JP2015041307A - Collation device and collation method and collation system and computer program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a collation device capable of more accurately detecting even rough impersonation causing the deterioration of "a likelihood of being the person", and collating an object to be collated.SOLUTION: A collation device 1 includes: a detection part 2 for detecting whether or not a predetermined operation is included in image information 101 on the basis of image information 101 including information configuring an image indicating an object; and a collation part 3 for collating the object on the basis of the image information 101 used when the detection is performed by the detection part 2 and a predetermined stored featured value, and for determining whether or not the object is a specific object on the basis of the collation result.

Description

  The present invention relates to a technical field in which an object is collated based on an image or the like obtained by capturing an object to be collated.

  A system that matches a person using a person's face or a part of the face is a so-called “spoofing” act of presenting a photograph of a face of another person who is not the person or a face photograph displayed by an image display device to the camera. There is a case where the person is collated as another person.

  As a method of detecting such “spoofing”, it is possible to detect whether or not the target of verification presented to the system (hereinafter referred to as “target for verification”) is a living body, Is generally known. Therefore, such a method is often called “liveness check” or “biological detection”.

  As a method for detecting “spoofing”, not only the above-described detection method but also various techniques have been used for a long time.

  For example, the detection method estimates the size of a facial part such as the face, mouth, and eyes based on the distance between the collation target such as the face and the camera. Further, as a result of estimating the size, the detection method determines whether or not the verification target object is appropriate as a human body.

  In addition, as an example, the detection method measures the temperature of the verification target object. The detection method determines whether or not the measured temperature is the temperature of the human body surface. The detection method determines whether or not the verification target is a three-dimensional object.

  As described above, the detection method is a technology for detecting that the verification target is a living body by determining whether or not the verification target is a human body based on the size or temperature of the verification target. It has been known.

  On the other hand, according to such a detection method, by detecting a “predetermined operation” by a person who is a target of collation (hereinafter referred to as “collation target”), the person who is the collation target is detected as a living body. Technology exists. That is, the detection method determines whether or not the person to be collated has a human-like movement.

  The technology for detecting the “predetermined motion” detects the “predetermined motion” by performing image processing on a human face image captured by a camera or the like, a part (part) of the face, and the periphery of the face. .

  In detecting the “predetermined operation”, the only sensing device used in the technology for detecting the “predetermined operation” is a camera. Therefore, the technology for detecting “predetermined operation” can keep the cost for introduction low. That is, a technique for detecting “predetermined operation” is very useful in industry.

  Here, for example, Patent Documents 1 to 3 are known as typical related technologies existing prior to the present application.

  Patent Document 1 discloses a technique relating to a personal identification device, a personal identification method, and a recording medium on which a personal identification program is recorded.

  The personal identification device disclosed in Patent Document 1 detects a blinking motion of a person to be collated using images captured in time series. The personal identification device determines that it is not “spoofing” when a blink operation is detected.

  The personal identification device collates a person to be collated when it is determined that it is not “spoofing”. The personal identification device has a disadvantage in that it cannot complete the process of collating the person to be collated until the person to be collated performs the blinking operation.

  Patent Document 2 discloses a technique related to a biometric matching device, a biometric matching system, a biometric matching method, and a program.

  In contrast to Patent Literature 1 that detects a blinking motion, Patent Literature 2 discloses a technology that requests a person to be collated to perform a specific motion.

  More specifically, the biometric matching system determines that it is not “spoofing” when a “predetermined action” is detected such that the person to be matched opens his eyes after closing his eyes.

  When it is determined that the biometric matching system is not “spoofing”, matching of a person who is a matching target is started.

  The techniques disclosed in Patent Document 1 and Patent Document 2 collate a person who is a collation target only when it is determined that it is not “spoofing”.

  On the other hand, the face authentication device disclosed in Patent Document 3 collates a person to be collated even when it is determined to be “spoofing”.

  More specifically, in the face authentication device, the person to be collated is registered based on an image obtained by capturing the person to be collated and an image representing the person to be collated registered in advance as a template image. It is determined whether or not the person is a matching target.

  Furthermore, the face authentication apparatus detects whether or not the person to be collated is a living body by detecting a change in a facial part such as an eye or a mouth.

  That is, even when it is detected that the person to be collated is not a living body, the face authentication apparatus can collate the person to be collated. As a result, when it is determined that the person to be collated is a registered person to be collated, the face authentication device sets a higher threshold to be used for identifying the living body. As a result, the face authentication apparatus can reduce the risk of erroneous matching.

Japanese Patent Laid-Open No. 11-339048 JP 2002-251614 A JP 2006-330936 A

  The personal identification device disclosed in Patent Document 1 generates a difference image between frame images based on different frame images taken in time series. The personal identification device determines whether or not the generated difference image includes a facial part representing, for example, the eyes and mouth of a person to be collated.

  For example, such a personal identification device determines that there is a facial part even when a shielding object such as paper is placed on the eyes of a person to be collated.

  The biometric matching device disclosed in Patent Literature 2 detects a partial image representing an eye from an image obtained by capturing a person who is a matching target. Furthermore, the biometric matching device performs matching based on the feature points extracted from the partial image representing the detected eye and the feature points stored in advance.

  Thereby, the biometric matching device can detect that a facial part such as an eye or a mouth has moved. That is, the biometric matching device can clearly distinguish between “a state where the eyes are open” and “a state where the eyes are closed”.

  However, Patent Document 2 does not take into account any technique for determining whether or not a rough camouflage that reduces the “personality” such as presenting a face photograph is described. Therefore, in such a biometric matching device, for example, even a hand-drawn eye pattern may detect that the eye has moved.

  The face authentication device disclosed in Patent Literature 3 generates a difference image between frame images based on a plurality of different frame images captured in time series. The face authentication apparatus estimates a region representing the eyes and mouth based on the difference image.

  The face authentication apparatus counts the number of pixels in each estimated area. Furthermore, the face authentication device determines that there has been movement (predetermined movement) in the eyes, for example, when the counted value exceeds a preset threshold value.

  The face authentication apparatus detects that the subject is a living body when it is determined that only one of the eyes and the mouth has moved. In addition, the face authentication apparatus collates a person to be collated by performing template matching on a template image registered in advance using the captured image.

  However, for example, when the face authentication device covers a mouth of a person to be collated with a shield such as paper, the face authentication device determines that only one of the eyes or the mouth has moved. That is, the face authentication device may still erroneously detect that it is a living body.

  As described above, the techniques disclosed in Patent Document 1 to Patent Document 3 are included in a face photograph when, for example, a rough camouflage such as presenting a face photograph that reduces the “personality” is missed. Match people. In other words, in such a technique, there is a risk of erroneous collation when a rough camouflage that reduces the “personality” is missed.

  For example, if the user already knows to detect a “predetermined action” such as opening his eyes after closing his eyes, avoiding the detection of such impersonation methods (ie, “spoofing”) Method) is easy to devise.

  In particular, a camouflage method using a facial photograph obtained by capturing a face of a person different from the person being “spoofed” can be easily devised and easily implemented. That is, a rough camouflage that reduces the “personality” is easily performed by a malicious third party.

  Further, in addition to the above-described camouflaging method, for example, as the camouflaging method, an image representing a third party who performs a “predetermined operation” is used. The camouflage method uses an image generated by combining a plurality of images so as to look like a “predetermined operation”. Alternatively, the camouflaging method includes a method of covering a face photograph obtained by capturing a person to be collated with a shield.

  More specifically, as an example, here, a case will be described in which a person to be collated detects “an eye closing operation” as a predetermined operation.

  For example, a third person who is different from the person to be collated (that is, the person to be impersonated) can easily obtain a facial photograph representing a state in which the person to be collated is open. Furthermore, it is easy for such a third party to obtain a facial photograph showing a state in which an unspecified person (a person different from the person to be collated) closed his eyes.

  First, the third party presents a facial photograph representing a state in which the person to be collated has opened eyes, for example, to a camera that captures the person to be collated. The third person replaces the facial photograph representing the state in which the person to be collated has eyes open with the facial photograph representing the state in which an unspecified person has closed eyes.

  Furthermore, the third party replaces a facial photograph representing a state in which an unspecified person has closed his eyes with a facial photograph representing a state in which the person to be collated has his eyes opened. That is, the third person presents a new face photograph representing the state of opening his eyes.

  As a result, the techniques disclosed in Patent Documents 1 to 3 detect a change in a facial photograph as a “predetermined operation (that is, an operation for closing eyes)”. That is, this technique erroneously detects that it is a living body. Further, the technology verifies that the person to be collated is a person to be collated based on a facial photograph representing a state in which the person to be collated is newly opened.

  The reason why such a problem occurs is that such technology collates a person to be collated after determining that the subject is a living body. For this reason, if it is erroneously detected that the subject is a living body by impersonation, such technology will erroneously collate as a person to be collated.

  Even if it is a case where it is detected whether the person who is a collation object is a living body after collating the person who is a collation object, the technique cannot solve the same problem.

  The main object of the present invention is to perform collation capable of more accurately detecting the impersonation and collating the target object to be collated even if it is a rough camouflage that reduces the “personality” The main purpose is to provide devices and the like.

  In order to achieve the above object, a collation apparatus according to the present invention has the following configuration.

That is, the verification device according to the present invention is
When it is detected that the image information includes the first action based on image information including information constituting an image representing the object to be collated, the object is detected as a living body, and the image When it is detected that the information does not include the first operation, a detection unit that detects that the information is not in the first operation;
When it is detected that it is the living body, the image information used when it is detected that the first action is included, and the image representing the first action by a specific object as a collation target held in the storage unit or A state in which the first operation is not performed by checking whether or not the object is the specific object based on a feature amount extracted based on any one of the images indicating that the first operation is not performed. If it is detected, the image information used when it is detected that the first operation is not included is compared with the feature amount based on the image information, and it is detected as the living body. A collation unit that determines whether or not the object is the specific object based on a result collated and a result collated when it is detected that the first operation is not performed When,
It is characterized by providing.

  Alternatively, this object can be achieved by a verification system including the verification device described above.

  In order to achieve the same object, the collation method according to the present invention is characterized by having the following configuration.

That is, the verification method according to the present invention is:
Depending on the information processing device,
When it is detected that the image information includes the first action based on image information including information constituting an image representing the object to be collated, the object is detected as a living body, and the image When it is detected that the information does not include the first action, it is detected that the first action is not performed,
When it is detected that it is the living body, the image information used when it is detected that the first action is included, and the image representing the first action by a specific object as a collation target held in the storage unit or A state in which the first operation is not performed by checking whether or not the object is the specific object based on a feature amount extracted based on any one of the images indicating that the first operation is not performed. If it is detected, the image information used when it is detected that the first operation is not included is compared with the feature amount based on the image information, and it is detected as the living body. Determining whether or not the object is the specific object based on the result of collation and the result of collation when detecting that the first operation is not performed,
It is characterized by that.

  This object is also achieved by a computer program that implements the collation apparatus and collation method having the above-described configurations by a computer, and a readable storage medium that stores the computer program.

  ADVANTAGE OF THE INVENTION According to this invention, even if it is a rough camouflage with which "personality" falls, the collation apparatus which can detect the camouflage more correctly and can collate the target object which is the object of collation Etc. can be provided.

It is a block diagram which shows the structure of the collation apparatus in the 1st Embodiment of this invention. It is a flowchart which shows the operation | movement which the collation apparatus in the 1st Embodiment of this invention performs. It is a block diagram which shows the structure of the collation apparatus in the 2nd Embodiment of this invention. It is a flowchart which shows the operation | movement which the collation apparatus in the 2nd Embodiment of this invention performs. It is a block diagram which shows the structure of the collation apparatus in the 3rd Embodiment of this invention. It is a flowchart which shows the operation | movement which the collation apparatus in the 3rd Embodiment of this invention performs. It is a block diagram which shows the structure of the collation system containing the collation apparatus in the 4th Embodiment of this invention. It is a flowchart which shows the operation | movement which the collation apparatus in the 4th Embodiment of this invention performs. It is a block diagram explaining illustratively the hardware constitutions of the information processor which can realize each embodiment concerning the present invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

<First Embodiment>
FIG. 1 is a block diagram showing a configuration of a collation device 1 according to the first embodiment of the present invention.

  In FIG. 1, the verification device 1 includes a detection unit 2 and a verification unit 3.

  More specifically, the detection unit 2 performs a predetermined operation (hereinafter, for convenience of explanation) on the image information 101 based on the image information 101 including an image (information constituting the image) representing an object to be collated. If it is detected that the image includes the “first operation”, the object included in the image represented by the image information 101 is detected as a living body.

  In addition, the detection unit 2 detects that the first operation is not being performed. That is, the detection unit 2 detects that it is in a state where it is easy to collate the target object to be collated (natural).

  Here, the state that is easy to collate (natural) includes, for example, a state that represents that the person has an open face and a mouthless expression when collating the person. However, the present invention described using this embodiment as an example is not limited to the configuration described above (hereinafter, the same applies to each embodiment).

  More specifically, when the detection unit 2 detects that the first operation is not included in the image information 101 based on the image information 101, the detection unit 2 detects that the first operation is not performed.

  The image information 101 is, for example, a moving image composed of a plurality of time-series image frames including an image representing a target object to be collated captured by a photographing device such as a camera.

  More specifically, the image information 101 includes, for example, a partial image representing a human body face to be collated or a part (part) of a face such as an eye or a mouth when collating a person.

  In the above-described embodiment, for convenience of explanation, the image information 101 is described as an example of a moving image composed of a plurality of time-series image frames. However, the embodiment according to the present invention is not limited to such a configuration.

  The image information 101 may employ an image composed of a plurality of time-series image frames. Further, for example, the image information 101 may employ a feature amount extracted from an image representing an object that is a collation target.

  The first operation is an operation of an object or a part of the object detected by the detection unit 2 based on the image information 101.

  More specifically, the first operation is, for example, an operation of a person's face or face part to be collated when the person is collated.

  As the first operation, for example, a person's face can be moved up and down, left and right, a face can be tilted left and right, eye (eyelid) opening and closing, eyeball (iris) movement, and mouth opening and closing operation can be adopted.

  The first operation may be any of the operations described above or a combination of one or more operations.

  More specifically, as an example, when an eye opening / closing operation is adopted as the first operation, the first operation represents an “eye closing” operation. When the detection unit 2 detects that the image information 101 includes a “close eye” operation, the detection unit 2 detects that the object included in the image represented by the image information 101 is a living body.

  Further, when the detection unit 2 detects that the image information 101 does not include the “close eye” operation, the detection unit 2 detects that the first operation is not performed. For example, when the detection unit 2 detects an “open eye” operation in the image information 101, the detection unit 2 may detect that the image information 101 does not include an “close eye” operation. For example, when the detection unit 2 detects the state of “open eyes” in the image information 101, the detection unit 2 may detect that the operation of “closing eyes” is not included in the image information 101.

  As an example, when the opening / closing operation of the mouth is adopted as the first operation, the first operation represents an operation of “opening the mouth”. When the detection unit 2 detects that the image information 101 includes an “open mouth” operation, the detection unit 2 detects that the object included in the image represented by the image information 101 is a living body.

  Further, when the detection unit 2 detects that the image information 101 does not include the “open mouth” operation, the detection unit 2 detects that the first operation is not performed. For example, the detection unit 2 may detect that the image information 101 does not include the “open mouth” operation when the image information 101 detects the “close mouth” operation. For example, when the detection unit 2 detects the state of “closing the mouth” in the image information 101, the detection unit 2 may detect that the operation of opening the mouth is not included in the image information 101.

  As a result, when the detection unit 2 identifies that “eyes are closed” based on the image information 101, the detection unit 2 can detect that the image information 101 includes the first action.

  For example, when the detection unit 2 identifies “open eyes” based on the image information 101, the detection unit 2 can detect that the first operation is not included in the image information 101. That is, the detection unit 2 can determine that the first operation has ended by detecting that the first operation is not included in the image information 101.

  Next, when the detection unit 2 detects that the object is a living body, the matching unit 3 holds the image information 101 used when the image information 101 includes the first operation and the storage unit 4. Based on the feature quantity 102 thus determined, it is verified whether or not the target object included in the image represented by the image information 101 is a specific target object to be verified.

  Further, when the detection unit 2 detects that the target object is not in the first motion, the image information 101 used when the verification unit 3 detects that the target is not in the first motion. Then, based on the feature quantity 102, it is verified whether or not the target object included in the image represented by the image information 101 is a specific target object to be verified.

  Here, the feature quantity 102 is data used when the collation unit 3 collates the target object to be collated. That is, when collating a person with an image, the feature quantity 102 is a numeric string (vector) converted so that it is easy to distinguish a person to be collated.

  More specifically, the feature quantity 102 is an image representing that the first action is not performed by a specific target object to be collated that is stored in the storage unit 4 in advance (that is, an action and a state different from the first action). It is the feature-value extracted based on.

  In the above-described embodiment, for convenience of explanation, as an example, the feature amount 102 is a feature amount extracted based on an image indicating that the first operation by the specific target object to be collated is not performed. This is explained in the example of adoption.

  However, the embodiment according to the present invention is not limited to such a configuration. As the feature quantity 102, a feature quantity extracted based on an image representing a first action by a specific target object to be collated may be adopted.

  In addition, since a technique for extracting a feature amount using an arbitrary statistical pattern recognition technique based on an image representing an object can be a general technique at present, a detailed description in the present embodiment is provided. Is omitted (hereinafter, the same applies to each embodiment).

  The collation unit 3 is a state in which the detection unit 2 detects that the target object is a living body (hereinafter referred to as “first collation result” for convenience of explanation) and the first operation is not performed. The object included in the image represented by the image information 101 is stored in advance as a collation target based on the collation result (hereinafter referred to as “second collation result” for convenience of explanation). It is determined whether or not the object is a specific object held in the unit 4.

  The storage unit 4 holds the feature amount 102. The storage unit 4 is a nonvolatile storage device that can read and write data by a computer.

  More specifically, as an example, the storage unit 4 may employ a nonvolatile storage device such as a hard disk drive (Hard_disk_drive: hereinafter referred to as “HDD”) mounted on an electronic device such as an information processing device. .

  As an example, the storage unit 4 may employ a storage device (not shown) connected to a communication network (not shown). However, the present invention described using this embodiment as an example is not limited to the above-described configuration (the same applies to the following embodiments).

  In the following description, the operation of the verification device 1 in the present embodiment will be described more specifically.

  FIG. 2 is a flowchart showing an operation performed by the collation device 1 according to the first embodiment of the present invention. The operation procedure of the collation apparatus 1 will be described along the flowchart.

  Here, for convenience of explanation, as an example, the collation apparatus 1 detects the opening / closing operation of the eyes of a person to be collated. The first operation represents an “eyes closing” operation. Further, the image information 101 includes a partial image representing a person to be collated.

  For convenience of explanation, the above-described configuration will be described as an example. However, the present invention described by taking this embodiment as an example is not limited to the above-described configuration (the same applies to the following embodiments).

Step S101:
Based on the image information 101, the detection unit 2 detects whether or not a person included in the image represented by the image information 101 is a living body. That is, the detection unit 2 detects whether or not the image information 101 includes the first operation based on the image information 101.

  More specifically, the detection unit 2 uses a generally known arbitrary statistical pattern recognition technique to detect that the person included in the image represented by the image information 101 is “first operation” by the generated classifier. It is identified whether or not it represents a state. When the detection unit 2 identifies that the state represents the “first operation”, the detection unit 2 detects that the image information 101 includes the first operation.

  As an arbitrary statistical pattern recognition technique, generally known support vector machines, neural networks, learning vector quantization, boosting, and the like can be adopted.

  For example, when identifying “open / closed eyes” in a person to be detected, the discriminator uses an arbitrary statistical pattern recognition technique to “close the eyes” and “open the eyes” including the periphery of the eyes. It is generated by learning a large number of human body images representing the “state” (the same applies to the following embodiments).

“YES” in step S102:
When the detection unit 2 detects that the person included in the image represented by the image information 101 is a living body, the detection unit 2 advances the processing to step S103 and step S104.

“NO” in step S102:
When the detection unit 2 detects that the person included in the image represented by the image information 101 is not a living body, the detection unit 2 proceeds with processing of the next video frame. That is, the detection unit 2 returns the process to step S101.

Step S103:
The collation unit 3 collates whether or not a person included in the image represented by the image information 101 used when the detection unit 2 detects a living body is a specific person to be collated. The collation unit 3 advances the process to step S107.

  More specifically, the collation unit 3 is based on the feature amount extracted from the image information 101 used when the detection unit 2 detects that it is a living body and the feature amount 102 held in the storage unit 4. It is verified whether or not the person included in the image represented by the image information 101 is a specific person to be verified.

  In addition, since the collation part 3 extracts the feature-value from the image information 101 and collates based on the extracted feature-value and the feature-value 102, since a general technique can be employ | adopted now, this implementation A detailed description of the embodiment will be omitted (hereinafter, the same applies to each embodiment).

Step S104:
Based on the image information 101, the detection unit 2 detects whether or not the image information 101 includes the first operation.

  More specifically, the detection unit 2 identifies whether or not a person included in the image represented by the image information 101 represents a “first operation”. The detection unit 2 detects that the first operation is not included in the image information 101 when it is identified that the state does not represent the “first operation”.

  Note that the image information 101 used when the detection unit 2 detects whether or not the image information 101 includes the first operation in step S104 is the image information 101 used by the detection unit 2 includes the first operation in the image information 101 in step S101. The image information is different from the image information 101 used when detecting whether or not. That is, the image information 101 used in step S101 and the image information 101 used in step S104 are different video frames (the same applies to the following embodiments).

  As a method for detecting whether the detection unit 2 includes the first operation in the image information 101, the method of detecting whether the image information 101 includes the first operation in step S101 described above can be employed. . Therefore, the overlapping description is omitted.

“NO” in step S105:
If the detection unit 2 detects that the first operation is not included in the image information 101, the process proceeds to step S106.

“YES” in step S105:
When the detection unit 2 detects that the image information 101 includes the first operation, the detection unit 2 proceeds with processing of the next video frame. That is, the detection unit 2 returns the process to step S101.

Step S106:
The collation unit 3 collates whether or not the person included in the image represented by the image information 101 used when the detection unit 2 detects that the first operation is not included is a specific person to be collated. The collation unit 3 advances the process to step S107.

  More specifically, the collation unit 3 is based on the feature amount extracted from the image information 101 used when the detection unit 2 detects that the first operation is not included, and the feature amount 102 held in the storage unit 4. Thus, it is verified whether or not the person included in the image represented by the image information 101 is a specific person to be verified.

Step S107:
Based on the result collated in step S103 (first collation result) and the result collated in step S106 (second collation result), the collation unit 3 collates a person included in the image represented by the image information 101. It is determined whether or not the target person is a specific person.

  More specifically, the collation unit 3 performs collation based on the feature amount extracted from the image information 101 used when it is detected as a living body and the feature amount 102 held in the storage unit 4 (first (Matching result), a result of matching based on the feature quantity extracted from the image information 101 used when it is detected that the first operation is not included, and the feature quantity 102 stored in the storage unit 4 (second matching result) ), It is determined whether or not the person included in the image represented by the image information 101 is a specific person to be collated.

  It should be noted that the determination results shown in steps S103, S106, and S107 described above are “true” indicating that the object is a specific object to be verified, or “false” indicating that the object is not a specific object that is a target to be verified. ”May be represented by a binary value. In addition, the determination result may be represented by a numerical value (collation score) that represents the object likeness to be collated.

  When a binary value representing “true” or “false” is adopted as the determination result, the collation unit 3 takes a logical product (AND) of the first collation result and the second collation result, and the result Judge according to.

  More specifically, when the first and second collation results represent “true”, the collation unit 3 determines “true”, that is, a specific object to be collated. On the other hand, if either of the first and second matching results represents “false”, the matching unit 3 determines that the information is not “false”, that is, a specific target object to be verified.

  When a numerical value (collation score) is adopted as the determination result, the collation unit 3 uses the linear sum of the first collation result and the second collation result or the comparison result between the minimum value and a predetermined threshold value. Accordingly, it is determined whether or not the object included in the image represented by the image information 101 is a specific object to be collated.

  More specifically, as an example, the collation unit 3 determines that the target is not a specific object to be collated when the linear sum or the minimum value is smaller than the threshold as a result of comparison. On the other hand, the collation part 3 may determine with it being the specific target object which is collation object, when it is more than a threshold value as a result of the comparison.

  Further, the matching unit 3 scales the value of the matching score between the first matching result and the second matching result as a real number from “0.0” to “1.0”, for example, so that it can be treated as a probability. .

  The matching unit 3 multiplies the scaled first matching result value and the second matching result value. The collation unit 3 determines whether or not the object included in the image represented by the image information 101 is a specific object to be collated, according to a result of comparing the value obtained by multiplication with a predetermined threshold value. Determine whether.

  The collation unit 3 subtracts the value obtained by multiplying the value of the first collation result scaled from the value “1” by the value of the second collation result. The collation unit 3 converts the value obtained by the subtraction using, for example, a softmax function. The collation unit 3 determines whether or not the object included in the image represented by the image information 101 is a specific object to be collated, according to a result of comparing the converted value with a predetermined threshold value. .

  As the predetermined threshold, for example, a configuration set in advance by an administrator may be employed. In addition, a configuration in which the predetermined threshold is set according to the determination result by the matching unit 3 may be adopted.

  In the present embodiment described above, for convenience of explanation, as an example, the collation apparatus 1 detects whether or not the image information 101 includes the first operation in step S101 and then performs the first operation on the image information 101 in step S104. The configuration for detecting whether or not it is included has been described as an example.

  However, the embodiment according to the present invention is not limited to such a configuration.

  More specifically, for example, in the present embodiment, the collation apparatus 1 adopts a configuration in which the processing described in steps S104 to S106 illustrated in FIG. 2 is performed before the processing described in steps S101 to S103. can do. That is, the detection unit 2 executes either one of the processing that detects that the image information 101 includes the first operation or the processing that detects that the image information 101 does not include the first operation, and either of them. A configuration in which one process is executed and then the other process is executed can be employed.

  In this case, in step S104 shown in FIG. 2, the detection unit 2 does not detect that the first operation has ended, but is in a state in which the target object to be verified can be verified (the first operation is not performed). It is possible to adopt a configuration that detects that the state is (the same applies to the following embodiments).

  Moreover, the collation apparatus 1 may employ | adopt the structure which performs a process in order of step S104 shown in FIG. 2, step S103, step S105, and step S106. Furthermore, the collation apparatus 1 may employ a configuration in which the processes described in step S103 and steps S104 to S106 are executed in substantially parallel.

  As described above, according to the collation device 1 according to the present embodiment, even if it is a rough camouflage such that the “personality” is reduced, the impersonation is detected more accurately, and the target to be collated You can collate things. The reason is as described below.

  That is, based on the image information 101, the detection unit 2 detects whether or not the image information 101 includes a predetermined operation (first operation). Thereby, the detection part 2 can detect more accurately whether the target object which is a collation object is a biological body. Further, the collation unit 3 is used when the detection unit 2 detects that the image information 101 includes a predetermined operation and the image information 101 that the image information 101 does not include the predetermined operation. Based on the image information 101, the object included in the image represented by the image information 101 is collated. The collation unit 3 includes the object included in the image represented by the image information 101 based on the collation result when it is detected that the predetermined operation is included and the collation result when it is detected that the predetermined operation is not included. This is because it can be determined whether or not is a specific target object to be collated.

  Thus, the collation apparatus 1 can verify the object (person) identity that is the collation target. Further, the collation device 1 can prevent the problem of collating an impersonated object, that is, erroneous collation even if it is a rough camouflage that reduces the “personality”.

<Second Embodiment>
Next, a second embodiment based on the verification device 1 according to the first embodiment of the present invention described above will be described. In the following description, the characteristic part according to the present embodiment will be mainly described. At this time, the same reference numerals are assigned to the same configurations as those in the above-described embodiments, and duplicate descriptions are omitted.

  A collation device 10 according to a second embodiment of the present invention will be described with reference to FIGS.

  FIG. 3 is a block diagram showing a configuration of the collation apparatus 10 according to the second embodiment of the present invention.

  In FIG. 3, the collation device 10 includes a first motion detection unit 11, a second motion detection unit 12, a first collation unit 13, a second collation unit 14, and a determination unit 15.

  More specifically, the first motion detection unit 11 and the second motion detection unit 12 correspond to the detection unit 2 described in the first embodiment.

  The first motion detection unit 11 executes the processing described in step S101 and step S102 illustrated in FIG.

  The second motion detection unit 12 executes the processing described in step S104 and step S105 illustrated in FIG.

  The 1st collation part 13, the 2nd collation part 14, and the determination part 15 are corresponded to the collation part 3 demonstrated in 1st Embodiment.

  The first verification unit 13 executes the process described in step S103 illustrated in FIG. The second collation unit 14 executes the processing described in step S106 shown in FIG. The determination unit 15 executes the process described in step S107 illustrated in FIG.

  More specifically, the first verification unit 13 includes the feature amount extracted from the image information 101 used when the first motion detection unit 11 detects a living body, and the first amount held in the storage unit 4. Based on the feature amount 103, it is verified whether or not the target object included in the image represented by the image information 101 is a specific target object to be verified. The first collation unit 13 gives the collation result (first collation result) to the determination unit 15.

  The second collation unit 14 extracts the feature amount extracted from the image information 101 used when the second motion detection unit 12 detects that the first motion is not included, and the second feature amount held in the storage unit 4. 104, it is verified whether or not the object included in the image represented by the image information 101 is a specific object to be verified. The second collation unit 14 gives the collation result (second collation result) to the determination unit 15.

  Based on the first collation result and the second collation result, the determination unit 15 selects a specific object in which the object included in the image represented by the image information 101 is stored in advance in the storage unit 5 as a collation target. It is determined whether or not.

  The storage unit 4 holds the first feature amount 103 and the second feature amount 104.

  In the present embodiment, the first feature amount 103 is a feature amount extracted based on an image representing a first action by a specific object that is a collation target. The second feature amount 104 is a feature amount extracted based on an image indicating that the first action is not performed by a specific target object to be collated.

  FIG. 4 is a flowchart showing an operation performed by the collation device 10 according to the second embodiment of the present invention. The operation procedure of the collation device 10 will be described along the flowchart.

  Note that the processing executed in step S201, step S202, step S204, and step S205 shown in FIG. 4 is the same as the processing described in step S101, step S102, step S104, and step S105 shown in FIG. Therefore, the overlapping description is omitted.

Step S203:
The first collation unit 13 includes the feature amount extracted from the image information 101 used when the first motion detection unit 11 detects that it is a living body, and the first feature amount 103 held in the storage unit 4. Based on this, it is verified whether or not the object included in the image represented by the image information 101 is a specific object to be verified.

  The first collation unit 13 gives the collation result (first collation result) to the determination unit 15. The first matching unit 13 advances the process to step S207.

Step S206:
The second collation unit 14 extracts the feature amount extracted from the image information 101 used when the second motion detection unit 12 detects that the first motion is not included, and the second feature amount held in the storage unit 4. 104, it is verified whether or not the object included in the image represented by the image information 101 is a specific object to be verified.

  The second collation unit 14 gives the collation result (second collation result) to the determination unit 15. The second collation unit 14 proceeds with the process to step S207.

Step S207:
The determination unit 15 determines whether or not the object included in the image represented by the image information 101 is a specific object that is a comparison target, based on the first matching result and the second matching result. .

  As described above, according to the collation device 10 according to the present embodiment, it is possible to enjoy the effects described in the first embodiment, and moreover, a rough disguise in which the “personality” is more quickly and accurately reduced. Can be detected, and an object to be verified can be verified.

  The reason is that the collation apparatus 10 executes the processing executed by the detection unit 2 described in the first embodiment in a distributed manner between the first operation detection unit 11 and the second operation detection unit 12. Furthermore, the collation apparatus 10 executes the processing executed by the collation unit 3 described in the first embodiment in a distributed manner to the first collation unit 13, the second collation unit 14, and the determination unit 15. For this reason, it is because the collation apparatus 10 can perform a process, without waiting for each process to complete | finish.

  In addition, the collation device 10 extracts feature amounts (first feature amount 103 and second feature amount 104) extracted based on an image representing the first operation and an image representing that the first operation is not performed, and image information. This is because collation is performed based on the feature amount extracted from 101, so that a specific object that is a collation target can be collated more accurately.

<Third Embodiment>
Next, a third embodiment based on the collation device 10 according to the second embodiment of the present invention described above will be described. In the following description, the characteristic part according to the present embodiment will be mainly described. At this time, the same reference numerals are assigned to the same configurations as those in the above-described embodiments, and duplicate descriptions are omitted.

  A collation device 20 according to a third embodiment of the present invention will be described with reference to FIGS.

  FIG. 5 is a block diagram showing a configuration of the collation device 20 according to the third embodiment of the present invention.

  In FIG. 5, the collation device 20 includes a first motion detection unit 21, a second motion detection unit 12, a first collation unit 13, a second collation unit 14, and a determination unit 15.

  In this embodiment, the collation apparatus 20 differs in the operation | movement of the 1st motion detection part 21 from the operation | movement of the 1st motion detection part 11 demonstrated in 2nd Embodiment.

  The first motion detection unit 11 described in the second embodiment indicates that the object included in the image represented by the image information 101 represents the “first motion” or “not performing the first motion”. The configuration for identifying one of the states to be expressed has been described as an example.

  In the present embodiment, the first motion detector 21 includes a state representing that the object included in the image represented by the image information 101 is “not a target”, a state representing “a first motion by the target”, and a “target A configuration is adopted in which one of the states indicating that the first action by the object is not performed ”is identified.

  Thereby, the first motion detector 21 can detect “spoofing” with higher accuracy.

  More specifically, as an example, when collating a person, the first motion detection unit 21 indicates that the person included in the image represented by the image information 101 is “not a person”, “first action by person”. And a state indicating that “the first action by a person is not performed” is identified.

  Thereby, the first motion detection unit 21 not only detects whether or not the target object included in the image represented by the image information 101 is a living body, but each part of the human body such as an eye, an eyeball, and a mouth is a human body. Or not. In other words, the first motion detection unit 21 can detect the disguise more accurately even when “not eyes” and “disguise as if it were eyes”.

  In the following description, the operation of the first motion detector 21 in the present embodiment will be described more specifically.

  FIG. 6 is a flowchart showing an operation performed by the collation device 20 according to the third embodiment of the present invention. The operation procedure of the collation device 20 will be described along the flowchart.

  Note that the processing executed in steps S302 to S307 shown in FIG. 6 is the same as the processing described in steps S202 to S207 shown in FIG. Therefore, the overlapping description is omitted.

Step S301:
Based on the image information 101, the first motion detection unit 21 determines that the object included in the image information 101 is a living body when it is detected that the image information 101 includes the first motion due to the object. .

  In the present embodiment, the first motion detection unit 21 uses a statistical pattern recognition technique to indicate that the object included in the image represented by the image information 101 is “not an object” by the generated classifier. Any one of the state, the state representing the “first operation by the object”, and the state representing “the first operation by the object is not performed” is identified.

  For example, when identifying “opening / closing of eyes” in a person to be detected, the discriminator determines that “not eyes” and “impersonated as if eyes” by any statistical pattern recognition technique. It is generated by learning a large number of human body images to be represented and a large number of human body images representing “a state where the eyes are closed” and “a state where the eyes are opened” including the periphery of the eyes.

  In the above-described embodiment, for convenience of explanation, as an example, the collation device 20 includes three (“not an object”, “first operation by the object”, and “object” by the first motion detection unit 21. In the above description, the configuration for identifying the state of “not performing the first operation according to the above” is taken as an example.

  However, the embodiment according to the present invention is not limited to such a configuration. The collation device 20 may employ a configuration in which the second motion detection unit 12 also identifies three states (hereinafter, the same applies to each embodiment).

  As described above, according to the collation device 20 according to the present embodiment, it is possible to enjoy the effects described in the embodiments and to detect “spoofing” with higher accuracy.

  The reason is that the first motion detection unit 21 of the verification device 20 can identify three states. That is, the first motion detection unit 21 not only detects the first motion, but also operates to verify whether or not the part that has performed the first motion is really an object. Therefore, for example, even when impersonation is performed using a non-target object, “spoofing” can be detected with higher accuracy. Thereby, the collation apparatus 20 can prevent the problem of collating the impersonated object, that is, erroneous collation.

  More specifically, as an example, even if the eyes are disguised as if the eyes are closed by putting a long piece of paper over the part representing both eyes in the face photograph representing the person who is the subject of “spoofing” The device 20 can accurately detect “spoofing”.

<Fourth Embodiment>
Next, an embodiment based on the collation device 10 according to the second embodiment of the present invention described above will be described. In the following description, the characteristic part according to the present embodiment will be mainly described. At this time, the same reference numerals are assigned to the same configurations as those in the above-described embodiments, and duplicate descriptions are omitted.

  A verification device 30 according to the fourth embodiment of the present invention will be described with reference to FIGS. 4, 7, and 8.

  FIG. 7 is a block diagram showing a configuration of a verification system 50 including a verification device 30 according to the fourth embodiment of the present invention.

  In FIG. 7, the collation system 50 is roughly provided with a first information processing device 51 including a collation device 30 and a second information processing device 52.

  The verification device 30 includes a first motion detection unit 11, a second motion detection unit 12, a first verification request unit 31, a second verification request unit 32, and a determination unit 15.

  In the present embodiment, the first verification unit 33, the second verification unit 34, and the storage unit 4 are provided in a second information processing device 52 that is different (that is, separate) from the first information processing device 51. .

  The first information processing apparatus 51 and the second information processing apparatus 52 are communicably connected via a communication network (not shown).

  More specifically, the first collation request unit 31 receives the second information processing apparatus in response to the first motion detection unit 11 detecting that the object included in the image represented by the image information 101 is a living body. 52 is requested to collate the object included in the image represented by the image information 101 (hereinafter simply referred to as “collation request”).

  The first verification request unit 31 uses the image information 101 used when the first verification unit 33 detects that the first motion detection unit 11 includes the first motion in the first verification unit 33 when the verification request is made. Send.

  The first verification request unit 31 receives the first verification result from the first verification unit 33 and gives the received first verification result to the determination unit 15.

  The second collation requesting unit 32 makes the second collation unit 34 of the second information processing device 52 respond to the second motion detection unit 12 detecting that the first motion is not being performed. It requests the object included in the image represented by the image information 101 to be collated.

  The second collation request unit 32 uses the image information used when the second motion detection unit 12 detects that the first motion is not included in the image information 101 with respect to the second collation unit 34 when the collation request is made. 101 is transmitted.

  The second collation request unit 32 receives the second collation result from the second collation unit 34 and gives the received second collation result to the determination unit 15.

  In the above-described embodiment, for convenience of explanation, as an example, the first collation request unit 31 and the second collation request unit 32 provide the first collation unit 33 or the second collation request when performing a collation request. The configuration for transmitting the image information 101 to the collation unit 34 has been described as an example.

  However, the embodiment according to the present invention is not limited to such a configuration. The first collation request unit 31 and the second collation request unit 32 are extracted based on the image information 101 with respect to the first collation unit 33 or the second collation unit 34 when requesting collation. A configuration for transmitting the feature amount may be employed.

  In that case, the first collation requesting unit 31 and the second collation requesting unit 32 can employ a configuration that extracts feature amounts based on the image information 101.

  Since the first collation requesting unit 31 and the second collation requesting unit 32 extract the feature amount based on the image information 101, a general technique can be adopted at present. Detailed description in this embodiment will be omitted (hereinafter, the same applies to each embodiment).

  The first collation unit 33 starts processing in response to receiving the collation request and the image information 101 from the first collation request unit 31.

  More specifically, the first matching unit 33 is a target included in an image represented by the image information 101 based on the feature amount extracted from the image information 101 and the first feature amount 103 held in the storage unit 4. Whether or not the object is a specific object to be verified is verified. Furthermore, the first collation unit 33 transmits the first collation result to the first collation request unit 31.

  The second verification unit 34 starts processing in response to receiving the verification request and the image information 101 from the second verification request unit 32.

  Based on the feature quantity extracted from the image information 101 and the second feature quantity 104 stored in the storage unit 4, the second collation unit 34 compares the object included in the image represented by the image information 101 with the collation target. It is checked whether or not the object is a specific object. Further, the second collation unit 34 transmits the second collation result to the second collation request unit 32.

  In the following description, the operation of the verification device 30 in the present embodiment will be described more specifically.

  FIG. 8 is a flowchart showing an operation performed by the collation device 30 according to the fourth embodiment of the present invention. The operation procedure of the collation device 30 will be described along the flowchart.

  Note that the processing executed in step S401, step S402, step S406, step S407, and step S411 shown in FIG. 8 is the same as the processing described in step S201, step S202, step S204, step S205, and step S207 shown in FIG. It is. Therefore, the overlapping description is omitted.

Step S403:
The first collation request unit 31 sends the collation request and the image information 101 used when the first motion detection unit 11 detects that the image information 101 includes the first motion to the first collation unit 33. To send.

Step S404:
The first verification unit 33 receives the verification request from the first verification request unit 31 and the image information 101 used when the first motion detection unit 11 detects that the image information 101 includes the first motion. To do.

  Based on the feature amount extracted from the received image information 101 and the first feature amount 103 held in the storage unit 4, the first collation unit 33 detects the object included in the image represented by the image information 101. It is collated whether or not it is a specific object to be collated. Next, the first collation unit 33 transmits the first collation result to the first collation request unit 31.

Step S405:
The first verification request unit 31 receives the first verification result from the first verification unit 33 and gives the received first verification result to the determination unit 15.

Step S408:
The second collation request unit 32 sends the collation request and the image information 101 used when the second motion detection unit 12 detects that the image information 101 does not include the first motion to the second collation unit 34. Send to.

Step S409:
The second verification unit 34 receives the verification request from the second verification request unit 32 and the image information 101 used when the second motion detection unit 12 detects that the image information 101 does not include the first motion. Receive.

  Based on the feature amount extracted from the received image information 101 and the second feature amount 104 held in the storage unit 4, the second collation unit 34 determines whether the object included in the image represented by the image information 101 is the target object. It is collated whether or not it is a specific object to be collated. Next, the second verification unit 34 transmits the second verification result to the second verification request unit 32.

Step S410:
The second collation request unit 32 receives the second collation result from the second collation unit 34 and gives the received second collation result to the determination unit 15.

  In the above-described embodiment, for convenience of explanation, as an example, the collation system 50 includes the first motion detection unit 11, the second motion detection unit 12, the first collation request unit 31, and the second collation request. The example comprised by the 1st information processing apparatus 51 containing the part 32 and the determination part 15, and the 2nd information processing apparatus 52 containing the 1st collation part 33, the 2nd collation part 34, and the memory | storage part 4 was demonstrated.

  However, the embodiment according to the present invention is not limited to such a configuration. The collation system 50 may be configured by, for example, the first information processing apparatus 51 including all the units configuring the collation apparatus 10 described in the second embodiment.

  As described above, according to the collation device 30 according to the present embodiment, the effect described in each embodiment can be enjoyed, and further, a rough camouflage in which the “personality” is quickly reduced is detected and collation is performed. It is possible to collate objects that are objects of

  The reason is that the first matching unit 33 and the second matching unit 34 operate in an information processing device 51 different from the matching device 30. That is, the collation device 30 performs processing executed in the first collation unit 33, the second collation unit 34, the first motion detection unit 11, the second motion detection unit 12, and the determination unit 15. Can be dispersed. Thereby, for example, even if the computing power of the first information processing device 51 is low, the collation device 30 can quickly execute each process.

  Further, according to the present embodiment, even when a system is operated in a plurality of verification devices 30, the cost for introducing the system can be kept low.

  The reason is that the collation device 30 can manage the feature quantities (first feature quantity 103 and second feature quantity 104) used when collating the object in one information processing apparatus.

  Furthermore, the collation device can be applied to an information terminal device such as a generally known personal computer, portable terminal device, or smartphone.

(Hardware configuration example)
Each unit shown in the drawings in the embodiment described above can be regarded as a function (processing) unit (software module) of a software program. Each of these software modules may be realized by dedicated hardware. However, the division of each part shown in these drawings is a configuration for convenience of explanation, and various configurations can be assumed in mounting. An example of the hardware environment in this case will be described with reference to FIG.

  FIG. 9 is a diagram exemplarily illustrating the configuration of an information processing device 300 (computer) that can execute the terminal specifying device according to the exemplary embodiment of the present invention. That is, FIG. 9 shows the whole or a part of the collation apparatus 1 shown in FIG. 1, or the collation apparatus 10 shown in FIG. 3, the collation apparatus 20 shown in FIG. 5, and the collation apparatus 30 shown in FIG. This is a configuration of a computer (information processing apparatus) such as a server capable of realizing a collation apparatus, and represents a hardware environment capable of realizing each function in the above-described embodiment.

  The information processing apparatus 300 illustrated in FIG. 9 includes a CPU (Central_Processing_Unit) 301, a ROM (Read_Only_Memory) 302, a RAM (Random_Access_Memory) 303, a hard disk 304 (storage device), and a communication interface (Interface: hereinafter “I”). / F ”) 305, and a reader / writer 308 capable of reading and writing data stored in a storage medium 307 such as a CD-ROM (Compact_Disc_Read_Only_Memory), and these components are connected via a bus 306 (communication line). It is a general computer.

  The present invention described by taking the above embodiment as an example is the block configuration diagram (FIGS. 1, 3, 5, and 7) referred to in the description of the information processing apparatus 300 shown in FIG. This is achieved by supplying a computer program capable of realizing the functions of the flowcharts (FIGS. 2, 4, 6, and 8) and then reading the computer program to the CPU 301 of the hardware and executing it. Further, the computer program supplied to the apparatus may be stored in a nonvolatile storage device such as a readable / writable temporary storage memory (RAM 303) or a hard disk 304.

  In the above case, the computer program can be supplied to the hardware by a method of installing the computer program via various storage media 307 such as a CD-ROM or a communication line such as the Internet. Currently, a general procedure can be adopted, such as a method of downloading from the outside. In such a case, the present invention can be regarded as being configured by a code constituting the computer program or a storage medium storing the code.

In addition, a part or all of each embodiment mentioned above can be described also as the following additional remarks. However, the present invention described by way of example with the above-described embodiments is not limited to the following. That is,
(Appendix 1)
When it is detected that the image information includes the first action based on image information including information constituting an image representing the object to be collated, the object is detected as a living body, and the image When it is detected that the information does not include the first operation, a detection unit that detects that the information is not in the first operation;
When it is detected that it is the living body, the image information used when it is detected that the first action is included, and the image representing the first action by a specific object as a collation target held in the storage unit or A state in which the first operation is not performed by checking whether or not the object is the specific object based on a feature amount extracted based on any one of the images indicating that the first operation is not performed. If it is detected, the image information used when it is detected that the first operation is not included is compared with the feature amount based on the image information, and it is detected as the living body. A collation unit that determines whether or not the object is the specific object based on a result collated and a result collated when it is detected that the first operation is not performed When,
A collation device comprising:
(Appendix 2)
The detector is
Executing either one of the processing to detect that the first operation is included or the processing to detect that the first operation is not included, and executing the other processing after executing one of the processing. ,
The collation device according to Supplementary Note 1, which is characterized.
(Appendix 3)
The collation unit
Furthermore, when it is detected that it is the living body, the image information used when it is detected that it includes the first action, and the first feature amount extracted based on the image representing the first action by the specific object, Whether or not it is the specific object, and if it is detected that the first operation is not performed, the image information used when detecting that the first operation is not included, and Supplementary note 1 or Supplementary note 2, wherein whether or not it is the specific target object is collated based on a second feature amount extracted based on an image indicating that the first action by the specific target object is not performed. The verification device described in 1.
(Appendix 4)
The detector is
Whether or not the first operation is represented by a discriminator generated by learning using a statistical pattern recognition technique based on the first operation and an image indicating that the first operation is not performed, or When the state indicating whether or not the first operation is being performed is identified, and when it is identified that the state represents the first operation, it is detected that the first operation is included, and 4. The collation device according to any one of supplementary notes 1 to 3, wherein when it is identified that the state represents that the first operation is not performed, it is detected that the first operation is not included.
(Appendix 5)
The detector is
Furthermore, based on the image information, any one of a state indicating that the object is not an object, a state indicating the first operation by the object, and a state indicating that the first operation by the object is not performed The collation apparatus according to appendix 4, wherein when the object is identified as being in a state representing the first action by the object, the object is detected as the living body.
(Appendix 6)
The object is a person,
The first operation includes
Any of the movement of the person detected by the detection unit based on the image information up and down, left and right, tilting the face left and right, eye opening / closing operation of the face, eyeball movement, and mouth opening / closing operation The collation device according to any one of appendix 1 to appendix 4, wherein the collation device is an operation in which one or more operations are combined.
(Appendix 7)
The detector is
When the eye opening / closing operation is detected, and the image information detects that the person includes an eye closing operation, the object is detected as a living body, and the image information includes an eye closing operation. 7. The collation device according to any one of appendix 1 to appendix 6, wherein when it is detected that the first operation is not performed, it is detected that the first operation is not performed.
(Appendix 8)
The detector is
When the opening / closing operation of the mouth is detected, and the image information includes the operation of opening the mouth of the person, the object is detected as a living body, and the image information includes an operation of opening the mouth. 7. The collation device according to any one of appendix 1 to appendix 6, wherein when it is detected that the first operation is not performed, it is detected that the first operation is not performed.
(Appendix 9)
A verification system in which a first information processing terminal and a second information processing terminal are communicably connected via a communication network,
The first information processing terminal
When it is detected that the image information includes the first action based on image information including information constituting an image representing the object to be collated, the object is detected as a living body, and the image When it is detected that the information does not include the first operation, a detection unit that detects that the information is not in the first operation;
A first collation requesting unit for requesting collation based on the image information used when it is detected as including the first action when it is detected as the living body;
A second collation requesting unit for requesting collation based on image information used when it is detected that the first operation is not included when it is detected that the first operation is not performed; When,
A determination unit configured to determine whether or not a specific target object is to be verified based on the first verification result and the second verification result.
In response to a request from the first verification request unit, a first extracted based on image information used when detecting that the first operation is included and an image representing the first operation by the specific object. A first collation unit that collates whether or not the object is the specific object based on a feature amount and transmits the first collation result;
In response to a request from the second verification request unit, the image information used when it is detected that the first operation is not included, and an image indicating that the first operation by the specific object is not performed A second collation unit that collates whether the object is the specific object based on the second feature amount extracted based on the second feature amount, and transmits the second collation result;
A collation system comprising:
(Appendix 10)
Depending on the information processing device,
When it is detected that the image information includes the first action based on image information including information constituting an image representing the object to be collated, the object is detected as a living body, and the image When it is detected that the information does not include the first action, it is detected that the first action is not performed,
When it is detected that it is the living body, the image information used when it is detected that the first action is included, and the image representing the first action by a specific object as a collation target held in the storage unit or A state in which the first operation is not performed by checking whether or not the object is the specific object based on a feature amount extracted based on any one of the images indicating that the first operation is not performed. If it is detected, the image information used when it is detected that the first operation is not included is compared with the feature amount based on the image information, and it is detected as the living body. Determining whether or not the object is the specific object based on the result of collation and the result of collation when detecting that the first operation is not performed,
The collation method characterized by this.
(Appendix 11)
When it is detected that the image information includes the first action based on image information including information constituting an image representing the object to be collated, the object is detected as a living body, and the image A function of detecting that the first operation is not performed when it is detected that the information does not include the first operation;
When it is detected that it is the living body, the image information used when it is detected that the first action is included, and the image representing the first action by a specific object as a collation target held in the storage unit or A state in which the first operation is not performed by checking whether or not the object is the specific object based on a feature amount extracted based on any one of the images indicating that the first operation is not performed. If it is detected, the image information used when it is detected that the first operation is not included is compared with the feature amount based on the image information, and it is detected as the living body. A function of determining whether or not the object is the specific object based on a result of collation and a result of collation when it is detected that the first operation is not performed ,
A computer program characterized in that a computer is realized.

DESCRIPTION OF SYMBOLS 1 Collation apparatus 2 Detection part 3 Collation part 4 Memory | storage part 10 Collation apparatus 11 1st motion detection part 12 2nd motion detection part 13 1st collation part 14 2nd collation part 15 Determination part 20 Collation apparatus 21 1st collation apparatus Motion detection unit 30 collation device 31 first collation request unit 32 second collation request unit 33 first collation unit 34 second collation unit 50 collation system 51 first information processing device 52 second information processing device 101 image Information 102 Feature amount 103 First feature amount 104 Second feature amount 300 Information processing device 301 CPU
302 ROM
303 RAM
304 Hard Disk 305 Communication Interface 306 Bus 307 Storage Medium 308 Reader / Writer

Claims (10)

When it is detected that the image information includes the first action based on image information including information constituting an image representing the object to be collated, the object is detected as a living body, and the image When it is detected that the information does not include the first operation, a detection unit that detects that the information is not in the first operation;
When it is detected that it is the living body, the image information used when it is detected that the first action is included, and the image representing the first action by a specific object as a collation target held in the storage unit or A state in which the first operation is not performed by checking whether or not the object is the specific object based on a feature amount extracted based on any one of the images indicating that the first operation is not performed. If it is detected, the image information used when it is detected that the first operation is not included is compared with the feature amount based on the image information, and it is detected as the living body. A collation unit that determines whether or not the object is the specific object based on a result collated and a result collated when it is detected that the first operation is not performed When,
A collation device comprising: The detector is
Executing either one of the processing to detect that the first operation is included or the processing to detect that the first operation is not included, and executing the other processing after executing one of the processing. ,
The collation device according to claim 1, wherein The collation unit
Furthermore, when it is detected that it is the living body, the image information used when it is detected that it includes the first action, and the first feature amount extracted based on the image representing the first action by the specific object, Whether or not it is the specific object, and if it is detected that the first operation is not performed, the image information used when detecting that the first operation is not included, and 2. The method according to claim 1, wherein whether or not the object is the specific object is collated based on a second feature amount extracted based on an image indicating that the first operation by the specific object is not performed. Item 3. The verification device according to Item 2. The detector is
Whether or not the first operation is represented by a discriminator generated by learning using a statistical pattern recognition technique based on the first operation and an image indicating that the first operation is not performed, or When the state indicating whether or not the first operation is being performed is identified, and when it is identified that the state represents the first operation, it is detected that the first operation is included, and The collation apparatus according to claim 1, wherein when it is identified that the first operation is not performed, it is detected that the first operation is not included. The detector is
Furthermore, based on the image information, any one of a state indicating that the object is not an object, a state indicating the first operation by the object, and a state indicating that the first operation by the object is not performed The collation apparatus according to claim 4, wherein when it is identified that the object is in a state representing the first action by the object, the object is detected as the living body. The object is a person,
The first operation is:
Any of the movement of the person detected by the detection unit based on the image information up and down, left and right, tilting the face left and right, eye opening / closing operation of the face, eyeball movement, and mouth opening / closing operation 5. The collation apparatus according to claim 1, wherein the collation apparatus is an operation in which one or more operations are combined. The detector is
When the eye opening / closing operation is detected, and the image information detects that the person includes an eye closing operation, the object is detected as a living body, and the image information includes an eye closing operation. The collation apparatus according to any one of claims 1 to 6, wherein when it is detected that the first operation is not performed, it is detected that the first operation is not performed. A verification system in which a first information processing terminal and a second information processing terminal are communicably connected via a communication network,
The first information processing terminal
When it is detected that the image information includes the first action based on image information including information constituting an image representing the object to be collated, the object is detected as a living body, and the image When it is detected that the information does not include the first operation, a detection unit that detects that the information is not in the first operation;
A first collation requesting unit for requesting collation based on the image information used when it is detected as including the first action when it is detected as the living body;
A second collation requesting unit for requesting collation based on image information used when it is detected that the first operation is not included when it is detected that the first operation is not performed; When,
A determination unit configured to determine whether or not a specific target object is to be verified based on the first verification result and the second verification result.
In response to a request from the first verification request unit, a first extracted based on image information used when detecting that the first operation is included and an image representing the first operation by the specific object. A first collation unit that collates whether or not the object is the specific object based on a feature amount and transmits the first collation result;
In response to a request from the second verification request unit, the image information used when it is detected that the first operation is not included, and an image indicating that the first operation by the specific object is not performed A second collation unit that collates whether the object is the specific object based on the second feature amount extracted based on the second feature amount, and transmits the second collation result;
A collation system comprising: Depending on the information processing device,
When it is detected that the image information includes the first action based on image information including information constituting an image representing the object to be collated, the object is detected as a living body, and the image When it is detected that the information does not include the first action, it is detected that the first action is not performed,
When it is detected that it is the living body, the image information used when it is detected that the first action is included, and the image representing the first action by a specific object as a collation target held in the storage unit or A state in which the first operation is not performed by checking whether or not the object is the specific object based on a feature amount extracted based on any one of the images indicating that the first operation is not performed. If it is detected, the image information used when it is detected that the first operation is not included is compared with the feature amount based on the image information, and it is detected as the living body. Determining whether or not the object is the specific object based on the result of collation and the result of collation when detecting that the first operation is not performed,
The collation method characterized by this. When it is detected that the image information includes the first action based on image information including information constituting an image representing the object to be collated, the object is detected as a living body, and the image A function of detecting that the first operation is not performed when it is detected that the information does not include the first operation;
When it is detected that it is the living body, the image information used when it is detected that the first action is included, and the image representing the first action by a specific object as a collation target held in the storage unit or A state in which the first operation is not performed by checking whether or not the object is the specific object based on a feature amount extracted based on any one of the images indicating that the first operation is not performed. If it is detected, the image information used when it is detected that the first operation is not included is compared with the feature amount based on the image information, and it is detected as the living body. A function of determining whether or not the object is the specific object based on a result of collation and a result of collation when it is detected that the first operation is not performed ,
A computer program characterized in that a computer is realized.

Images