JP2020009291A - Face image selection device - Google Patents

Abstract

To provide a face image selection device capable of selecting a face image with little effect of glasses from multiple images for face authentication which are face images oriented in an appropriate direction.SOLUTION: The face image selection device includes: a face orientation estimation part 12 that estimates the angle of the face direction in the user's face image captured by a camera 100 and generates a piece of face orientation information; an overexposure information generation part 13 that generates a piece of overexposure information representing the percentage of overexposed areas due to the reflection of glasses in the face image; and a face image selection part 15 that selects a face image that satisfies the predetermined condition for the face orientation information and the overexposure information.SELECTED DRAWING: Figure 1

Description

  The present disclosure relates to a face image selection device that selects a human face image.

  As an individual authentication method, a face authentication technology for performing authentication based on a face image has been widely used. In the face authentication technology, it is necessary to register a face image to be used for authentication in advance.

  As a face authentication technique, it is general to extract a feature point such as an outer corner of the eye or a corner of a mouth in a face image and perform authentication based on a distance between the feature points, an image feature amount of a peripheral region including the feature point, and the like. Therefore, as a face image used for face authentication, an image in which feature points of the face are clear is desirable. An image with clear feature points is likely to be obtained when the user's face is appropriately oriented with respect to the camera that captures the face image. As a technique for photographing a face image facing an appropriate direction, for example, there is a technique disclosed in Patent Document 1.

  Patent Document 1 discloses the following technology. Analyzing the face image of a person, detecting the eyes and pupils as feature points, comparing the difference between the eye position parameter and the pupil position parameter calculated based on the detected eyes and pupils with a target value, and detecting the face direction Is determined. The direction of the face of the person is determined at the time of photographing, and when it is determined that photographing is impossible, the correction content is transmitted to the person, so that a face photograph with the proper direction can be obtained.

JP 2006-268248 A

  By the way, when capturing a face image used for face authentication, for example, when the user wears glasses, depending on the position of the light source, the position corresponding to the glasses shines white due to reflection (the pixel at the position corresponding to the glasses). In some cases, a face image with a value of saturated whiteout is captured. With such a face image, accurate face authentication is difficult.

  An object of the present disclosure is to provide a face image selection device capable of selecting a face image having a small influence of reflection of glasses from a plurality of face authentication images captured.

  A face image selection device according to an embodiment of the present disclosure includes a face direction information acquisition unit that acquires face direction information indicating an angle of a face direction in a user's face image captured by an imaging device, and overexposure due to reflection of glasses in the face image. And a face image selection unit that selects a face image in which the face orientation information and the whiteout information satisfy predetermined conditions.

  ADVANTAGE OF THE INVENTION According to this invention, the face image with little influence of the reflection of glasses can be selected from the several images for face authentication which were image | photographed.

1 is a diagram illustrating a configuration example of a face image selection device according to an embodiment of the present disclosure. Diagram for explaining face angle Diagram for explaining whiteout information FIG. 7 is a diagram for explaining a predetermined angle range instructed by a face direction instructing unit; Flow chart for describing a first operation example of a face image selection process of the face image selection device Flow chart for describing a first operation example of a face image selection process of the face image selection device Flowchart for describing a second operation example of the face image selection process of the face image selection device The figure which illustrated the hardware constitutions of the computer which implement | achieves each function of a face image selection device by a program.

  Hereinafter, each embodiment of the present disclosure will be described in detail with reference to the drawings. However, an unnecessary detailed description, for example, a detailed description of a well-known matter or a redundant description of substantially the same configuration may be omitted. The following description and the drawings referred to are provided for those skilled in the art to understand the present disclosure, and are not intended to limit the scope of the present disclosure.

[Configuration example]
FIG. 1 is a diagram illustrating a configuration example of a face image selection device 10 according to an embodiment of the present disclosure. As shown in FIG. 1, the face image selecting device 10 includes a face image obtaining unit 11, a face direction estimating unit 12, a whiteout information generating unit 13, an evaluation function value calculating unit 14, a face image selecting unit 15, a face direction instructing unit. 16 and a processing control unit 17.

  The face image acquisition unit 11 acquires a face image generated by the camera 100 provided outside the face image selection device 10 by photographing a user's face. The camera 100 captures an image based on visible light or infrared light incident on an image sensor (not shown). The camera 100 is an example of an imaging device according to the present disclosure. The camera 100 may capture an image based on reflected light such as infrared light emitted from a light source provided near the camera 100. Further, instead of providing the camera 100 outside the face image selection device 10, the camera may be built in the face image selection device 10.

  The face direction estimating unit 12 extracts the user's face in the face image acquired by the face image acquiring unit 11, and estimates the direction in which the user's face is facing (hereinafter, referred to as the face direction). As a method in which the face direction estimating unit 12 estimates the user's face direction based on the face image, for example, there is a method based on machine learning. The face direction estimating method by machine learning is to estimate a user's face direction in an image based on a model obtained by learning a lot of face directions of a face image and a face in the face image. Is the way.

  In addition, the face direction estimating unit 12 acquires face direction information, which is information on the direction of the face in the face image, based on the estimation result. The face direction information is information indicating, for example, an angle (hereinafter, referred to as a face direction angle) of a face direction in an image based on a face direction in which the user is directly facing the camera 100. is there. The face direction estimating unit 12 is an example of a face direction information acquiring unit according to the present disclosure.

  FIG. 2 is a diagram for explaining a face direction angle. FIG. 2 is a diagram in which the positional relationship between the user's head and the camera 100 is viewed from above the user's head. In FIG. 2, the direction of the arrow indicating “face direction” is the direction in which the user's face is facing. The angle between this direction and the direction in which the camera 100 is viewed from the user is the face direction angle.

  In FIG. 2, only the face direction angle in the left-right rotation direction (yaw direction) around the vertical direction of the user's head is considered, but the present disclosure is not limited to this. As the face direction angle of the present disclosure, for example, the face direction angle in the vertical rotation direction (pitch direction) around the left-right direction of the user's head, or the left-right rotation direction (roll direction) around the front-rear direction of the user's head May be the face direction angle.

  The whiteout information generation unit 13 generates whiteout information indicating the degree of whiteout in the face image acquired by the face image acquisition unit 11. Note that overexposure means that light reflected by eyeglasses or the like is incident on the image sensor of the camera 100, so that a white pixel (having a luminance value near the maximum) occurs in the captured image, and the gradation of that portion is lost. It is a phenomenon that is done. The degree of overexposure is, for example, the ratio of pixels occupied by a pixel whose luminance value is equal to or higher than a predetermined value near the maximum value among all pixels constituting the face image. For example, in the case of 256 gradations, it is only necessary to determine that a pixel having a luminance of 255 (maximum) is overexposed. Alternatively, a range may be provided for the luminance determined to be overexposed, and, for example, a pixel having a luminance of 245 to 255 may be determined to be overexposed.

  FIG. 3 is a diagram for explaining the whiteout information. When the overexposed area exists in the face image as shown in FIG. 3, the overexposed information generation unit 13 divides the number of pixels occupied by the overexposed area by the number of pixels of the entire face image and occupies the overexposed area. By calculating a ratio (hereinafter, referred to as a whiteout ratio), whiteout information including the calculated whiteout ratio is generated.

  In a face image in which overexposure occurs due to light reflected on glasses, the gradation of pixels corresponding to eyes and pupils may be lost. For this reason, it is desirable that the face direction estimating method used in the face direction estimating unit 12 described above does not use the feature points corresponding to the eyes and the pupils.

  The evaluation function value calculation unit 14 defines an evaluation function indicating how appropriate the face image acquired by the face image acquisition unit 11 is as a face image used for face authentication. The evaluation function is a function defined by using, as parameters, face direction information regarding the face direction angle estimated by the face direction estimating unit 12 and overexposure information regarding overexposure generated by the overexposed information generating unit 13. The evaluation function value calculation unit 14 is an example of an evaluation value calculation unit according to the present disclosure.

Specifically, the evaluation function _fv is given, for example, by the following equation (1).
f _v = A × | d _f | / 360 ° + B × p _w (1)

However, in equation (1), _df is the face direction angle when the camera direction is set to 0, and _pw is the ratio of the number of overexposed pixels to the number of pixels of the entire face image. The coefficients A and B are weighting coefficients. The value of such evaluation function _{f v,} for example the face of the user is directly facing the full camera 100 when _{(| = 0 | d f)} , and overexposure pixel number is 0 _(p w = 0) Becomes the minimum (0). That is, the closer to the value of the evaluation function f _v is 0, the face image is an appropriate image more face recognition. The value of the evaluation function f _v is an example of the evaluation value of the present disclosure.

  The face image selecting unit 15 selects a face image in which the face direction information and the whiteout information satisfy predetermined conditions from a plurality of face images, and outputs the selected face image to the outside. The external is, for example, a face authentication device that performs face authentication using a face image captured by the face image selection device 10.

The face image in which the face direction information and the overexposed information satisfy the predetermined condition is, specifically, a face image in which the value of the evaluation function described above is as small as possible. The face image whose evaluation function value is as small as possible is a face image in which the face direction is not so far from the direction of the camera 100 and the ratio of the number of overexposed pixels is small to some extent. As explained above, the value of the evaluation function f _v, for example the face of the user is directly facing the full camera 100 (d f _{= 0),} and overexposure pixel count is ₀ (p w ₌ 0) In this case, the value becomes the minimum (0), and the face image indicating this value is the image most suitable for face authentication.

However, for example, when the light source is provided near the camera 100, the more the user's face faces the camera 100, the more likely overexposure occurs in the face image. For this reason, overexposed proportion p _w increases when you try to reduce the face angle d _f, overexposed when you try to reduce the proportion p _w face angle d _f increases. In such a case where is the face angle d _f and overexposed proportion p _w there is a trade-off, the value of the evaluation function f _v is difficult to 0. In this case, the face image selecting section 15 selects the face image as close to 0 can the value of the evaluation function f _v.

  The face direction instructing unit 16 instructs the user on the face direction based on the face direction information generated by the face direction estimating unit 12. The face direction instructing unit 16 instructs the user by, for example, voice or the like. This is because, for example, when an instruction is given using a display device such as a small display, the user's eyes turn to the display device, and it is difficult to obtain an image suitable for face authentication. Therefore, in the present disclosure, any instruction method that does not affect the user's line of sight or face direction may be used, and it is not necessary to use voice. Specifically, for example, a method may be adopted in which light sources are provided on the left and right sides of the user, and the light sources in the direction to be directed emit light.

  The face direction instructing unit 16 instructs the user to face the camera 100 and the face. More specifically, the face direction instructing unit 16 instructs the user such that the face direction angle falls within a predetermined angle range.

  FIG. 4 is a diagram for describing a predetermined angle range specified by the face direction specifying unit 16. As shown in FIG. 4, the predetermined angle range is an angle range centered on the direction in which the user's face faces the camera 100. That is, if the face orientation angle falls within the predetermined angle range, the user's face is almost directly facing the camera, and the camera 100 can take an image suitable for face authentication. Note that a specific example of the predetermined angle range may be, for example, ± 5 ° with the camera 100 direction being 0 °.

  The processing control unit 17 controls the entire processing of the facial image selection device 10 in each of the above-described configurations.

[First operation example]
Next, a face image selecting process of the face image selecting apparatus 10 having the above-described configuration will be described. FIGS. 5 and 6 are flowcharts for describing a first operation example of the face image selection process of the face image selection device 10.

  In FIG. 5, the processing from step S1 to step S6 is repeated until YES is selected in step S7.

  First, the face direction instructing unit 16 instructs the user such that the face direction is within a predetermined angle range (see FIG. 4) (step S1). Then, the face image acquisition unit 11 acquires a face image of the user captured by the camera 100 (Step S2). In step S2, the face image acquisition unit 11 may acquire one face image for each instruction in step S1, or when the camera 100 continuously captures a plurality of face images, Alternatively, a plurality of face images may be acquired at one time.

  The face direction estimating unit 12 estimates a face direction angle in the acquired face image and generates face direction information (step S3). The face direction estimating unit 12 determines whether or not the estimated face direction angle is within a predetermined angle range (Step S4). If the face orientation angle is within the predetermined angle range (step S4: YES), the process proceeds to step S5, otherwise (step S4: NO), the process returns to step S1.

  By the processing up to step S4, a face image whose face direction angle is surely within the predetermined angle range can be obtained. Next, the overexposed information generation unit 13 calculates the ratio of the number of overexposed pixels to the total number of pixels of the face image, and generates overexposed information (step S5). Further, the evaluation function value calculation unit 14 calculates an evaluation function value defined based on the face direction information and the overexposed information (Step S6).

  Then, the processing control unit 17 determines whether or not the processing from step S1 to step S6 has already been repeated a specified number of times (step S7). The prescribed number is not particularly limited, but is, for example, about three times. If it has been repeated the specified number of times (step S7: YES), the process proceeds to step S11 in FIG. 6 (see R1 in FIGS. 5 and 6). On the other hand, if it has not been repeated the specified number of times (step S7: NO), the process returns to step S1.

  As described above, while the processing from step S1 to step S6 is repeated a specified number of times, the face direction is within the predetermined angle range, whiteout information is generated, and a plurality of face images for which the evaluation function value is calculated are obtained. . Each face image is stored in, for example, a storage unit not shown in FIG. This storage unit may be provided outside the face image selection device 10 or may be built in the face image selection device 10. Through such processing, a plurality of face images each having a slightly different face direction and degree of overexposure are obtained.

  The processing control unit 17 determines whether or not there is a face image whose evaluation function value is equal to or less than a predetermined threshold value among the plurality of face images (step S11). The predetermined threshold is a threshold for determining whether or not the face image is appropriate for face authentication, and is determined by the authentication accuracy of the face authentication device that performs face authentication using the face image selected by the face image selection device 10. Value. When there is no image whose evaluation function value is equal to or smaller than the predetermined threshold (step S11: NO), the process proceeds to step S12, and when there is an image whose evaluation function value is equal to or smaller than the predetermined threshold (step S11: YES), The process proceeds to step S14.

  The processing control unit 17 increases the specified number of times that the processing of steps S1 to S6 is repeated for each component of the face image selection device 10 and / or causes the face direction instruction unit 16 to: The predetermined angle range is increased. Then, the processing control unit 17 determines whether each value has become equal to or more than a predetermined limit value as a result of increasing the specified number of times and / or the predetermined angle range (Step S13). If the specified number of times and / or the predetermined angle range is equal to or larger than the predetermined limit value (step S13: YES), the process proceeds to step S14, otherwise (step S13: NO), returns to step S1 in FIG. (See R2 in FIGS. 5 and 6).

  When the specified number of times is increased, the number of times that the processing from step S1 to step S6 in FIG. 5 is repeated increases. In this case, since the number of trials simply increases, the possibility that a more appropriate face image (small evaluation function value) is obtained increases. When the predetermined angle range is increased, the range in which the face direction instruction unit 16 turns the face toward the user by the face direction instruction in step S1 of FIG. 5 is widened. In this case, the face direction of the face image acquired in step S2 is slightly away from the direction of the camera 100, but since the face direction is different from the direction up to that time, the overexposure ratio may be improved. There is. For this reason, the possibility that a more appropriate face image is acquired increases. Then, by simultaneously increasing the prescribed number of times and the predetermined angle range, it is possible to further increase the possibility of obtaining an appropriate face image.

  In step S12, the processing control unit 17 does not necessarily need to increase both the specified number of times and the predetermined angle range, and may increase only one of them.

  Further, the limit value used for the determination in step S13 is a predetermined number of times and / or a maximum value in a predetermined angle range. When the initial value of the prescribed number of times is three, for example, the increment of one degree may be set to two and the limit value may be set to nine. That is, the prescribed number of times when increased once is five times, seven times when increased twice, and nine times when further increased three times. When the initial value of the predetermined angle range is ± 5 °, for example, the increment of one degree may be set to ± 5 ° and the limit value may be set to ± 15 °. That is, the predetermined angle range when the angle is increased once is ± 10 °, and when the angle is further increased twice, it is ± 15 °. By preliminarily setting the limit value in this way, it is possible to prevent the processing from taking too much time or selecting a face image that does not face the camera 100 at all. In addition, the initial value and the limit value of the specified number of times and the predetermined angle range exemplified above are merely examples, and the present disclosure is not limited thereto.

  Then, the face image selecting unit 15 selects a face image having the smallest evaluation function value from among the plurality of face images obtained by the processing so far (step S14). As described above, the face image having the smallest evaluation function value is an image suitable for face authentication, in which the face faces the camera 100 and has few overexposed areas. In step S14, the face image selecting unit 15 may select a predetermined number of face images instead of selecting only one face image having the smallest evaluation function value, or may select a predetermined number of face images. All face images that are equal to or smaller than the threshold value may be selected.

  As described above, according to the first operation example of the face image selection device 10, a face image in which the face direction angle of the user falls within the predetermined angle range is acquired without fail. Then, the evaluation function value is calculated based on the overexposure information including the ratio of the number of overexposed pixels included in the face image and the face direction information including the face direction angle, and the face image with the smallest evaluation function value is selected. As a result, a face image suitable for face authentication is selected. Furthermore, if the evaluation function value of the face image required by the face image selection device 10 does not satisfy the predetermined threshold, the process is repeated by changing conditions such as the number of repetitions and the range of the face orientation angle. Increase the likelihood of getting

[Second operation example]
Next, a second operation example of the face image selection device 10 will be described. FIG. 7 is a flowchart for describing a second operation example of the face image selection process of the face image selection device 10.

  First, the face direction instructing unit 16 instructs the user so that the face direction is within a predetermined angle range (see FIG. 4) (step S21). Then, the face image acquisition unit 11 continuously acquires the user's face image captured by the camera 100 (Step S22). In the second operation example, it is assumed that the camera 100 continuously captures images at a predetermined minute time interval. In step S22, the face image acquisition unit 11 changes the face direction according to the instruction in step S21. It is possible to continuously obtain a face image of a user who is in the presence of the user.

  The face direction estimating unit 12 estimates a face direction angle for all continuously acquired face images and generates face direction information (step S23). The face direction estimating unit 12 determines whether or not the estimated face direction angle is within a predetermined angle range in each image (step S24). If the face image has a face orientation angle within the predetermined angle range (step S24: YES), the process proceeds to step S25; otherwise (step S24: NO), the process returns to step S21.

  By the processing up to step S24, a plurality of face images in which the face orientation angle is within the predetermined angle range are obtained. Next, in each face image, the whiteout information generation unit 13 calculates the ratio of the number of whiteout pixels to the total number of pixels in the face image, and generates whiteout information (step S25). The whiteout information generation unit 13 determines whether the whiteout ratio is within a predetermined ratio range in each face image (step S26). For a face image whose whiteout ratio is not within the predetermined ratio range (step S26: NO), the process proceeds to step S27; otherwise (step S26: YES), the process proceeds to step S210.

  If the whiteout ratio is not within the predetermined ratio range, the face image currently being processed does not satisfy the suitability of the face image required by the face image selection device 10. Therefore, in the following steps S27 to S29, the whiteout ratio of the face image acquired later is improved. The whiteout information generation unit 13 determines whether the whiteout ratio of the face image currently being processed is smaller than the whiteout ratio of the previously processed face image among a plurality of continuous face images ( Step S27). If the whiteout ratio of the face image currently being processed is smaller than that of the previously processed face image (step S27: YES), the process proceeds to step S28; otherwise (step S27: NO), the process proceeds. Goes to step S29.

  If the overexposed ratio in the face image currently being processed is smaller than the overexposed ratio of the previously processed face image, the overexposed ratio is gradually improved by the current face direction instruction (step S21). Become. For this reason, the face direction instructing unit 16 continues the face direction instruction at the time when the currently processed face image is captured (step S28). Then, the process returns to step S22.

  On the other hand, if the overexposed ratio in the face image that is currently being processed is not less than the overexposed ratio of the previously processed face image, the overexposed ratio is deteriorating by the current face direction instruction (step S21). Become. For this reason, the face direction instructing unit 16 returns to the direction in which the face direction was instructed at the time when the face image processed immediately before was photographed, and gives an instruction again (step S29). Then, the process returns to step S22. As a result, it is possible to prevent the whiteout ratio from further increasing.

  As described above, by performing the processing from step S27 to step S29, the overexposed ratio of the face image acquired later can be improved.

  In step S26, when it is determined that the overexposed ratio in the face image currently being processed is within the predetermined ratio range, the image is considered to satisfy the appropriateness of the face image required by the face image selecting device 10. For this reason, the face image selecting unit 15 selects the latest face image at that time, that is, the face image in which the whiteout ratio is within the predetermined ratio range (step S210). Thereby, the face image selecting unit 15 can select a face image suitable for face authentication.

  As described above, according to the second operation example of the face image selection device 10, the face images of the user whose face direction is being changed by the face direction instruction are continuously obtained. Then, it is determined whether or not the face direction angle is within a predetermined angle range and whether or not the overexposed ratio is within a predetermined ratio range, and a face image that is within the range is selected as a face authentication image. Is done. On the other hand, when a face image whose whiteout ratio is not within the predetermined ratio range is obtained, the whiteout ratio of the face image obtained thereafter is improved. Specifically, by changing the pointing direction in the face direction instruction based on the change in the whiteout ratio between the face image processed immediately before and the face image being processed, the whiteness of the face image obtained thereafter is changed. The jump ratio is improved.

  As described above, in the second operation example, the pointing direction in the face direction instruction is changed based on the overexposed ratio in the obtained face image so that the overexposed ratio is improved. As a result, an image suitable for face authentication can be quickly selected.

<Modification>
Although various embodiments have been described with reference to the drawings, the present disclosure is not limited to such examples. It will be apparent to those skilled in the art that various changes or modifications can be made within the scope of the claims, and these naturally belong to the technical scope of the present disclosure. I understand. Further, each component in the above embodiment may be arbitrarily combined without departing from the spirit of the disclosure.

  In the embodiment described above, the camera 100 is provided outside the face image selection device 10, but the present disclosure is not limited to this. The camera 100 may be included in the face image selection device 10.

  In the above-described embodiment, the face direction estimating unit 12 estimates the user's face direction and generates face direction information based on the result, but the present disclosure is not limited to this. The face direction estimating unit 12 may acquire the already generated face direction information.

<Action / Effect>
According to the face image selection device 10 of the present disclosure, the face direction estimating unit 12 that acquires face direction information indicating the angle of the face direction in the user's face image captured by the camera 100, The image processing apparatus includes a whiteout information generation unit 13 that generates whiteout information indicating a ratio of a whiteout region, and a face image selection unit 15 that selects a face image in which face direction information and whiteout information satisfy predetermined conditions.

  Further, according to face image selecting device 10 of the present disclosure, face image selecting unit 15 selects a face image from a plurality of face images based on the value of an evaluation function defined using face direction information and overexposed information. Then, a face image whose face direction information and overexposed information satisfy predetermined conditions is selected.

  With such a configuration, it is possible to select a face image suitable for face authentication, in which the face direction faces the camera 100 as much as possible and the whiteout is as small as possible from a plurality of face images.

  Furthermore, when the evaluation function value is not less than or equal to the predetermined threshold, the face image selection device 10 of the present disclosure changes the conditions such as the number of repetitions and the range of the face orientation angle and performs the process again. This can increase the possibility of obtaining an appropriate face image.

  Further, the face direction instructing unit 16 instructs that the angle of the face direction is within a predetermined angle range, and when the direction of the user's face changes due to the instruction, the ratio of the overexposed area is changed after the change before the change. Is larger, the user is instructed to turn the face in the direction before the change.

  With such a configuration, the pointing direction in the face direction instruction is changed based on the overexposed ratio in the obtained face image so that the overexposed ratio is improved. As a result, an image suitable for face authentication can be quickly selected.

<Example of computer>
FIG. 8 is a diagram exemplifying a hardware configuration of a computer that realizes each function of the face image selecting device 10 in the above-described embodiment by a program.

  As shown in FIG. 8, a computer 2100 includes an input device 2101, such as an input button and a touch pad, an output device 2102, such as a display and a speaker, a CPU (Central Processing Unit) 2103, a ROM (Read Only Memory) 2104, and a RAM (Random). Access Memory) 2105. Further, the computer 2100 is a reading device that reads information from a recording medium such as a hard disk device, a storage device 2106 such as an SSD (Solid State Drive), a DVD-ROM (Digital Versatile Disk Read Only Memory), and a USB (Universal Serial Bus) memory. 2107, and a transmission / reception device 2108 for communicating via a network. The above-described units are connected by a bus 2109.

  Then, the reading device 2107 reads the program for realizing the function of each unit from a recording medium on which the program is recorded, and causes the storage device 2106 to store the program. Alternatively, the transmission / reception device 2108 communicates with a server device connected to the network, and causes the storage device 2106 to store a program downloaded from the server device for realizing the function of each unit.

  Then, the CPU 2103 copies the program stored in the storage device 2106 to the RAM 2105, and sequentially reads and executes the instructions included in the program from the RAM 2105, thereby realizing the functions of the above-described units. When the program is executed, information obtained by various processes described in each embodiment is stored in the RAM 2105 or the storage device 2106, and is appropriately used.

<Summary of this disclosure>
A face image selection device according to the present disclosure includes a face direction information acquisition unit that acquires face direction information indicating an angle of a face direction in a user's face image captured by an imaging device, and indicates a degree of overexposure in the face image. The image processing apparatus includes a whiteout information generation unit that generates whiteout information, and a face image selection unit that selects a face image in which the face direction information and the whiteout information satisfy predetermined conditions.

  In the face image selecting device according to the present disclosure, the face image selecting unit, based on a function defined using the face direction information and the overexposed information, from among a plurality of face images, the face direction information and The face image that satisfies the predetermined condition with the whiteout information is selected.

  In the face image selection device according to the present disclosure, an evaluation value calculation that calculates, for each of the face images, an evaluation value indicating a degree of appropriateness as a face image used for face authentication, using the face direction information and the overexposure information. A face image selection unit that selects a face image whose evaluation value is within a predetermined range as an image satisfying the predetermined condition.

  The face image selection device according to the present disclosure further includes a face direction instruction unit that instructs the user such that the angle of the face direction falls within a predetermined angle range based on the direction of the imaging device.

  In the face image selection device according to the present disclosure, when the face direction information and the whiteout information do not satisfy the predetermined condition, the face direction instruction unit instructs again to satisfy the predetermined condition.

  In the face image selecting apparatus according to the present disclosure, the face direction instructing unit instructs the angle of the face direction, the face direction information acquiring unit acquires the face direction information, and the whiteout information generating unit outputs the whiteout information. And a processing control unit that repeats a process including generating the evaluation value and calculating the evaluation value by the evaluation value calculation unit up to a specified number of times.

  In the face image selection device according to the present disclosure, the processing control unit determines whether or not there is a face image whose evaluation value is within the predetermined range. The specified number of times and / or the predetermined angle range in the processing are increased.

  In the face image selecting device according to the present disclosure, the face direction instructing unit may be configured such that, when a change occurs in the user's face direction due to the instruction, a ratio of the overexposed area is greater after the change than before the change. In this case, the user is instructed to turn the face in the direction before the change.

  In the face image selecting device according to the present disclosure, the face direction instructing unit may be configured such that, when a change occurs in the user's face direction due to the instruction, the ratio of the overexposed area is smaller after the change than before the change. , The above instruction is continued.

  INDUSTRIAL APPLICABILITY The present disclosure is useful for a face image selection device that selects a suitable face image for face authentication.

REFERENCE SIGNS LIST 10 face image selecting device 11 face image obtaining unit 12 face direction estimating unit 13 overexposed information generating unit 14 evaluation function value calculating unit 15 face image selecting unit 16 face direction instructing unit 17 processing control unit

Claims (9)

A face orientation information acquisition unit that acquires face orientation information indicating an angle of the face orientation in a user's face image captured by the imaging device,
A whiteout information generating unit that generates whiteout information indicating a degree of whiteout in the face image,
A face image selecting unit that selects a face image in which the face direction information and the overexposed information satisfy a predetermined condition,
A face image selection device having: The face image selecting unit, based on a function defined by using the face direction information and the overexposed information, the face direction information and the overexposed information from the plurality of face images, the predetermined Selecting the face image satisfying the condition,
The face image selection device according to claim 1. Using the face direction information and the overexposed information, further comprising an evaluation value calculation unit that calculates an evaluation value indicating an appropriate degree as a face image used for face authentication for each of the face images,
The face image selecting unit selects a face image whose evaluation value is within a predetermined range as an image that satisfies the predetermined condition,
The face image selection device according to claim 1. The apparatus further includes a face direction instruction unit that instructs the user such that the angle of the face direction falls within a predetermined angle range based on the direction of the imaging device.
The face image selection device according to claim 3. The face direction instructing unit, when the face direction information and the overexposed information do not satisfy the predetermined condition, instructs again to satisfy the predetermined condition,
The face image selection device according to claim 4. An instruction for the angle of the face direction by the face direction instruction unit, an acquisition of the face direction information by the face direction information acquisition unit, an generation of the overexposure information by the overexposure information generation unit, and an evaluation value calculation unit Calculating the evaluation value according to, further comprising a processing control unit that repeats the processing including the specified number of times,
The face image selection device according to claim 5. The processing control unit determines whether or not there is a face image whose evaluation value is within the predetermined range, and when it is determined that the face image does not exist, the specified number of times and / or in the processing after the determination. Increasing the predetermined angular range,
The face image selection device according to claim 6. When the face direction of the user is changed by the instruction, if the ratio of the overexposed area is larger after the change than before the change, the face direction instructing unit is configured to change the face of the user. Is again instructed to point in the direction before the change,
The face image selection device according to claim 4. The face direction instructing unit, when a change occurs in the face direction of the user due to the instruction, if the ratio of the overexposed area is smaller after the change than before the change, continue the instruction,
The face image selection device according to claim 4.

Images