Object recognition device, control method for object recognition device

Abstract

To shorten the time taken for face recognition, so that a user can start work early.SOLUTION: A face recognition device (10) includes: a collation unit (14) that collates a plurality of partial captured images obtained by extracting an image of a predetermined portion of a subject from the captured image of the subject with a plurality of partial registered images, which are images of predetermined portions of a recognition subject stored in a storage unit (12), in order according to the priority associated with each partial registered image; and a recognition output unit (15) that outputs a recognition result of the subject when the collation in the collation unit satisfies a predetermined condition.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an object recognition device.

Background Art

[0002] Conventionally, there has been a face recognition system that performs face recognition of a user and identifies whether the user has the authority to perform a specific operation. For example, the face recognition system is used as a permission device for entering and leaving a highly secure area or a device that gives permission to use a specific device.

[0003] Patent Document 1 discloses a technique that enables normal face recognition even when the user is wearing a mask or has changed their hairstyle. By emphasizing the feature amount of a certain part and reducing the feature amount of other parts, normal face recognition can be performed even when there are differences from the data registered in the database.

[0004] Patent Document 2 is an invention for face recognition when the captured image of the user's face is different from the image registered in the database due to a mask, sunglasses, a hat, etc. As described in the above example, when there is a difference between the captured image and the registered image in the database, instead of immediately failing face recognition, the captured image is divided into a plurality of areas, and the user is notified of the area where the collation has failed to prompt re-recognition, thereby achieving successful face recognition.

[0005] Patent Document 3 is an invention in a digital camera. By defining the priority of a subject for each shooting mode, the subject can be preferentially recognized, and optical adjustment can be performed early according to the subject.

[0006] Patent Document 4 is an invention for reducing the capacity of a biological pattern image of a living body such as a vein pattern. By recognizing similar patterns for each region divided into a plurality of regions and storing the numbers corresponding to the patterns for each region.

[0007] Patent Document 5 uses a low-resolution image to perform face recognition. When it is determined that the images are the same, face recognition is performed on a part of a higher-resolution image, thereby gradually improving the accuracy of face recognition while ensuring speed.

Prior Art Documents

Patent Documents

[0008]

Patent Document 1

Patent Document 2

Patent Document 3

Patent Document 4

Patent Document 5

Summary of the Invention

Problems to be Solved by the Invention

[0009] However, the above-described conventional technologies take time for face recognition due to the size of the face part to be collated in face recognition or the number of face recognition targets. Therefore, the user cannot move on to the work early, which causes stress to the user.

[0010]

[0011] Therefore, one aspect of the present invention aims to realize an apparatus that can perform object recognition at high speed while maintaining recognition accuracy.

Means for Solving the Problems

[0012] To solve the above problems, an object recognition apparatus according to an aspect of the present invention includes an image acquisition unit that acquires a captured image obtained by capturing a target object, an image extraction unit that creates a plurality of partial captured images by extracting images of a plurality of predetermined parts of the target object from the captured image, a plurality of partial registration images obtained by extracting images of a plurality of predetermined parts of a registered object as a recognition target stored in a storage unit, and a collation unit that collates the partial registration images with the partial captured images in an order according to priorities respectively associated with the partial registration images, and a recognition output unit that outputs a recognition result of the target object when the collation in the collation unit satisfies a predetermined condition.

[0013] According to the above configuration, by repeatedly collating between the partial captured image and the partial registration image in the order of the priorities associated with the partial registration images until a predetermined condition is satisfied, it is possible to determine whether the target object is a registered object. Since the recognition result is output when the predetermined condition is satisfied, the recognition result of the target object can be output at an early stage.

[0014] The collation unit may perform collation by calculating a similarity between the partial registration image and the partial captured image and comparing it with a predetermined threshold value.

[0015] According to the above configuration, the partial registration image and the partial captured image can be collated by comparing the similarity with the threshold value.

[0016] When the collation unit performs collation in the order according to the priorities, and determines that the partial registration image and the partial captured image corresponding to the partial registration image substantially match, the recognition output unit may output the recognition result of the target object.

[0017] According to the above configuration, when the collation unit determines that there is a partial captured image that substantially matches the partial registration image, the recognition result that the target object is a registered object can be output.

[0018] The memory unit stores, for a plurality of the registered objects, the partial registered images and the priorities of the partial registered images. The matching unit may perform matching for the partial registered images of all the registered objects at a certain priority, and then perform matching for the partial registered images of all the registered objects at the next priority.

[0019] According to the above configuration, for all the registered objects, matching of the partial captured images is performed in the order of priorities, and when it is determined that there is a substantial match, a recognition result that the target object is a registered object can be output.

[0020] When the matching unit performs matching in the order according to the priorities and determines that the partial registered image and the partial captured image corresponding to the partial registered image do not substantially match, the recognition output unit may output the recognition result of the target object.

[0021] According to the above configuration, when the matching unit determines that there is a partial captured image that does not substantially match the partial registered image, the recognition result that the target object is not a registered object can be output.

[0022] The memory unit stores, for a plurality of the target objects, the partial registered images and the priorities of the partial registered images. The matching unit performs matching for the partial registered images of all the registered objects at a certain priority, and then performs matching for the partial registered images of all the registered objects at the next priority. When it is determined that the partial registered image and the partial captured image corresponding to the partial registered image do not substantially match, the matching for the corresponding registered object at the next priority may not be performed.

[0023] According to the above configuration, for all the registered objects, matching of the partial captured images is performed in the order of priorities, and when it is determined that all the registered objects do not substantially match, a recognition result that the target object is not a registered object can be output.

[0024] The matching unit may further include a priority setting unit that performs a process of storing the calculation result of the similarity as a similarity history and sets the priority of the partial registration image based on the similarity history.

[0025] According to the above configuration, the priority of the partial registration image can be automatically set according to the similarity history. Therefore, by setting the priority of the partial registration image according to the trend of the similarity, it is possible to output the recognition result of whether the target object is a registered object earlier.

[0026] The apparatus may further include a priority setting unit that sets the priority of the partial registration image based on a user input.

[0027] According to the above configuration, the priority of the partial registration image can be arbitrarily set based on the user input. Therefore, it is possible to change the priority of a part for which the priority is intentionally lowered or a part for which the priority is intentionally raised.

[0028] The target object may be a human face.

[0029] According to the above configuration, for example, the object may be a human face and can be used when performing face recognition at an early stage.

[0030] An object recognition apparatus according to another aspect includes an image acquisition step of acquiring a captured image of a target object, an image extraction step of creating a plurality of partial captured images by extracting images of a plurality of predetermined parts of the target object from the captured image, a plurality of partial registration images obtained by extracting images of a plurality of predetermined parts of a registered object as a recognition target stored in a storage unit, and a matching step of performing matching between the partial captured images in an order according to priorities respectively associated with the partial registration images, and a recognition output step of outputting a recognition result of the target object when the matching in the matching unit satisfies a predetermined condition.

[0031] The object recognition device according to each aspect of the present invention may be implemented by a computer. In this case, an object recognition program of the object recognition device for implementing the object recognition device by operating the computer as each part (software element) included in the object recognition device, and a computer-readable recording medium on which the program is recorded also fall within the scope of the present invention.

Effects of the Invention

[0032] According to one aspect of the present invention, a partial captured image extracted from a captured image and a pre-registered partial registered image are collated with a priority associated with the partial registered image, and when the collation result satisfies a predetermined condition, recognition can be terminated early.

Brief Description of the Drawings

[0033]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Modes for Carrying Out the Invention

[0034] [Embodiment 1] Hereinafter, an embodiment according to one aspect of the present invention (hereinafter also referred to as "this embodiment") will be described with reference to the drawings. In the drawings, the same or corresponding parts are denoted by the same reference numerals and their description will not be repeated.

[0035] §1. Application Example FIG. 1 is a block diagram showing the configuration of a main part of a face recognition system 100 according to Embodiment 1. The face recognition system 100 includes a camera 20 that captures a captured image including a face, a face recognition device (object recognition device) 10 that performs face recognition on the captured image, and an electromagnetic lock 30 controlled by the face recognition device 10.

[0036] The face recognition device 10 extracts a partial captured image corresponding to a pre-registered partial registered image from the captured image, and performs collation by comparing the partial captured image and the partial registered image. Here, a plurality of partial registered images are prepared, and a priority is set for each partial registered image. The collation for each partial registered image is performed in the order according to the priority.

[0037] The face recognition device 10 has roughly two modes, a mode that prioritizes the recognition speed of an authorized person who is recognized as an authorized person for face recognition, and a mode that prioritizes the recognition speed of a non-authorized person who is recognized as not being an authorized person for face recognition. In the mode that prioritizes the recognition speed of an authorized person, when a pair of a set of partial captured images and partial registered images can be recognized as the same image, it is recognized as substantially the same face. On the other hand, in the mode that prioritizes the recognition speed of a non-authorized person, when a pair of a set of partial captured images and partial registered images can be recognized as different images, it is recognized as not being substantially the same face. When recognized as substantially the same face, the face recognition device 10 releases the electromagnetic lock 30.

[0038] That is, the face recognition device 10 performs collation in the order according to the priority, and can output the face recognition result when the recognition result of a set of partial captured images and partial registered images is obtained. Therefore, the recognition result can be output earlier than when performing face recognition on the entire captured image.

[0039] §2. Configuration Example (Configuration of Face Recognition Device) Based on FIG. 1, the configuration of the face recognition system 100 will be described. The face recognition device 10 includes an image acquisition unit 11, a storage unit 12, an image extraction unit 13, a matching unit 14, and a recognition output unit 15.

[0040] The camera 20 is a camera that captures a captured image including a face to be subjected to face recognition. The camera 20 outputs the captured image to the image acquisition unit 11.

[0041] The electromagnetic lock 30 is an electronic lock controlled by the face recognition device 10. The electromagnetic lock 30 is not limited to an electronic lock, and can be any program-controlled device or program, etc. If it is not unlocked, the functions of the device or program, etc. cannot be used.

[0042] For example, when the electromagnetic lock 30 is provided on a door, it is impossible to open the door and enter the interior unless the electromagnetic lock 30 is unlocked by face recognition. Also, in an electronic device incorporating the camera 20 such as a smartphone, the smartphone cannot accept operations unless the electromagnetic lock 30 is unlocked (the electromagnetic lock 30 is released) by face recognition.

[0043] The image acquisition unit 11 inputs the captured image. The image acquisition unit 11 outputs the input captured image to the image extraction unit 13.

[0044] The storage unit 12 stores a plurality of partial registration images for each registered person who is the recognition target of face recognition. Also, the storage unit 12 stores the priority together with each partial registration image, and the order of performing face recognition on the partial registration images is determined. Note that the information on the partial registration images and the priority is stored in a storage unit external to the face recognition device 10, and the face recognition device 10 may read the information on the partial registration images and the priority via various communication means.

[0045] The image extraction unit 13 extracts a portion corresponding to the partial registration image from the input captured image, and uses it as a partial captured image. The image extraction unit 13 outputs the extracted multiple partial captured images to the collation unit 14.

[0046] The collation unit 14 compares the input partial captured image with the partial registration image, and calculates a numerical value called the similarity indicating the degree of similarity between them. Then, the collation unit 14 determines whether the images are substantially the same by comparing the similarity with a predetermined threshold value. Also, the partial registration image of a certain priority is collated with the partial captured image corresponding to the partial registration image, and if necessary, the partial registration image of the next priority and the partial captured image corresponding to the partial registration image are collated. After performing the collation necessary for face recognition and the collation result is determined, the collation unit 14 outputs the collation result to the recognition output unit 15.

[0047] The recognition output unit 15 determines whether to issue a signal to unlock the electromagnetic lock 30 upon receiving the input collation result. When the collation result that the subject of face recognition is a registered person is obtained, the recognition output unit 15 unlocks the electromagnetic lock 30. On the contrary, when the collation result that the subject of face recognition is not a registered person is obtained, the recognition output unit 15 does not unlock the electromagnetic lock 30.

[0048] (Extraction of partial captured image in captured image) FIG. 2 is a model diagram showing the captured image 111 and the partial captured image 112 according to Embodiment 1. As shown in FIG. 2, the captured image 111 is an image in which the entire face of the subject to be face-recognized is shown. The partial captured image 112 is, for example, an image of a part of five captured images 111, and is the partial captured images 112a to 112e. Each partial captured image is an image that partially shows a part representing the facial features.

[0049] For example, the partial captured image 112a captures both eyes, the partial captured image 112b captures the mouth, the partial captured image 112c captures the hair, the partial captured image 112d captures the nose, and the partial captured image 112e captures the forehead. The ranges of the partial captured images 112 (collectively referred to as the partial captured images 112 including the partial captured images 112a to 112e) may overlap with each other. For example, in the example of FIG. 2, both the partial captured image 112a and the partial captured image 112e include eyebrows.

[0050] The partial captured image 112 is extracted from the captured image 111 by the image extraction unit 13, and the extracted range is the range corresponding to the partial registration image.

[0051] That is, the image extraction unit 13 extracts the partial captured images 112 corresponding to the partial registration images from the captured image 111. At this time, in order to extract the partial captured image 112, pattern matching may be performed on the captured image 111 to derive the range of the partial captured image 112.

[0052] Moreover, the method for extracting the partial captured image 112 is not limited to this method. The position of the partial area to be extracted from the captured image 111 may be determined in advance, and the corresponding partial area may be extracted as the partial captured image 112.

[0053] (Method for registering partial registration image) FIG. 3 is a flowchart showing the process of registering the partial registration image. In order to perform face recognition, it is necessary to register in advance the partial registration images obtained by dividing the face image by part.

[0054] In S11, the camera 20 captures the face image of the registrant to be registered. In S12, the captured face image is extracted for each part of the face. The parts of the face include eyes, mouth, nose, eyebrows, forehead, hair, etc., and the image may be extracted so that the periphery is also captured for each part. This operation may be performed by the user operating the face recognition device 10, or automatically by the face recognition device 10 according to a predetermined rule.

[0055] In S13, for each divided part, it is registered in the storage unit 12 as a partial registered image. In S14, for each partial registered image, a priority indicating the order for performing face recognition is assigned. Here, the assignment of priority may be manually performed by the user, or may be automatically performed by the face recognition device 10 according to a predetermined rule.

[0056] §3. Operation Example Hereinafter, four types of operation examples will be described with specific examples.

[0057] (Operation Example 1: Prioritize the Personal Recognition Speed for One Registered Person) FIG. 4 is a flowchart showing the operation flow of Operation Example 1 according to Embodiment 1. Operation Example 1 is an operation that prioritizes the personal recognition speed for the registered person of the partial registered image when the partial registered image to be recognized is a face image of one person. For simplicity of explanation, it is assumed that there are five partial registered images.

[0058] In S21, the camera 20 captures a face image, and the image acquisition unit 11 acquires the face image as a captured image.

[0059] In S22, the recognition output unit 15 performs a collation loop for collating the partial registered images for each part from i = 1 to i = 5. Here, i is a natural number indicating the priority of the partial registered image.

[0060] In S23, the image extraction unit 13 extracts a partial captured image corresponding to the partial registered image with the i-th priority.

[0061] In S24, the collation unit 14 compares and collates the partial captured image with the partial registered image with the i-th priority. When the partial captured image and the partial registered image substantially match, it proceeds to S25. When the partial captured image and the partial registered image do not substantially match, it proceeds to S26.

[0062] In S25, since the matching unit 14 determines that the partial captured image and the partial registered image substantially match, it recognizes that the subject is the same person as the single registrant and exits the matching loop. Thereafter, the recognition output unit 15 unlocks the electromagnetic lock 30.

[0063] In S26, when i = 5, the matching unit 14 exits the matching loop or increments i = i + 1 and moves to the next matching loop.

[0064] In S27, since the recognition output unit 15 determines that the subject is not the same person as the registrant, it recognizes that the subject is a different person from the registrant. Therefore, the recognition output unit 15 does not unlock the electromagnetic lock 30.

[0065] According to the above processing, when it is determined that one partial captured image substantially matches the partial registered image, it is possible to quickly recognize that the subject is the same person as the registrant. Also, in order to ensure security, it is preferable to set a strict threshold for the similarity for determining that the partial captured image and the partial registered image substantially match each other. By setting the threshold strictly, it is possible to suppress the possibility that a subject other than the registrant is recognized as the same person by face recognition.

[0066] Here, in S24, conditional branching is performed depending on whether or not one partial captured image substantially matches the partial registered image. When the condition is satisfied (Yes in S24), the electromagnetic lock 30 is released (S25), but it is not necessarily released based on the recognition result of one partial captured image. That is, when a predetermined number of partial captured images substantially match the partial registered image, it may be recognized that the subject is the same person as the single registrant. By increasing the number of partial registered images required to determine that the person is the same, security is improved. Here, the number of partial registered images required to determine that the person is the same is preferably less than the number of partial registered images stored in the storage unit 12. Thereby, compared with the case of collating with all partial registered images, the collation result can be output at high speed.

[0067] As an actual example of Operation Example 1, it is preferably used when performing face recognition with a device used basically only by the registrant. For example, in an electronic device such as a smartphone owned by an individual, there is a case where face recognition is used for user recognition.

[0068] (Operation Example 2: Prioritize the recognition speed of others for one registrant) FIG. 5 is a flowchart showing the operation flow of Operation Example 2 according to Embodiment 1. Operation Example 2 is an operation that prioritizes the recognition speed of others who are recognized as not being the registrant of the partial registration image when the partial registration image to be subjected to face recognition is a face image of one person. For the sake of simplicity of explanation, it is assumed that there are five partial registration images.

[0069] The processes of S21 to S23 are the same as those in Operation Example 1, so the description thereof is omitted.

[0070] In S34, the collation unit 14 compares and collates the partial captured image with the partial registration image having the i-th priority. When the partial captured image and the partial registration image substantially match, the process proceeds to S36. When the partial captured image and the partial registration image do not substantially match, the process proceeds to S35.

[0071] In S35, since the partial captured image and the partial registration image do not substantially match, the collation unit 14 recognizes that the target person is a person different from the single registrant and exits the collation loop. Thereafter, the recognition output unit 15 does not unlock the electromagnetic lock 30.

[0072] In S36, when i = 5, the collation unit 14 exits the collation loop and proceeds to S37, or the collation unit 14 increments i = i + 1 and moves to the next collation loop.

[0073] In S37, since the partial captured images corresponding to all the partial registration images substantially match, the recognition output unit 15 recognizes that the target person is the same person as the registrant. Therefore, the recognition output unit 15 unlocks the electromagnetic lock 30.

[0074] According to the above processing, when it is determined that one partial captured image does not substantially match the partial registered image, it is possible to quickly recognize that the subject is not the same person as the registrant. In addition, when the subject is the registrant, in order to suppress the misjudgment of determining that the subject is not the registrant, it is preferable to set a relatively loose similarity threshold for determining that the partial captured image and the partial registered image match each other. Although the relatively loose setting of the threshold may result in a determination that even a subject who is not the registrant substantially matches in the comparison of each part, by using a large number of parts in the comparison loop, it is possible to output that the subject who is not the registrant is a different person from the registrant as a result.

[0075] Here, in S34, conditional branching is performed depending on whether or not one partial captured image substantially matches the partial registered image. When the condition is not satisfied (No in S34), the recognition process ends, but it is not necessarily required to end with the recognition result based on one partial captured image. That is, when a predetermined number of partial captured images do not match the partial registered images, it may be recognized that the subject is not the same person as the registrant. By increasing the number of necessary partial registered images, misjudgment can be suppressed. Here, the number of partial registered images required to determine that the subject is not the same person is preferably less than the number of partial registered images stored in the storage unit 12. Thereby, compared with the case of comparing with all partial registered images, the comparison result can be output at high speed.

[0076] As an actual example of Operation Example 2, it is preferably used basically when preventing the abuse by others of a device used only by the registrant. For example, it is used for improving the security for unlocking by face authentication when an electronic device such as a smartphone owned by an individual is lost.

[0077] (Operation Example 3: Prioritizing the personal recognition speed in multiple registrants) FIG. 6 is a flowchart showing the operation flow of operation example 3 according to Embodiment 1. Operation example 3 is an operation that prioritizes the person recognition speed for the person of the partial registration image when a plurality of face images of people are registered as the partial registration images for face recognition. For simplicity of explanation, it is assumed that there are three registered persons and there are five partial registration images for each registered person.

[0078] In S41, the camera 20 captures a face image, and the image acquisition unit 11 acquires the face image as a captured image.

[0079] In S42, the collation unit 14 performs a collation loop for collating partial registration images for each part from i = 1 to i = 5. Here, i is a natural number indicating the priority of the partial registration image.

[0080] In S43, the collation unit 14 performs a registrant loop for collating each registrant from j = 1 to j = 3. Here, j is a natural number representing the registrant.

[0081] In S44, the image extraction unit 13 extracts a partial captured image corresponding to the partial registration image of the i-th priority for the j-th registrant.

[0082] In S45, the collation unit 14 compares and collates the partial captured image with the partial registration image of the i-th priority for the j-th registrant. When the partial captured image and the partial registration image substantially match, the process proceeds to S46. When the partial captured image and the partial registration image do not substantially match, the process proceeds to S47.

[0083] In S46, since the partial captured image and the partial registration image substantially match, the collation unit 14 recognizes that the target person is the same person as one of the registrants and exits the collation loop. Thereafter, the recognition output unit 15 unlocks the electromagnetic lock 30.

[0084] In S47, when j = 3, the collation unit 14 exits the registrant loop, or increments j = j + 1 and moves to the next registrant loop.

[0085] In S48, when i = 5, the collation unit 14 exits the collation loop or increments i = i + 1 and moves to the next collation loop.

[0086] In S49, since the target person does not substantially match all of the multiple registered persons, the recognition output unit 15 recognizes that the target person is a person different from the multiple registered persons. Therefore, the recognition output unit 15 does not unlock the electromagnetic lock 30.

[0087] According to the above processing, when it is determined that one partial captured image substantially matches any one of the partial registered images of multiple registered persons, it is recognized that the target person is the same person as one of the registered persons, and face recognition is terminated. Therefore, face recognition of the target person can be performed at an early stage. Also, in order to ensure security, it is preferable to set a strict similarity threshold for determining that the partial captured image and the partial registered image substantially match each other. By setting the threshold strictly, it is possible to suppress the possibility that a target person who is not one of the multiple registered persons is recognized as the same person as any one of the multiple registered persons by face recognition.

[0088] As an actual example of operation example 3, it is a device in which the possibility of collation by a target person other than the registered person is low, and it is preferably used when performing face recognition on a device used by multiple registered persons. For example, it can be considered for use in applications such as an entrance gate where basically only authorized personnel are expected to enter, and entry is permitted if the person is a registered person.

[0089] (Operation Example 4: Prioritize recognition speed of others in multiple registered persons) FIG. 7 is a flowchart showing the operation flow of operation example 4 according to Embodiment 1. Operation example 4 is an operation that prioritizes the recognition speed of others excluding the person of the partial registered image for which face recognition is to be performed when face images of multiple persons are registered as the partial registered image. For simplicity of explanation, it is assumed that there are three registered persons, and there are five partial registered images for each registered person.

[0090] The processes of S41 to S44 are the same as those in Operation Example 3, and thus the description thereof is omitted.

[0091] In S55, the collation unit 14 compares and collates the partial captured image with the i-th priority partial registered image of the j-th registered person. If the partial captured image and the partial registered image do not substantially match, the process proceeds to S56. If the partial captured image and the partial registered image substantially match, the process proceeds to S57.

[0092] In S56, the collation unit 14 determines that the target person is not the j-th registered person. Therefore, in the subsequent processes, the j-th registered person is removed from the candidates, and the subsequent collation is not performed.

[0093] In S57, when j = 3, the collation unit 14 exits the registered person loop and proceeds to S58, or increments j = j + 1 and moves to the next registered person loop.

[0094] In S58, the collation unit 14 determines whether there are remaining candidate registered persons among the multiple registered persons. If there are no remaining candidate registered persons, the process proceeds to S59. If there are remaining candidate registered persons, the process proceeds to S60.

[0095] In S59, since there is no partial registered image that substantially matches the partial captured image, the collation unit 14 recognizes that the target person is someone other than any one of the multiple registered persons and exits the collation loop. Thereafter, the recognition output unit 15 does not unlock the electromagnetic lock 30.

[0096] In S60, when i = 5, the collation unit 14 exits the collation loop and proceeds to S61, or increments i = i + 1 and moves to the next collation loop.

[0097] In S61, since there is a registered person whose partial captured image corresponding to all partial registered images substantially matches, the recognition output unit 15 recognizes that the target person is the same person as any one of the multiple registered persons. Therefore, the recognition output unit 15 unlocks the electromagnetic lock 30.

[0098] According to the above processing, when it is determined that a single partial captured image does not substantially match any one of the partial registered images of multiple registered persons, it can be recognized that the subject is different from the registered person, so the subject can be recognized as someone else at an early stage. However, in order not to misidentify a registered person as someone else, it is preferable to set a relatively loose similarity threshold for determining that the partial captured image and the partial registered image substantially match each other. Because the threshold setting is loose, a subject who is not a registered person may be determined to substantially match in the per-site collation, but by using a large number of sites in the collation loop, it is possible to output that a subject who is not a registered person is a different person from the registered person as a result.

[0099] Also, when the partial captured image and the partial registered image are collated and it is determined that they do not substantially match, the process is to remove the registered person from the candidates of multiple registered persons. Therefore, as the collation progresses, the number of candidates for multiple registered persons will be narrowed down. As a result, even when the subject is the same person as one of multiple registered persons and the electromagnetic lock is to be unlocked, it is possible to unlock it earlier than in the case of collating one by one.

[0100] As an actual example of Operation Example 4, it is a device in which a subject other than the registered person is highly likely to be collated, and it is preferably used when face recognition is performed on a device used by multiple registered persons. For example, it can be considered for use in applications such as allowing only registered persons to enter at an entrance gate where a large number of unspecified people are expected to enter. Since it is determined to be the same person by substantially matching in all parts, it is used as an electromagnetic lock for an entrance / exit gate with high security.

[0101] §4. Function and Effect The collation unit can determine whether the partial registered image and the partial captured image corresponding to the partial registered image substantially match based on the calculated similarity.

[0102] When there is only one registered person, if the recognition speed of the person is prioritized, the target person is recognized as the same person as the registered person when a combination in which at least one partial registration image and a partial captured image substantially match is obtained. On the other hand, if the recognition speed of other people is prioritized, the target person is recognized as not the same person as the registered person when a combination in which at least one partial registration image and a partial captured image substantially match is obtained.

[0103] Furthermore, when there are multiple registrants, if the recognition speed of the individual is prioritized, the target person is recognized as the same person as the registrant when a combination is obtained in which at least one partial registration image of a certain body part substantially matches a partial captured image. On the other hand, if the recognition speed of other people is prioritized, if a combination is obtained in which at least one partial registration image of a certain body part substantially does not match a partial captured image, the target person is determined not to be the same person as the registrant, and in the subsequent processing, the registrant is removed from the candidates for face recognition.

[0104] As described above, the results of face recognition can be influenced by image recognition of a certain part, which makes it possible to terminate face recognition early.

[0105] [Embodiment 2] In the first embodiment, a method for increasing the speed of face recognition has been described. In contrast, in the second embodiment, a method for optimizing the priority of partial registration images in order to increase the speed of face recognition will be described.

[0106] 8 is a block diagram showing the configuration of the main parts of a face recognition system 100a according to embodiment 2. The face recognition system 100a differs from the face recognition system 100 in that a priority setting unit 16 is provided in the face recognition device 10a.

[0107] The priority setting unit 16 is a functional block that optimizes the priority for each partial registration image based on the history of the similarity of face recognition results performed in the past. The priority setting unit 16 first learns the tendency of similarity for each partial registration image based on the history of similarity. As an example, the priority setting unit 16 calculates the average similarity based on the history of similarity. When prioritizing the recognition speed of the person himself (operation examples 1 and 3), the priority setting unit 16 assigns priorities to each partial registration image in the order in which the average similarity decreases. As a result, the partial registration image with a high priority is likely to have a high similarity, and face recognition can be completed early. Similarly, when prioritizing the recognition speed of others (operation examples 2 and 4), the priority setting unit 16 assigns priorities in the order in which the average similarity decreases. As a result, the partial registration image with a high priority is likely to have a low similarity to others, and face recognition can be completed early.

[0108] Also, the optimization process of the priority may be performed at any time while the face recognition device 10a is not performing face recognition processing, so that face recognition can always be performed with the optimal priority.

[0109] Furthermore, the optimization of the priority in the priority setting unit 16 does not have to be performed automatically, and may be manually set in an arbitrary priority order by a user operation. In this case, the priority setting unit 16 may display a list and an image of the partial registration image on a display device provided in the face recognition device 10a, and receive a priority setting input from the user for each partial registration image by an input device provided in the face recognition device 10a. Also, the priority setting unit 16 may acquire and set the priority setting information received from the user in an external terminal device via communication means.

[0110] As a specific example of priority assignment, for example, since the eyes are a part where a person's features are strongly shown, the priority of the eyes is set to be high, and since the mouth area is often masked, the priority of the mouth area is set to be low. Also, it becomes possible to set the priority according to one's own features and preferences. For example, it is possible to set the priority of the eyes to be low because the user often wears sunglasses, or to set the priority of the hair to be low because the user often changes their hairstyle.

[0111] 〔Embodiment 3〕 In Embodiment 1, the description focused on face recognition, but the object to be recognized is not limited to the face. In Embodiment 3, object recognition with the object to be recognized being an object will be described.

[0112] FIG. 9 is a model diagram showing the captured image 113 and the partial captured images 114 according to Embodiment 3. The captured image 113 is an image in which an arbitrary object is captured, and may be, for example, an image of a car as shown in FIG. 9. The partial captured images 114 are, for example, images of a part of four captured images 113, and are partial captured images 114a to 114d. Each partial captured image represents a feature of the object.

[0113] For example, the partial captured image 114a is an image of the side glass, the partial captured image 114b is an image of the front glass, the partial captured image 114c is an image of the headlight, and the partial captured image 114d is an image of the tire. The ranges of the partial captured images 114 (collectively referred to as the partial captured images 114a to 114d and denoted as the partial captured image 114) may overlap each other. The method of creating these partial captured images 114 and the method of assigning priorities to the partial captured images 114 are the same as the flowchart of the operations in face recognition shown in FIG. 3.

[0114] Also, the flowchart of the operations in object recognition according to Embodiment 3 is the same as the flowchart of the operations in face recognition shown in FIGS. 4 to 7. By the same method as the method described above, the target object is imaged to obtain a captured image, the partial captured image corresponding to the partial registered image in the registered object is extracted from the captured image, and the partial registered image and the partial captured image are collated.

[0115] When priority is given to the speed at which the object is recognized as a registered object based on the matching result, the target object is recognized as the same object as the registered object when the matching result in at least one part substantially matches, as shown in the above-mentioned Operation Example 1. On the other hand, when priority is given to the speed at which the object is recognized as not being a registered object, the target object is recognized as not being the same object as the registered object when the matching result in at least one part substantially does not match, as shown in the above-mentioned Operation Example 2.

[0116] Next, a case where there are multiple registered objects will be described. When priority is given to the speed at which the target object is recognized as one of the registered objects, as shown in the above-mentioned Operational Example 3, the target object is recognized as being the same as one of the multiple registered objects when the target object substantially matches one of the multiple registered objects in the matching of parts by priority. When priority is given to the speed at which the target object is recognized as being none of the registered objects, as shown in the above-mentioned Operational Example 4, the target object is recognized as being the same as all of the multiple registered objects when the target object substantially does not match one of the multiple registered objects in the matching process of parts by priority.

[0117] [Software implementation example] The functions of the face recognition device 10 / 10a and the object recognition device (hereinafter referred to as the "device") can be realized by a program that causes a computer to function as the device, and a program that causes a computer to function as each control block of the device (especially each part included in the face recognition device 10 / 10a).

[0118] In this case, the device includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the program. The control device and storage device execute the program, thereby realizing the functions described in each of the above embodiments.

[0119] The above program may be recorded on one or more computer-readable recording media, rather than temporarily. This recording medium may or may not be provided in the above device. In the latter case, the above program may be supplied to the above device via any wired or wireless transmission medium.

[0120] In addition, part or all of the functions of each of the above control blocks can also be realized by a logic circuit. For example, an integrated circuit in which a logic circuit functioning as each of the above control blocks is formed is also included in the scope of the present invention. In addition to this, for example, it is also possible to realize the functions of each of the above control blocks by a quantum computer.

[0121] Also, each process described in each of the above embodiments may be executed by AI (Artificial Intelligence). In this case, the AI may operate in the above control device, or may operate in another device (for example, an edge computer or a cloud server, etc.).

[0122] 〔Supplementary Notes〕 The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope shown in the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention.

Explanation of Reference Numerals

[0123] 10, 10a Face recognition device 12 Storage unit 13 Image extraction unit 14 Matching unit 15 Recognition output unit 16 Priority setting unit 20 Camera 30 Electromagnetic lock 100, 100a Face recognition system 111, 113 Captured image 112, 112a~112e, 114, 114a~114d Partial captured image

Claims

1. An image acquisition unit that acquires a captured image of a target object; An image extraction unit that creates a plurality of partial captured images by extracting images of a plurality of predetermined parts of the target object from the captured image; A collation unit that collates a plurality of partial registration images obtained by extracting images of a plurality of predetermined parts of a registered object as a recognition target stored in a storage unit with the partial captured images in an order according to priorities respectively associated with the partial registration images; A recognition output unit that outputs a recognition result of the target object when the collation in the collation unit satisfies a predetermined condition, comprising: The collation unit performs collation in the order according to the priorities, and when it is determined that the partial registration image and the partial captured image corresponding to the partial registration image substantially match, the recognition output unit outputs the recognition result of the target object; The storage unit stores the partial registration images and the priorities of the partial registration images for a plurality of the registered objects; The collation unit performs collation on the partial registration images of all the registered objects at a certain priority, and then performs collation on the partial registration images of all the registered objects at the next priority, an object recognition device.

2. An image acquisition unit that acquires a captured image of a target object; An image extraction unit that creates a plurality of partial captured images by extracting images of a plurality of predetermined parts of the target object from the captured image; A collation unit that collates a plurality of partial registration images obtained by extracting images of a plurality of predetermined parts of a registered object as a recognition target stored in a storage unit with the partial captured images in an order according to priorities respectively associated with the partial registration images; A recognition output unit that outputs a recognition result of the target object when the collation in the collation unit satisfies a predetermined condition, comprising: The collation unit performs collation in the order according to the priorities, and when it is determined that the partial registration image and the partial captured image corresponding to the partial registration image do not substantially match, the recognition output unit outputs the recognition result of the target object; The storage unit stores the partial registration images and the priorities of the partial registration images for a plurality of the target objects; After performing collation on the partial registered images of all the registered objects at a certain priority level, the collation unit performs collation on the partial registered images of all the registered objects at the next priority level. When it is determined that the partial registered image and the partial captured image corresponding to the partial registered image do not substantially match, the collation regarding the corresponding registered object at the next priority level is not performed. An object recognition device.

3. The object recognition device according to claim 1 or 2, wherein the collation unit calculates a similarity between the partial registered image and the partial captured image and compares the similarity with a predetermined threshold value to perform collation.

4. The collation unit performs a process of storing the calculation result of the similarity as a similarity history, The object recognition device according to claim 3, further comprising a priority setting unit that sets a priority of the partial registered image based on the similarity history.

5. The object recognition device according to any one of claims 1 to 4, further comprising a priority setting unit that sets a priority of the partial registered image based on an input of a user.

6. The object recognition device according to any one of claims 1 to 5, wherein the target object is a human face.

7. An image acquisition step of acquiring a captured image of a target object, An image extraction step of creating a plurality of partial captured images by extracting images of a plurality of predetermined parts of the target object from the captured image, A collation step of collating a plurality of partial registered images obtained by extracting images of a plurality of predetermined parts of a registered object as a recognition target stored in a storage unit with the partial captured images in an order according to priorities respectively associated with the partial registered images, A recognition output step of outputting a recognition result of the target object when the collation in the collation step satisfies a predetermined condition, including: In the collation step, collation is performed in the order according to the priority, and when it is determined that the partial registered image and the partial captured image corresponding to the partial registered image substantially match, the recognition result of the target object is output in the recognition output step. The storage unit stores the partial registered image and the priority of the partial registered image for a plurality of the registered objects. A control method for an object recognition device, wherein in the collation step, after performing collation on the partial registered images of all the registered objects at a certain priority level, collation is performed on the partial registered images of all the registered objects at the next priority level.

8. An image acquisition step of acquiring a captured image of a target object; An image extraction step of creating a plurality of partial captured images by extracting images of a plurality of predetermined parts of the target object from the captured image; A matching step of performing matching between a plurality of partial registered images obtained by extracting images of a plurality of predetermined parts of a registered object as a recognition target stored in a storage unit and the partial captured images in an order according to the priorities respectively associated with the partial registered images; A recognition output step of outputting a recognition result of the target object when the matching in the matching step satisfies a predetermined condition, and including: In the matching step, matching is performed in the order according to the priorities, and when it is determined that the partial registered image and the partial captured image corresponding to the partial registered image do not substantially match, the recognition result of the target object is output in the recognition output step; The storage unit stores the partial registered images and the priorities of the partial registered images for a plurality of the target objects; In the matching step, after performing matching for the partial registered images of all the registered objects at a certain priority, matching for the partial registered images of all the registered objects at the next priority is performed. When it is determined that the partial registered image and the partial captured image corresponding to the partial registered image do not substantially match, matching for the corresponding registered object at the next priority is not performed. A control method for an object recognition device.

9. An object recognition program for causing a computer to function as the object recognition device according to Claim 1, the object recognition program for causing a computer to function as the image acquisition unit, the image extraction unit, the storage unit, the matching unit, and the recognition output unit.

10. An object recognition program for causing a computer to function as the object recognition device according to Claim 2, the object recognition program for causing a computer to function as the image acquisition unit, the image extraction unit, the storage unit, the matching unit, and the recognition output unit.

Images