Authentication apparatus, imaging apparatus, method for controlling these apparatuses, program, and storage medium

Abstract

To achieve, in authentication of a person, both prevention of reception of another person caused by takeover of an authentication state and rejection of a person himself / herself caused by re-authentication.SOLUTION: An authentication apparatus comprises: a detection unit that detects a subject from an input image; a collation unit that collates the subject detected by the detection unit with an authentication target registered in advance and outputs a collation score indicating the similarity between the subject and the authentication target; an update unit that updates an authentication score being an evaluation value indicating a degree at which the subject matches the authentication target on the basis of the collation score; and an authentication unit that authenticates the subject on the basis of the authentication score. The update unit changes a method for updating the authentication score on the basis of the magnitude relationship between the collation score and the authentication score.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present invention relates to an authentication device for authenticating a person.

Background Art

[0002] Conventionally, in imaging devices such as digital cameras, products equipped with a tracking AF (auto focus) mode have been put into practical use. The tracking AF mode is a mode in which a person's face and pupils are detected from images continuously output from an imaging element, and the focus state and exposure state are continuously optimized for the detected person's face or pupils. Further, in order to select a desired tracking target from a plurality of subjects, a technique of registering a person in advance and selecting a registered person using face authentication is described in Patent Document 1.

[0003] Generally, in the registration process of face authentication, feature information is extracted from a face image and stored in a non-volatile memory. At the time of shooting, a face is detected from images continuously output from the imaging element, and feature information of the detected face is extracted. It is determined whether the person is the registered person by comparing the similarity between this feature information and the feature information stored in the non-volatile memory.

[0004] In recent years, it has become common to use deep learning as a face authentication algorithm. While the performance has improved, the processing load has also increased. Especially in devices such as digital cameras with limited resources and using embedded software, it is difficult to perform authentication processing with a large processing load in real time within the limited resources.

[0005] Therefore, in Patent Document 2, as a method for efficiently performing authentication processing with limited resources, a method has been proposed in which a subject that has succeeded in authentication once is not subjected to collation processing thereafter, and the authentication state is taken over by tracking.

Prior Art Documents

Patent Documents

[0006]

Patent Document 1

[0007] However, in the method of taking over the authentication state by tracking, there is a possibility that a different subject may be mistakenly tracked due to the intersection or crowding of subjects, and in such a case, the authentication state will be taken over by another person (false person acceptance state). For this reason, even if a subject has been successfully authenticated once, it is necessary to perform re-matching as appropriate, and if the authentication state is in a false acceptance state, it is necessary to cancel the authentication state. On the other hand, when re-matching is performed, even if the subject should maintain the authentication state, the matching score may decrease depending on the state, and the person may be rejected (false person rejection).

[0008] The present invention has been made in consideration of the above-mentioned problems, and its purpose is to achieve both prevention of acceptance of a false identity due to handover of authentication status and prevention of rejection of the true identity due to re-authentication in person authentication. [Means for solving the problem]

[0009] The authentication device of the present invention comprises a detection means for detecting a subject from an input image, a matching means for matching the subject detected by the detection means with a pre-registered authentication target and outputting a matching score indicating the similarity between the subject and the authentication target, an update means for updating an authentication score which is an evaluation value indicating the degree to which the subject matches the authentication target based on the matching score, and an authentication means for authenticating the subject based on the authentication score, wherein the update means changes a method for updating the authentication score based on the magnitude relationship between the matching score and the authentication score. Effect of the Invention

[0010] According to the present invention, in the authentication of a person, it is possible to achieve both the suppression of acceptance of others due to the transfer of the authentication state and the suppression of rejection of the person himself / herself due to re-authentication.

Brief Description of the Drawings

[0011]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Mode for Carrying Out the Invention

[0012] Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the invention according to the claims. Although a plurality of features are described in the embodiments, not all of these plurality of features are essential to the invention, and the plurality of features may be arbitrarily combined. Further, in the accompanying drawings, the same or similar configurations are denoted by the same reference numerals, and duplicate explanations are omitted.

[0013] FIG. 1 is a diagram showing the configuration of a digital single-lens reflex camera (hereinafter, also simply referred to as a camera) 100 which is an embodiment of the imaging device of the present invention. FIG. 2 is a diagram showing the configuration related to the control of the digital single-lens reflex camera 100.

[0014] In the camera 100 of this embodiment, as shown in FIG. 1, a detachable and replaceable lens unit 120 is mounted on the front side (subject side) of the camera body 101. The lens unit 120 has a focus lens 121, a diaphragm 122, etc., and is electrically connected to the camera body 101 via a mount contact portion 123. Through this electrical connection, the control unit 201 (see FIG. 2) of the camera body 101 controls the lens unit 120, and can adjust the amount of light and the focal position taken into the camera body 101. Note that the focus lens 121 can also be manually adjusted by the user.

[0015] The image sensor 104 that captures the subject image is composed of a CCD, a CMOS sensor, etc., and includes an infrared cut filter, a low-pass filter, etc. The image sensor 104 photoelectrically converts the subject image formed by passing through the imaging optical system of the lens unit 120 during shooting, and transmits signal information for generating a captured image to the arithmetic unit 102. The arithmetic unit 102 generates a captured image from the received signal information, stores it in an external storage device 107 (see FIG. 2), and displays it on a display unit 105 such as an LCD. The shutter 103 shields the image sensor 104 during non-shooting, and opens during shooting to expose the image sensor 104.

[0016] Next, the configuration related to the control of the camera 100 will be described with reference to FIG. 2.

[0017] The arithmetic unit 102 includes a multi-core CPU, a RAM, a ROM, which can process a plurality of tasks in parallel, and a dedicated circuit for executing specific arithmetic processing at high speed. The arithmetic unit 102 has a control unit 201, a main subject arithmetic unit 202 for detecting a subject, a tracking arithmetic unit 203, a focus arithmetic unit 204, an exposure arithmetic unit 205, etc. The control unit 201 controls each part of the camera body 101 and the lens unit 120.

[0018] The main subject calculation unit 202 includes a subject detector 211 that detects a subject, a main subject determination unit 212 that performs authentication and determination of the main subject, a matching target selection unit 213 that selects a matching target, an authentication score update unit 214 that updates an authentication score, and a matching unit 215 that performs matching.

[0019] The subject detector 211 is sequentially input with continuous images acquired from the imaging device 104, and performs processing to detect one or more subject regions from each image. In this embodiment, the target subject is a human face. As the detection method, any known method such as AdaBoost or a convolutional neural network can be used. Also, as the implementation form, it may be a program running on a CPU, dedicated hardware, or a combination thereof. The subject detection result obtained from the subject detector 211 is sent to the main subject determination unit 212 and the matching target selection unit 213.

[0020] The matching target selection unit 213 selects a matching target from the detected subjects. The number of subjects to be matched at one time can be determined according to the processing speed of the matching unit 215 and the frame rate of the continuous images, and may be one or multiple. Also, there may be a frame in which no matching target is selected even though a subject is detected. The selected matching target is sent to the matching unit 215 after the face region is trimmed and scaled to a predetermined size.

[0021] When the matching unit 215 receives a face region image, it extracts feature information from the image and performs a matching process by comparing the similarity with the feature information of the subjects registered in advance as authentication targets in the database 216. Using the matching result, it performs authentication as to which registered subject the matching target is or whether there is no corresponding registered subject. When a series of matching processes is completed, the matching result is sent to the authentication score update unit 214.

[0022] The database 216 can register a plurality of subjects, and for each subject, subject information, feature information, and an image of the subject area are stored. Also, the user can set a priority for each registered subject, and the priority is stored in each subject information.

[0023] The authentication score update unit 214 updates the authentication status and authentication score of each subject using the received matching result. Based on this authentication status and subject detection result, the main subject determination unit 212 determines the main subject.

[0024] The tracking calculation unit 203 calculates the AF area and AE (auto exposure) area on the LV (live view) image (that is, the imaging element 104) so as to track the main subject determined by the main subject determination unit 212.

[0025] The focus calculation unit 204 acquires focus information (contrast evaluation value of the LV image or defocus amount of the imaging optical system) in the AF area. The control unit 201 transmits a focus instruction for controlling the position of the focus lens 121 to the lens unit 120 based on the focus information. The lens unit 120 drives the focus lens 121 in response to the focus instruction. Thereby, tracking AF as focus control for the main subject is performed.

[0026] Also, the exposure calculation unit 205 acquires luminance information in the AE area. The control unit 201 transmits an aperture instruction for controlling the opening amount of the aperture 122 to the lens unit 120 based on the luminance information. The lens unit 120 drives the aperture 122 in response to the aperture instruction. Thereby, tracking AE as exposure control for the main subject is performed.

[0027] Next, with reference to FIGS. 3 and 4, the main subject determination process and the matching process will be described in detail. FIG. 3 is a flowchart for explaining the control from image acquisition to main subject determination, and FIG. 4 is a flowchart for explaining the matching process in more detail. The processes of these flowcharts are realized by the control unit 201 expanding and executing the control program stored in the ROM in the arithmetic unit 102 in the RAM.

[0028] In step S301, the control unit 201 acquires one frame of an image from the imaging device 104 as a series of processing target images. The acquired image is converted into a resolution and file format suitable for subsequent detection processing and matching processing.

[0029] In step S302, the control unit 201 inputs the processing target image to the subject detector 211 and obtains the detection result. The detection result consists of the coordinates of the detected subject region and a score representing the reliability of the subject. The subject region is rectangular, and its coordinates can be expressed as a combination of the center coordinates and the width and height, or the upper left coordinates and the lower right coordinates. Since the detected subjects may include false detections, only the detection results with a reliability score equal to or higher than a predetermined threshold are filtered by referring to the reliability score. A unique subject ID is assigned to each detected subject. Authentication information such as the authentication target, authentication score, and authentication state is managed for each subject ID.

[0030] In this step, further, MOT (Multi Object Tracking) is performed by associating the subjects detected in the previous frame with the subjects detected in the current frame. That is, the detection results in both frames are compared, and for the detection results with similar subject regions, the subject ID is inherited from the previous frame assuming they are the same subject. As a method for determining the similarity of the subject regions, for example, IoU (Intersection over Union) is calculated, and if this is equal to or higher than a predetermined threshold, it is determined that they are similar.

[0031] In step 303, the control unit 201 determines whether a subject is detected with reference to the detection result. If no subject exists, since there is no main subject either, the control unit 201 ends the process. If a subject exists, the control unit 201 proceeds to step S304 with the process.

[0032] In step S304, the control unit 201 selects a subject to be collated in the current frame from the detected subjects. Since the collation unit 215 in the present embodiment can collate one subject per frame, when there are a plurality of detected subjects, it is necessary to select one of them. The method for selecting the collation target may be any method. For example, the frame last selected as the collation target is stored, and a method of selecting the subject for which the most time has elapsed, or a method of preferentially selecting the subject closer to the center of the image can be mentioned.

[0033] In step S305, the control unit 201 checks whether a collation target is selected in step S304. If the collation target is selected, the control unit 201 proceeds to step S306 with the process, and if not selected, the control unit 201 proceeds to step S315 with the process.

[0034] In step S306, the control unit 201 performs a collation process.

[0035] Here, the collation process in step S306 will be described with reference to FIG. 4.

[0036] In step S401, the control unit 201 reads the face region image of the subject selected in step S304. The face region image determines a trimming region with reference to the coordinates of the detection result and cuts it out from the processing target image. The cut-out image is pre-processed for subsequent processing, such as scaling to a predetermined resolution.

[0037] In step S402, the control unit 201 detects organs such as eyes, nose, and mouth using the pre-processed face region image. The organs are detected as points, and an organ point score representing their coordinates and reliability is output.

[0038] In step S403, the control unit 201 refers to the organ point score to determine whether the detected organ is valid or invalid. If the organ is valid, the control unit 201 proceeds to step S404; if not, it proceeds to step S405.

[0039] In step 404, the control unit 201 normalizes the image using the coordinates of the obtained organ points. During normalization, an affine transformation is performed so that each organ can be arranged at a predetermined position.

[0040] In step S405, the control unit 201 calculates feature information from the normalized face region image. Calculation methods for feature information include algorithms that obtain feature information based on rules from the coordinates of feature points of face parts such as eyes, nose, and mouth, and algorithms that input the image into a neural network and obtain feature information as its output. In this embodiment, the face region image is input into a pre-trained DNN (Deep Neural Network) to extract the feature information of the subject.

[0041] In step S406, the control unit 201 reads the feature information of the authentication target stored in the database 216. The authentication target is a subject pre-registered by the user. Similar to the collation target, a face region image is acquired, and after preprocessing, it is input into the DNN to obtain the feature information. Multiple authentication targets can be registered, and in this step, the feature information of all authentication targets is read.

[0042] In step S407, the control unit 201 compares the similarity between the feature information of all authentication targets obtained from the database 216 and the feature information of the collation target obtained in step S405. In this embodiment, the cosine similarity is used for similarity comparison, and a real number from -1 to +1 is output as the similarity score. The collation is performed in a 1:N manner, and for one collation target, the similarity scores with each of the N authentication targets are calculated.

[0043] In step S408, the control unit 201 searches for the authentication target with the highest similarity score from the N similarity scores, and uses this as the matching score. The control unit outputs the matching score, the authentication target corresponding to the matching score, and the organ detection result obtained in step 402 to end a series of matching processes.

[0044] Returning to Figure 3, in step S307, the control unit 201 determines whether the object to be matched is appropriate. Here, the determination is made using the organ detection result obtained as the result of the authentication process in step 306. If one or some of the detected organ points have low scores, it is possible that part of the face is hidden by an obstacle or the subject is facing backward. Therefore, each organ point score is checked, and if there is an organ point below a predetermined threshold, it is determined that the object to be matched is not appropriate, and the control unit 201 proceeds to step S308. The threshold may be determined for each organ point, or may be determined with respect to the sum or average of all organ point scores. Also, the threshold may be dynamically changed, such as relaxing the threshold for the other eye if the score of one eye is high.

[0045] In step S309, the control unit 201 checks the past matching history from the authentication information of the subject to be matched, and checks whether it has been matched at least once. If the control unit 201 has never been matched and this is the first match, it proceeds to step S310. If it has been matched, it proceeds to step S311.

[0046] In step S311, the control unit 201 determines whether the matching score obtained in the current frame from the authentication information of the subject to be matched is lower than the authentication score, which is an evaluation value indicating the degree to which the subject matches the authentication target (the magnitude relationship between the matching score and the authentication score). If the matching score is lower than the authentication score, the control unit 201 proceeds to step S312. If it is equal to or higher than the authentication score, it proceeds to step S313.

[0047] In step S312, the control unit 201 updates the authentication score using the first update method. In step S313, the control unit 201 updates the authentication score using the second update method. In step S308, the control unit 201 updates the authentication score using the third update method. In step S310, the control unit 201 updates the authentication score using the fourth update method. Details of each of the first to fourth authentication score update methods will be described later.

[0048] In step S314, the control unit 201 updates the authentication status of each subject based on the updated authentication score. The authentication status is a flag indicating whether the subject is a person registered in the database. If the authentication score of the subject to be collated is equal to or higher than a predetermined threshold, the authentication is successful; if it is less than the predetermined threshold, the authentication fails. At the same time, the authentication target is also updated. However, it is not allowed for a plurality of different subjects to be in a successful authentication state for the same authentication target. Therefore, when updating to the successful authentication state, it is necessary to check whether there is another subject in the successful authentication state for the same authentication target. If there is, the one with the higher authentication score is updated to the successful authentication state, and the one with the lower authentication score is updated to the failed authentication state.

[0049] In step S315, the control unit 201 determines the main subject using the main subject determination unit 212 based on the updated authentication status. First, the number of subjects in the successful authentication state is confirmed among all the subjects. If there is one subject in the successful authentication state, this subject is set as the main subject. If there are multiple subjects in the successful authentication state, the subject with the highest priority of the authentication target preset (designated) by the user is set as the main subject.

[0050] Subsequently, the first to fourth authentication score update methods will be described. In the first update method, the authentication score is updated by exponential moving average using the authentication score N(n - 1) in the previous frame of the subject and the collation score S(n) obtained in the current frame.

[0051] Assuming the smoothing coefficient is α, the authentication score N(n) of the current frame is obtained by N(n)=αS(n)+(1-α)N(n - 1). Here, the smoothing coefficient α is a real number between 0 and 1. The first update method is adopted when S(n)<N(n - 1) or S(n)≦N(n - 1) is satisfied. Thereby, even when an extremely low matching score is obtained instantaneously, the decrease in the authentication score becomes gentle, so it is possible to suppress the rejection of the user due to an incorrect authentication failure state.

[0052] In the second update method as well, the authentication score is updated by exponential moving average in the same way. However, the smoothing coefficient is β which is different from α, and the authentication score is updated by N(n)=βS(n)+(1-β)N(n - 1). In this case, α < β. Thereby, when the authentication score decreases, the change becomes gentle, while when the authentication score increases, the change becomes sensitive, so it is possible to quickly transition to the authentication success state.

[0053] As described in step S302, the subject is performing MOT across the frame. However, in cases where the subjects are interleaved or crowded, mis - tracking may occur, and the authentication success state may be inherited by the wrong subject, resulting in an unauthorized acceptance state. By providing the second update method, even when an unauthorized acceptance state occurs, it is possible to quickly return to the correct authentication state. β may be set to 1, in which case the matching score is directly adopted as the authentication score.

[0054] The third update method is a method of updating the authentication score without using the matching score, and the score is updated by subtraction or multiplication using a predetermined value γ. When subtraction is performed, an arbitrary real number γ is used, and it is updated by N(n)=N(n - 1)-γ. When multiplication is performed, a real number γ between 0 and 1 is used, and it is updated by N(n)=γN(n - 1). The following description assumes that subtraction is performed.

[0055] The third update method is selected when the collation target is inappropriate. At this time, since the collation score is unreliable, it is necessary to update the authentication score without using the collation score. As time passes, the risk of false tracking occurring in the MOT increases, so it is appropriate to lower the authentication score when the collation score cannot be used.

[0056] The fourth update method is represented by N(n)=S(n), which uses the collation score as the authentication score as it is. The fourth update method is selected when the collation target is the first collation. In this case, since N(n - 1) is an indefinite value, it cannot be used. This applies to a newly framed subject, etc. In this case, it is possible to improve user usability by immediately raising the authentication score and quickly achieving the authentication success state.

[0057] Based on the above, the changes in the authentication score and authentication state will be described using the specific examples shown in FIGS. 5 and 6. However, the collation score and authentication score are integers from 0 to 1000, and the threshold value of the authentication score is 500. Assume that subjects A to C appear, and only subject A is registered in the database. The smoothing coefficient α of the first update method is 0.3, the smoothing coefficient β of the second update method is 0.9, and the predetermined value γ of the third update method is 10. Also, a rectangular frame is drawn for the main subject in each frame.

[0058] FIG. 5 is a diagram showing a specific example of a scene in which the rejection of the genuine user can be suppressed as an effect of the present embodiment. However, for the sake of easy understanding of the explanation, frames in which subjects other than subject A are the collation target are omitted, and subject A is selected as the collation target in all frames in which subject A exists.

[0059] In frame 1, there is no authentication target in the processing target image. In this case, since the main subject is determined according to the composition and the size of the subject, subject B is selected as the main subject.

[0060] In frame 2, subject A enters the image to be processed and is selected as the object to be verified. As a result of the authentication process, a matching score of 800 is obtained. Since this is the first verification for subject A, the fourth update method is selected, and the authentication score also becomes the same as the matching score, which is 800. Since the authentication score of subject A exceeds the threshold of 500, subject A is in a successful authentication state and becomes the main subject.

[0061] In frame 3, since a part of subject A is hidden, the matching score drops to 400. Since it is lower than the authentication score of 800 in the previous frame, the first update method is selected. From the formula of the exponential moving average with a smoothing factor of 0.3, the authentication score drops to 680. Although the matching score is lower than the threshold of 500, since the authentication score exceeds the threshold, the successful authentication state can be maintained, and subject A continues to be the main subject.

[0062] In frame 4, the hidden area of subject A further increases, and the matching score drops to 20. At this time, since the left eye is completely hidden, the organ point score drops below the threshold. As a result, it is determined that the object to be verified is not appropriate, and the third update method is selected.

[0063] In this case, the authentication score is updated to 670 by subtracting a predetermined value of 10 from the authentication score of 680 in the previous frame. If the first update method had been selected in this frame, the authentication score would have been 482, which is lower than the threshold, resulting in a failed authentication state, and the main subject would have changed to subject B. That is, because the third update method was selected, subject A can maintain the successful authentication state.

[0064] In frame 5, the hidden state of subject A is resolved, and the matching score becomes 800 again. Since it exceeds the authentication score of 670 in the previous frame, the second update method is selected. From the formula of the exponential moving average with a smoothing factor of 0.9, the authentication score rises to 787. Even if the authentication score had been lower than the threshold in frame 3 or frame 4, due to the large smoothing factor, it is possible to transition to the successful authentication state in frame 5, and subject A can become the main subject again.

[0065] FIG. 6 is a diagram showing a specific example of a scene in which acceptance of others can be suppressed as an effect of the present embodiment.

[0066] In frame 1, it is assumed that the subject A being tracked as ID:1 is selected as the collation target. In this frame, the authentication score of the subject A is updated by the fourth update method and becomes the authentication success state.

[0067] In frame 2, the collation target switches to the subject B being tracked as ID:2. Since the subject B is not registered in the database and will be compared with the feature information of the subject A in the database, the collation score shows a low value. The authentication score is updated by the fourth update method.

[0068] In frame 3, the subject A and the subject B intersect. In this case, since only one subject is detected, either the subject with ID:1 or ID:2 disappears. Here, the main subject in the previous frame is prioritized and ID:1 is left. As a result, the subject B is recognized as ID:1 and false tracking occurs. The collation target is the subject B, and it shows a low collation score as in the previous frame. At this time, the authentication score is updated by the second update method. However, since false tracking has occurred, the authentication score in the previous frame inherits the score of the subject A. As a result of the calculation by the exponential moving average, the authentication score becomes 710. Since this authentication score exceeds the threshold value of 500, it becomes the authentication success state and the main subject. Since the subject B is not actually a registered subject, it is in the state of accepting others.

[0069] In frame 4, subject A reappears, and ID:3 is assigned as a new subject. In this frame, subject A is selected as the collation target. The authentication score is updated by the fourth update method, and the authentication score becomes 750. Since this score exceeds the threshold of 500, it is possible to transition to the authentication success state. However, since subject B was also in the authentication success state in the previous frame, one of them needs to be set to the authentication failure state. When comparing the authentication scores of each subject, since the authentication score of subject A is higher, subject A is set to the authentication success state and subject B is set to the authentication failure state. By reducing the authentication score of the subject in the acceptance state of others by the second update method in frame 3 and directly accepting the collation score as the authentication score by the fourth update method in frame 4, when the correct authentication target appears, it becomes possible to immediately correct the acceptance state of others.

[0070] In this example, since only subject A was registered in the database as the authentication target, only one subject could achieve the authentication success state. However, if multiple subjects are registered in the database, there will be a subject in the authentication success state for each registered authentication target. That is, if N subjects are registered in the database, at most N subjects will be in the authentication success state. In this case, the primary subject is determined according to the priority order of the subjects registered in the database set by the user in advance.

[0071] As described above, in this embodiment, in the face authentication system, one of a plurality of update methods for the authentication score is selected according to the situation, and the authentication score in the current frame is updated using the collation score and the authentication score in the previous frame. As a result, it becomes possible to stably and sensitively transition to the authentication success state in response to a temporary decrease in the collation score due to changes in the subject's expression, hiding, etc., and it becomes possible to achieve both suppression of accepting others and suppression of rejecting the user himself.

[0072] The disclosure of this specification includes the following authentication device, authentication method, imaging device and its control method, program, and storage medium.

[0073] (Item 1) Detection means for detecting a subject from the input image, Matching means for comparing the subject detected by the detection means with a pre-registered authentication target and outputting a matching score indicating the degree of similarity between the subject and the authentication target, Updating means for updating an authentication score, which is an evaluation value indicating the degree to which the subject matches the authentication target, based on the matching score, Authentication means for authenticating the subject based on the authentication score, comprising, The updating means is characterized by changing the updating method of the authentication score based on the magnitude relationship between the matching score and the authentication score. An authentication device.

[0074] (Item 2) The detection means is characterized by detecting a subject from each of the input continuous images. The authentication device according to Item 1.

[0075] (Item 3) The authentication device according to Item 1 or 2, further comprising selection means for selecting an object to be compared with the authentication target from the subject detected by the detection means.

[0076] (Item 4) The updating means updates the authentication score using a first updating method when the matching score is smaller than the authentication score, and updates the authentication score using a second updating method when the matching score is equal to or greater than the authentication score. The authentication device according to any one of Items 1 to 3.

[0077] (Item 5) The first updating method and the second updating method are characterized by updating the authentication score by an exponential moving average with different smoothing coefficients. The authentication device according to Item 4.

[0078] (Item 6) The authentication device according to item 5, wherein the smoothing coefficient in the first update method is smaller than the smoothing coefficient in the second update method.

[0079] (Item 7) The authentication device according to any one of items 1 to 6, wherein when the collation score of the subject cannot be obtained, the update means updates the authentication score using a third update method.

[0080] (Item 8) The authentication device according to item 7, wherein in the third update method, the authentication score is updated by multiplying the authentication score by a predetermined coefficient.

[0081] (Item 9) The authentication device according to item 7, wherein in the third update method, the authentication score is updated by subtracting a predetermined value from the authentication score.

[0082] (Item 10) The authentication device according to any one of items 1 to 9, wherein for a subject for which the collation with the authentication target is performed for the first time, the update means updates the authentication score by a fourth update method in which the collation score is used as the authentication score.

[0083] (Item 11) A detection step of detecting a subject from the input image, A collation step of collating the subject detected in the detection step with a pre-registered authentication target and outputting a collation score indicating the similarity between the subject and the authentication target, An update step of updating an authentication score, which is an evaluation value indicating the degree of coincidence between the subject and the authentication target, based on the collation score, An authentication step of authenticating the subject based on the authentication score, and In the update step, an authentication method characterized by changing the update method of the authentication score based on the magnitude relationship between the collation score and the authentication score.

[0084] (Item 12) Detection means for detecting a subject from a captured image, Matching means for matching the subject detected by the detection means with a pre-registered authentication target and outputting a matching score indicating the degree of similarity between the subject and the authentication target, Updating means for updating an authentication score, which is an evaluation value indicating the degree to which the subject matches the authentication target, based on the matching score, Determination means for determining a main subject based on the authentication score, and comprising, The updating means changes the updating method of the authentication score based on the magnitude relationship between the matching score and the authentication score. An imaging device characterized by this.

[0085] (Item 13) The detection means detects a subject from each of the captured consecutive images. The imaging device according to Item 12, characterized by this.

[0086] (Item 14) The imaging device according to Item 12 or 13, further comprising selection means for selecting a target to be matched with the authentication target from the subject detected by the detection means.

[0087] (Item 15) The updating means updates the authentication score using a first updating method when the matching score is smaller than the authentication score, and updates the authentication score using a second updating method when the matching score is equal to or greater than the authentication score. The imaging device according to any one of Items 12 to 14, characterized by this.

[0088] (Item 16) The first updating method and the second updating method update the authentication score by exponential moving averages with different smoothing coefficients. The imaging device according to Item 15, characterized by this.

[0089] (Item 17) The imaging device according to item 16, wherein the smoothing coefficient in the first update method is smaller than the smoothing coefficient in the second update method.

[0090] (Item 18) The imaging device according to any one of items 12 to 17, wherein when the update means cannot obtain the collation score of the subject, the authentication score is updated using a third update method.

[0091] (Item 19) The imaging device according to item 18, wherein in the third update method, the authentication score is updated by multiplying the authentication score by a predetermined coefficient.

[0092] (Item 20) The imaging device according to item 18, wherein in the third update method, the authentication score is updated by subtracting a predetermined value from the authentication score.

[0093] (Item 21) The imaging device according to any one of items 12 to 20, wherein when the update means collates with the subject to be authenticated for the first time, the authentication score is updated by a fourth update method in which the collation score is used as the authentication score.

[0094] (Item 22) The imaging device according to any one of items 12 to 21, wherein when a plurality of subjects are authenticated as the same authentication target, the determination means determines the subject with the higher authentication score as the main subject.

[0095] (Item 23) The imaging device according to any one of items 12 to 21, wherein when a plurality of subjects are authenticated as different authentication targets, the determination means determines the subject authenticated as the authentication target with the higher priority specified by the user as the main subject.

[0096] (Item 24) The imaging device according to any one of items 12 to 23, characterized in that focus control or exposure control is preferentially performed on the subject determined as the main subject by the determination means.

[0097] (Item 25) A detection step of detecting a subject from the captured image, A collation step of collating the subject detected in the detection step with a pre-registered authentication target and outputting a collation score indicating the similarity between the subject and the authentication target, An update step of updating an authentication score, which is an evaluation value indicating the degree of coincidence between the subject and the authentication target, based on the collation score, A determination step of determining the main subject based on the authentication score, having, In the update step, the method for updating the authentication score is changed based on the magnitude relationship between the collation score and the authentication score. A control method for an imaging device.

[0098] (Item 26) A program for causing a computer to execute each step of the method according to item 11 or 25.

[0099] (Item 27) A computer-readable storage medium storing a program for causing a computer to execute each step of the method according to item 11 or 25.

[0100] (Other embodiments) The present invention can also be realized by supplying a program that realizes one or more functions of the above-described embodiments to a system or device via a network or a storage medium, and having one or more processors in the computer of the system or device read and execute the program. It can also be realized by a circuit (for example, an ASIC) that realizes one or more functions.

[0101] The invention is not limited to the above embodiments, and various changes and modifications can be made without departing from the spirit and scope of the invention. Therefore, the claims are appended to disclose the scope of the invention.

Explanation of Signs

[0102] 100: Digital single-lens reflex camera, 102: Arithmetic unit, 104: Image sensor, 201: Control unit, 202: Main subject arithmetic unit, 211: Subject detector, 212: Main subject determination unit, 213: Collation target selection unit, 214: Authentication score update unit, 215: Collation unit, 216: Database

Claims

1. Detection means for detecting a subject from an input image, Collation means for collating the subject detected by the detection means with a pre-registered authentication target and outputting a collation score indicating the degree of similarity between the subject and the authentication target, Updating means for updating an authentication score, which is an evaluation value indicating the degree to which the subject matches the authentication target, based on the collation score, Authentication means for authenticating the subject based on the authentication score, Comprising, The updating means is characterized in that it changes the method for updating the authentication score based on the magnitude relationship between the collation score and the authentication score. An authentication device.

2. The detection means is characterized in that it detects a subject from each of the input continuous images. The authentication device according to claim 1.

3. The authentication device according to claim 1, further comprising selection means for selecting a target to be collated with the authentication target from the subject detected by the detection means.

4. The updating means updates the authentication score using a first updating method when the collation score is smaller than the authentication score, and updates the authentication score using a second updating method when the collation score is equal to or greater than the authentication score. The authentication device according to claim 1, characterized by this.

5. The first updating method and the second updating method are characterized in that the authentication score is updated by an exponential moving average with different smoothing coefficients. The authentication device according to claim 4.

6. The authentication device according to claim 5, characterized in that the smoothing coefficient in the first updating method is smaller than the smoothing coefficient in the second updating method.

7. The updating means is characterized in that when the collation score of the subject cannot be obtained, the authentication score is updated using a third updating method. The authentication device according to claim 1.

8. In the third updating method, the authentication score is updated by multiplying the authentication score by a predetermined coefficient. The authentication device according to claim 7, characterized by this.

9. In the third updating method, the authentication score is updated by subtracting a predetermined value from the authentication score. The authentication device according to claim 7, characterized by this.

10. For a subject for which collation with the authentication target is performed for the first time, the updating means updates the authentication score by a fourth updating method in which the collation score is set as the authentication score. The authentication device according to claim 1, characterized by this.

11. A detection step of detecting a subject from the input image; A matching step of matching the subject detected in the detection step with a pre-registered authentication target and outputting a matching score indicating the similarity between the subject and the authentication target; An update step of updating an authentication score, which is an evaluation value indicating the degree to which the subject matches the authentication target, based on the matching score; An authentication step of authenticating the subject based on the authentication score; comprising; In the update step, the authentication method is characterized in that the update method of the authentication score is changed based on the magnitude relationship between the matching score and the authentication score.

12. Detection means for detecting a subject from the captured image; Matching means for matching the subject detected by the detection means with a pre-registered authentication target and outputting a matching score indicating the similarity between the subject and the authentication target; Update means for updating an authentication score, which is an evaluation value indicating the degree to which the subject matches the authentication target, based on the matching score; Determination means for determining a main subject based on the authentication score; comprising; The update means is characterized in that the update method of the authentication score is changed based on the magnitude relationship between the matching score and the authentication score. An imaging device.

13. The imaging device according to claim 12, wherein the detection means detects a subject from each of the captured consecutive images.

14. The imaging device according to claim 12, further comprising selection means for selecting a target to be matched with the authentication target from the subject detected by the detection means.

15. The imaging device according to claim 12, wherein the update means updates the authentication score using a first update method when the matching score is smaller than the authentication score, and updates the authentication score using a second update method when the matching score is equal to or greater than the authentication score.

16. The imaging device according to claim 15, wherein the first update method and the second update method update the authentication score by exponential moving averages with different smoothing coefficients.

17. The imaging device according to claim 16, wherein the smoothing coefficient in the first update method is smaller than the smoothing coefficient in the second update method.

18. The imaging device according to claim 12, wherein the updating means updates the authentication score using a third updating method when the matching score of the subject cannot be obtained.

19. The imaging device according to claim 18, wherein in the third updating method, the authentication score is updated by multiplying the authentication score by a predetermined coefficient.

20. The imaging device according to claim 18, wherein in the third updating method, the authentication score is updated by subtracting a predetermined value from the authentication score.

21. The imaging device according to any one of claims 12 to 20, wherein the updating means updates the authentication score by a fourth updating method in which the matching score is set as the authentication score for a subject for which the matching with the authentication target is performed for the first time.

22. The imaging device according to claim 12, wherein when a plurality of subjects are authenticated as the same authentication target, the determination means determines the subject with the higher authentication score as the main subject.

23. The imaging device according to claim 12, wherein when a plurality of subjects are authenticated as different authentication targets, the determination means determines the subject authenticated as the authentication target with the higher priority specified by the user as the main subject.

24. The imaging device according to claim 12, wherein priority focus control or exposure control is preferentially performed on the subject determined as the main subject by the determination means.

25. A detection step of detecting a subject from the captured image; A matching step of matching the subject detected in the detection step with a pre-registered authentication target and outputting a matching score indicating the similarity between the subject and the authentication target; An updating step of updating an authentication score, which is an evaluation value indicating the degree of coincidence between the subject and the authentication target, based on the matching score; A determination step of determining a main subject based on the authentication score; and In the updating step, a control method of an imaging device is characterized in that the updating method of the authentication score is changed based on the magnitude relationship between the matching score and the authentication score.

26. A program for causing a computer to execute each step of the method according to claim 11 or 25.

27. A computer-readable storage medium storing a program for causing a computer to execute each step of the method according to claim 11 or 25.

Images