and changes in the building structure. In particular, in a high illuminance environment, there is a problem that the authentication accuracy of the face image system deteriorates, and conventionally, there has been a problem that equipment for shielding natural light has to be installed around the face authentication device. According to the face authentication method and system 1 according to the present embodiment, an ND filter having a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value is selected from a plurality of ND filters having different transmittances, and face authentication can be performed with high accuracy regardless of the installation conditions of the face authentication device 10 and the surrounding environment.

[0062] Also, in the face authentication method and system 1 according to the present embodiment, as described above, the illuminance limit of the face authentication device 10 is preset, and in the step of selecting an ND filter with a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters with different transmittances, an ND filter with a transmittance such that the luminous flux divergence does not exceed the illuminance limit of the face authentication device 10 may be selected. In this case, by selecting an ND filter with a transmittance such that the luminous flux divergence does not exceed the illuminance limit of the face authentication device 10, face authentication can be performed with high accuracy regardless of the installation conditions of the face authentication device 10 and the surrounding environment.

[0063] Also, in the face authentication method and system 1 according to the present embodiment, as described above, in the step of selecting an ND filter with a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters with different transmittances, an ND filter with a transmittance of 5% or more and 50% or less may be selected. In this case, by selecting an ND filter with a transmittance of 5% or more and 50% or less, face authentication becomes possible.

[0064] Also, in the face authentication method and system 1 according to the present embodiment, as described above, in the step of selecting an ND filter with a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters with different transmittances, an ND filter with a transmittance of 10% or more and 40% or less may be selected. In this case, by selecting an ND filter with a transmittance of 10% or more and 40% or less, face authentication becomes possible.

[0065] Also, in the face authentication method and system 1 according to the present embodiment, as described above, in the step of selecting an ND filter with a transmittance such that the luminous flux divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters with different transmittances, an ND filter with a transmittance of 15% or more and 35% or less may be selected. In this case, by selecting an ND filter with a transmittance of 15% or more and 35% or less, face authentication becomes possible.

[0066] Note that the face recognition method and system 1 according to this embodiment are not limited to the above-described manner, and various modifications can be made.

[0067] For example, in this embodiment, the computer 20 extracts a hash value from the face image data using a predetermined hash function. However, the face authentication device 10 may extract a hash value from the face image data using a predetermined hash function. Also, the computer 20 performs face authentication on the face image captured by the camera. However, the face authentication device 10 may perform face authentication on the face image captured by the camera.

Explanation of Signs

[0068] 1: System 10: Face authentication device 11: Control unit 12: Display unit 13: Operation unit 14: Imaging unit 15: Storage unit 16: Communication unit 20: Computer 21: Control unit 22: Storage unit 23: Communication unit 24: Display unit 25: Operation unit 211: Face image data reception means 212: Hash value registration means 213: Hash value extraction means 214: Authentication means 215: Entrance restriction release means

Claims

1. A step of installing a face recognition device having a camera at a predetermined location; A step of selecting an ND filter having a transmittance such that the light beam divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters having different transmittances; A step of attaching the selected ND filter to the lens of the camera; A step of extracting a hash value from a face image captured by the camera to which the ND filter is attached by a predetermined hash function by the face recognition device; A step of performing face recognition on the face image captured by the camera by the face recognition device based on the extracted hash value and the hash value corresponding to a pre-registered face image; comprising A method of selecting an ND filter having a transmittance such that the light beam divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters having different transmittances, and selecting an ND filter having a transmittance of 10% or more and 40% or less.

2. The method according to claim 1, wherein in the step of selecting an ND filter having a transmittance such that the light beam divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value from a plurality of ND filters having different transmittances, an ND filter having a transmittance of 15% or more and 35% or less is selected.

3. A system comprising a face recognition device having a camera installed at a predetermined location, an ND filter is attached to the lens of the camera, and the transmittance of the ND filter is such that the light beam divergence, which is the illuminance after transmission of the ND filter, does not exceed a predetermined value, the face recognition device extracts a hash value from a face image captured by the camera by a predetermined hash function, face recognition is performed on the face image captured by the camera by the face recognition device based on the extracted hash value and the hash value corresponding to a pre-registered face image, a system in which the transmittance of the ND filter attached to the lens of the camera is 10% or more and 40% or less.

4. The system according to claim 3, wherein the transmittance of the ND filter attached to the lens of the camera is 15% or more and 35% or less.

Images