Method for setting the camera angle for capturing facial images of attendees and attendee authentication device

Abstract

We propose a method for setting the shooting angle to obtain optimal facial image information of multiple attendees in a classroom using a small number of cameras, along with an attendee authentication device. [Solution] When capturing facial images of attendees seated in chairs in a classroom having multiple desks neatly arranged from front to back, and chairs positioned behind each desk, using a facial image acquisition device (11) mounted on the upper part of the vertical front wall of the classroom, the shooting angle setting method according to the present invention calculates the appropriate facial image shooting angle y (degrees) of the facial image acquisition device (11) using a logarithmic function: y = Aln(X) + B, where A is the arrangement coefficient and B is the arrangement increment.

Description

Technical Field

[0001] The present invention relates to a method for setting a shooting angle of a camera that shoots a face image of an attendee sitting on a chair in a classroom and an attendee authentication device.

Background Art

[0002] The following Patent Document 1 shows an attendance management server by a face imaging device that uses a face imaging device by a camera arranged indoors where a plurality of attendees gather, shoots a plurality of face images in a seated state, and sends the data to an attendance management server to grasp the presence or absence of registered persons. In this face imaging device, a plurality of cameras for face imaging devices are installed indoors, and face images of attendees are automatically acquired in a seated state. Also, face images of attendees are registered in advance in the face imaging device, the acquired data is transmitted to the attendance management server, and attendance presence / absence data is managed.

[0003] Patent Document is discloses a technique for evaluating behavior at an individual or group level based on image data analysis using a social behavior recognition system 10 including a single or a plurality of PZT cameras 12 that shoot still images or video images and a computing device that implements a smartphone, a smartwatch, a tablet, a laptop computer, and a desktop computer.

[0004] Patent Document 3 discloses an imaging device that can continue tracking a subject even if tracking fails when tracking a subject based on video analysis. The imaging device, which is capable of driving pan, tilt and zoom and can communicate with a wireless device held by the subject, includes: a first identification means that performs video analysis on image data obtained by the imaging means of the imaging device to identify the relative position of the subject with respect to the imaging device; a second identification means that identifies the relative position of the subject with respect to the imaging device based on the direction of incoming radio waves received by the wireless device; a determination means that determines whether or not tracking of the subject has failed as a result of tracking the subject using the first identification means; a position identification control means that executes either the first identification means or the second identification means based on the determination result of the determination means; and a PTZ control means that performs pan, tilt or zoom drive control at the position identified by the first identification means or the second identification means. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2012-43389 [Patent Document 2] Japanese Patent Publication No. 2018-124972 [Patent Document 3] Japanese Patent Publication No. 2023-44924 [Overview of the project] [Problems that the invention aims to solve]

[0006] Conventional attendance management servers that authenticate the faces of all students in a school classroom are equipped with a facial imaging system that places one camera in front of each student. However, this facial imaging system has the drawback of being expensive to implement because it increases the number of cameras, network devices, and facial imaging system servers. In addition, a facial imaging system that places cameras above the front of each row of desks where students are seated has the problem that the faces of students become obscured as they move towards the back of the classroom, making it difficult to capture the faces of students. Therefore, there is a need for a new method of setting the shooting angle and an attendee authentication device that can obtain optimal facial image information of attendees with a small number of cameras, regardless of the classroom floor structure, including stepped floor structures and flat floor structures. [Means for solving the problem]

[0007] In a classroom where multiple desks are neatly arranged from front to back, and attendees are seated in chairs behind each desk, the present invention provides a method for setting the shooting angle for capturing facial images of attendees using a facial image acquisition device (camera) mounted on the upper part of the vertical front wall of the classroom. If A is the arrangement coefficient and B is the arrangement sum, the appropriate facial image shooting angle y (degrees) for the camera is calculated using a logarithmic function: y = Aln(X) + B, where X is the row order of the desks. The logarithmic function ln(X) is the natural logarithm function log with base e (Napier's number). e (X)

[0008] In a classroom where multiple desks are neatly arranged from front to back, and attendees are seated in one or more chairs behind each desk, the arrangement of the desks, the chairs behind each desk, and the facial positions of the seated attendees are predetermined or planned, or similar, so that attendees can effectively learn the lecture content.

[0009] When multiple desks and chairs behind each desk are arranged in a sequential line, and a camera is mounted on the front wall of a classroom where attendees seated in each chair can effectively learn the lecture content, the inventors have found that clear facial images of multiple attendees can be captured with one or a few cameras by using a given function that represents the relationship between the row order of the chairs and the angle (depression angle) of the camera. The distance of the camera from the attendees' faces is expressed as a function of the row order X of the desks.

[0010] For example, when capturing facial images of multiple attendees seated in chairs from the first row to the tenth row using a camera, the inventors experimentally determined the optimal camera angle for capturing the attendees' faces and plotted the camera angles for each chair in each row on a graph. The resulting graph uses the logarithmic function y = Aln(X) + B to approximate the appropriate facial image capture angle y (degrees) for the camera relative to the row order X of the desks. The appropriate facial image capture angle y (degrees) allows for reliable and clear recognition of the attendees from the digital facial image information captured by the camera. [Brief explanation of the drawing]

[0011] [Figure 1] Block diagram showing the basic configuration of the attendee authentication device according to the present invention [Figure 2] Cross-sectional view showing the angle of a camera mounted on the upper part of the front wall of a tiered-floor classroom relative to the horizontal ceiling surface and the faces of the attendees. [Figure 3] Cross-sectional view showing the angle of the camera relative to the horizontal ceiling surface and the faces of the attendees. [Figure 4] A cross-sectional view showing the camera angle relative to the attendees' faces, with respect to the side wall positioned perpendicular and perpendicular to the front wall of the classroom. [Figure 5] A logarithmic graph of Example 1 showing the optimal camera angle characteristics relative to the faces of attendees in each row in a flat-floor classroom structure. [Figure 6] Logarithmic graphs of Examples 2 and 3 showing the optimal camera angle characteristics relative to the faces of attendees in each row in a tiered-floor classroom. [Figure 7]Logarithmic graphs of Examples 4 and 5 showing the optimal camera angle characteristics for the faces of attendees in the left and right seats of each row in a tiered-floor classroom. [Figure 8] Logarithmic graphs of Examples 6 and 7 showing the optimal camera angle characteristics for the faces of attendees in the left and right seats of each row in a flat-floor classroom. [Modes for carrying out the invention]

[0012] Embodiments and examples of the shooting angle setting method and attendee authentication device according to the present invention will be described below with reference to the drawings.

[0013] The attendee authentication device of the present invention shown in Figure 1 comprises one or more facial image acquisition devices (11), i.e., cameras (11); one or more card readers (12) that read card information from authentication cards owned by each attendee; a network system (10) to which the cameras (11) and card readers (12) are connected for mutual communication; an attendance management server (13) that receives facial image information from the cameras (11) and attendee card information from the card readers (12) via the network system (10); and a facial image authentication server (14) that stores facial image information of a large number of attendees and is connected to the network system (10). The camera (11) is used in the attendee authentication device to confirm the attendance of attendees by performing the shooting angle setting method according to the present invention and taking a picture of the attendee's face at an appropriate shooting angle; and after taking a picture of the attendee's face, the camera (11) sends the facial image information to the network system (10).

[0014] The camera (11) is a camera with a PTZ function that controls the panning of the camera to cover a wide shooting area by remote operation. PTZ is an abbreviation of the first letters of Pan, Tilt, and Zoom. Pan refers to horizontal panning, Tilt refers to vertical panning, and Zoom refers to the functions of zooming in (telephoto) and zooming out (wide angle). The PTZ function of the camera (11) is controlled via the network system (10) by the attendance management server (13), the face image authentication server (14), and / or a personal computer or smartphone.

[0015] The attendance management server (13) of the attendee authentication device compares the face image information of the attendee transmitted from the camera (11) with the face image information stored in the face image authentication server (14), and when these pieces of information match and the card information from the card reader (12) matches the appropriate information of the attendee, it approves the presence of the attendee.

[0016] The inventors have learned the fact that if a camera installed in a classroom where a plurality of desks and the chairs arranged behind each desk are properly arranged is set at a given angle, clear face images of all attendees can be taken with a single or a small number of cameras. For example, in a flat-floor classroom or a stepped-floor classroom, the camera is installed on the front wall at a height of 2800 mm or 3000 mm from the floor shown in FIGS. 2 and 3, and the camera angle to the front-row attendees with respect to the ceiling plane is set at 27° or 30°. FIG. 2 shows a part of a stepped-floor classroom, but the present invention can be similarly and properly implemented in a flat-floor structure classroom. FIG. 4 shows an example in which the camera angle is set to 29° or less with respect to the face of the front-row attendee with respect to the side wall perpendicular to the front wall.

[0017] The inventors measured the camera angles for taking clear face image information of the attendees sitting on the chairs in each row in a classroom provided with desks in the first to tenth rows and the chairs arranged behind the desks. The actually measured values of the camera angles are shown in Table 1.

[0018]

Table 1

[0019] A first embodiment of the present invention includes Examples 1 to 8 in which the appropriate facial image capture angle y of the camera is displayed as a logarithmic function of the row order X of the desks. In the first embodiment of the appropriate facial image capture angle y of the camera, it is displayed as a logarithmic curve. In contrast, a second embodiment of the present invention includes Example 8 in which the appropriate facial image capture angle y is displayed as the area of ​​the region between logarithmic functions, which is sandwiched between a plurality of logarithmic curves displayed as logarithmic functions. The arrangement coefficient A and arrangement sum B for each embodiment are shown in Figures 5 to 8 as calculated numbers with two or three decimal places, but in the embodiments, the number is rounded to two decimal places to display the arrangement coefficient A and arrangement sum B as practical values.

[0020] Figure 2 shows a partial cross-sectional view showing the angle between the horizontal ceiling surface and the faces of attendees, with the camera used in implementing the shooting angle setting method of the present invention mounted on the vertical front wall of a tiered-floor classroom. Figure 2 is a cross-sectional view showing only the first and second rows, with the third row onwards omitted from the tenth row. In actual Figure 3, the sixth row onwards is omitted, and the first to fifth rows are shown.

[0021] Example 1 Example 1, shown in Figure 5, is an actual classroom with a stepped floor structure, where desks and chairs are arranged from front to back in rows 1 through 10. A camera is installed on the front wall at a height of 2800 mm or 3000 mm, and the relationship between the angle of the attendees' faces to the camera and each row is measured. If y (degrees) is the appropriate angle for capturing facial images with the camera for a given row order X of desks, then the approximation curve connecting the multiple measured values ​​plotted in Figure 5 is a logarithmic function: Example 1: y = -8.2 l n(X) + 26.8 It was found that this method allows for the calculation of an approximate value y of the angle at which the attendees' faces are captured by the camera.

[0022] Examples 2 and 3 Examples 2 and 3 in Figure 6, represented by numbers 2 and 3 in Table 1, show measured values ​​of the relationship between the angle of the attendees' faces to the camera and each row, measured by placing cameras on the front wall at camera angles of 30° and 27°, respectively, relative to the horizontal plane of the ceiling, with desks and chairs arranged in rows 1 through 10 with a height difference of 100 mm. If y (degrees) is the appropriate angle for capturing facial images with the camera for a given row order X of desks, then the logarithmic functions of the approximation curve connecting the multiple measured values ​​plotted in Figure 6 are: Example 2: y = -11.1 ln(X) + 30.2 Example 3: y = -10.4 ln(X) + 27.1 It was found that this can be expressed as a formula, and an approximate value y of the angle at which the attendee's face image is captured relative to the camera can be calculated.

[0023] Examples 4 and 5 Examples 4 and 5 in Figure 7, represented by numbers 4 and 5 in Table 1, show measured values ​​of the relationship between the angle of the attendees' faces to the camera and each row, measured by placing cameras on the front wall at camera angles of 30° and 27°, respectively, relative to the horizontal plane of the ceiling, with desks and chairs arranged in rows 1 through 10 with a height difference of 100 mm. If y (degrees) is the appropriate angle for capturing facial images with the camera for a given row order X of desks, then the logarithmic function of the multiple measured value connection approximation curves plotted in Figure 7 is: Example 4: y = -8.9 ln(X) + 29.5 Example 5: y = -3.7ln(X) + 12.2 It was found that this method allows for the calculation of an approximate value y of the angle at which the attendees' faces are captured by the camera.

[0024] Examples 6 and 7 Examples 6 and 7 in Figure 8, represented by numbers 6 and 7 in Table 1, show measured values ​​obtained by arranging desks and chairs in rows 1 through 10 in a flat-floor classroom, and mounting cameras on the front wall at camera angles of 30° and 27°, respectively, relative to the horizontal plane of the ceiling, to measure the relationship between the angle of the attendees' faces to the cameras and each row. If y (degrees) is the appropriate angle for capturing facial images with the camera for a given row order X of desks, then the logarithmic functions of the approximation curve connecting the multiple measured values ​​plotted in Figure 8 are: Example 6: y = -10.8 ln(X) + 36.9 Example 7: y = -6.6 l n(X) + 21.1 It was found that this method allows for the calculation of an approximate value y of the angle at which the attendees' faces are captured by the camera.

[0025] Example 8 In Examples 4 and 5 of Figure 7, the appropriate facial image capture angles y (degrees) for the left and right seats of the camera are expressed by logarithmic functions: y = -8.9ln(X) + 29.5 and y = -3.7ln(X) + 12.2, respectively. However, in Example 7, it was found that the camera angle corresponding to the logarithmic region between the two logarithmic curves of Examples 4 and 5 can be adopted as the appropriate facial image capture angle.

[0026] Example 9 In Examples 6 and 7 of Figure 8, the appropriate facial image capture angles y (degrees) for the left and right seats of the camera are expressed on logarithmic curves represented by the logarithmic functions y = -10.8ln(X) + 36.9 and y = -6.6ln(X) + 21.1, respectively. However, in Example 8, similar to Examples 4 and 5, it was found that the camera angle corresponding to the logarithmic region between the two logarithmic curves of Examples 6 and 7 can be adopted as the appropriate facial image capture angle.

[0027] According to the above-described examples 1 to 7, if A is the arrangement coefficient and B is the arrangement sum, the general formula for the appropriate camera facial image capture angle y (degrees) is expressed as the logarithmic function of the desk row order X as y = Aln(X) + B. The arrangement function A and the arrangement sum B are constants that differ depending on the size of the classroom, the arrangement of desks and chairs, and the angle of the attendees' faces relative to the camera, but the angle of the camera relative to the faces of seated attendees in a classroom where a lecture is being held does not differ significantly.

[0028] The embodiments and models of the present invention are not limited to those described above and can be further modified. For example, even in a classroom where desks and chairs are arranged in fewer than 10 rows or 11 or more rows, the appropriate facial image capture angle y (degrees) of the camera can be displayed as a logarithmic function, allowing attendees to be recognized from the captured facial images. The appropriate facial image capture angle y (degrees) of the camera can also be displayed as a logarithmic function in a classroom where 1 to 4 attendees are seated in chairs behind a single desk.

[0029] The appropriate facial image capture angle y for the logarithmic function is the angle at which the camera lens focuses on the attendee's face. However, even if the angle does not fall on the logarithmic function curve and is slightly off-focus from the camera lens, an angle with an angular error value within 10% of the logarithmic function curve can still be selected as a clear, appropriate facial image capture angle image that can identify the attendee's face. Figures 2 and 3 illustrate a classroom with a stepped floor structure, but it should be understood that the shooting angle setting method and attendee authentication device of the present invention can be similarly applied to classrooms with a flat floor structure.

[0030] The present invention's method for setting the shooting angle provides the following effects and benefits. 1. By determining the seating position and orientation relative to the camera, selecting the placement coefficient A and placement sum B, and setting the appropriate facial image capture angle y for one or a few cameras using a logarithmic function, clear facial images of all attendees can be captured. This allows for confirmation of the presence or absence of attendees. 2. The camera (11) that controls the camera's orientation, or the attendance management server (13), stores a logarithmic function to determine the placement function A and placement summation number B according to the classroom conditions, allowing the camera to capture clear facial images of all attendees. 3. In a typical classroom, an optical magnification of 20x to 30x for a camera with PTZ functionality is usually sufficient. 4. The system can capture clear facial images of attendees in classrooms with flat floors or floors with steps of 100mm, 150mm, or any other desired step height. [Industrial applicability]

[0031] The shooting angle setting method and attendee authentication device according to the present invention can be used to verify the attendance status of meetings and conferences within various organizations, including universities, associations, and companies, including medical and pharmaceutical institutions. [Explanation of Symbols]

[0032] (10) Network system, (11) PTZ camera (facial image acquisition device), (12) Card reader, (13) Attendance management server, (14) Facial image authentication server

Claims

1. When a facial image acquisition device mounted on the upper part of the vertical front wall of a classroom captures facial images of attendees seated in chairs in a classroom having multiple desks arranged neatly from front to back, and chairs placed behind each desk, Let A be the placement coefficient and B be the placement sum. For the row order X of the desks, the logarithmic function is: y = Aln(X) + B A method for setting the shooting angle, characterized by calculating the appropriate facial image shooting angle y (degrees) for a facial image acquisition device.

2. The logarithmic function representing the appropriate facial image acquisition angle y (degrees) for a facial image acquisition device is: y = -8.2 ln(X) + 26.8 y=-11.1ln(X)+30.2 y=-10.4ln(X)+27.1 y = -8.9 ln(X) + 29.5 y = -3.7 ln(X) + 12.2 y = -10.8 ln(X) + 36.9 or y = -6.6 ln(X) + 21.1 The method for setting the shooting angle according to claim 1, which is displayed in any of the following ways.

3. When attendees are seated in the left seat, located behind each of the multiple desks, and in the right seat, located adjacent to the left seat, the appropriate facial image acquisition angle y (degrees) for the left and right seats of the facial image acquisition device is expressed as a logarithmic function: y = -10.8 ln(X) + 36.9 and y = -6.6 ln(X) + 21.1 The method for setting the shooting angle according to claim 2, wherein the camera angle in the region corresponding to the area between the double logarithmic function curves is adopted as the appropriate facial image shooting angle.

4. When attendees are seated in the left seat, located behind each of the multiple desks, and in the right seat, located adjacent to the left seat, the appropriate facial image acquisition angle y (degrees) for the left and right seats of the facial image acquisition device is expressed as a logarithmic function: y = -8.9 ln(X) + 29.5 y = -3.7 ln(X) + 12.2 The method for setting the shooting angle according to claim 2, wherein the camera angle in the region corresponding to the area between the double logarithmic function curves is adopted as the appropriate facial image shooting angle.

5. The method for setting the shooting angle according to claim 1, wherein even for facial image shooting angles that do not lie on a logarithmic function curve, an error value within 10% of the logarithmic function angle is adopted as the appropriate facial image shooting angle.

6. The method for setting the shooting angle according to claim 1, wherein the angle of the camera mounted on the vertical front wall of the classroom to the desk of the front seat is 29 degrees or less with respect to a vertical side wall positioned perpendicular to the front wall.

7. The method for setting the shooting angle according to claim 1, wherein the classroom has a plurality of desks arranged in an orderly manner from front to back, and chairs placed behind each desk, and the classroom has a flat floor structure or a stepped floor structure.

8. The method for setting the shooting angle according to claim 7, wherein the stepped floor structure has a step height of 100 mm, 150 mm, or other heights.

9. A facial image acquisition device that performs the shooting angle setting method described in any one of claims 1 to 8 to capture facial image information of attendees at an appropriate facial image shooting angle y, A card reader into which each attendee enters the card information of the card they own, An attendance management server that receives facial image information from a facial image acquisition device and attendee card information from a card reader, A facial image authentication server that stores facial image information of a large number of attendees, The system comprises a network system that interconnects a facial image acquisition device, a card reader, an attendance management server, and a facial image authentication server. An attendee authentication device characterized in that the attendance management server compares the attendee's facial image information transmitted from the facial image acquisition device with the facial image information stored in the facial image authentication server, and approves the attendee's attendance when these pieces of information match and the card information from the card reader matches the attendee's information.

Images