JP2005094710A - Reflex action information acquisition apparatus and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information acquisition apparatus for improving identifiability since a conventional information acquisition apparatus only passively photographs an identification target and when only such information disadvantageous for identification is obtained that the identification target turns its back to a camera, the occasional occurrence of a state advantageous for detecting the identification target is to be just waited, so that identification accuracy becomes low. <P>SOLUTION: In the reflex action information acquisition apparatus, identification accuracy is improved by actively creating a status advantageous for identification by inducing the reflex action of the identification target and the reflex action information acquisition apparatus comprises: an image input part 100 for inputting information of the identification target; an image display part 101 for displaying the inputted image of the identification target; an instruction input part 102 for inputting an instruction for the induction of the reflex action; a reflex action induction part 103 for applying a physical operation to induce the reflex action of the identification target; a reaction result input part 105 for inputting a result of the reflex action as an image; and a reaction result presentation part 106 for performing at least one of display, storage and recognition of the reaction result. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an apparatus that acquires information about an object and human reflex behavior, and is used for a monitoring apparatus, an identification apparatus, and the like.

  As a typical conventional monitoring device, there is a video monitoring device composed of a camera installed in a bank or a store of a store and a video deck for recording captured images. In the video monitoring apparatus, when the angle of view of the camera is widened, a wide range can be monitored, but there is a problem that analysis becomes difficult later because the resolution becomes low. On the other hand, if the angle of view of the camera is narrowed, the resolution increases, but the monitoring range becomes narrow. Therefore, a large number of monitoring cameras have to be installed, and there is a problem that costs increase.

  The anti-counterfeit monitoring devices of Patent Document 1 and Patent Document 2 prevent forgery of a monitoring image by storing a special pattern together with a laser beam in an image. Not used to cause.

The fingerprint authentication device of Patent Document 3 is an identification device that determines whether or not a detected object is a counterfeit by applying a stimulus such as heat to a finger while collecting a fingerprint image and causing the detected object to respond. As a modification, there is one that applies electrical stimulation to cause a conditional reflection to occur on the detected object and detects this, but the conditional reflection is triggered only to determine whether the detected object is a counterfeit. For example, it is not used for pressing the sensor with an appropriate pressure in order to obtain an advantageous state for the fingerprint recognition process.
Japanese Patent No. 3083784 US Pat. No. 5,764,769 Japanese Patent No. 2943337

  As described above, when the conventional identification device only captures the identification object passively and only obtains information that is disadvantageous for identification, such as the identification object is facing backward with respect to the camera. There is a problem that the identification accuracy is low because there is no choice but to wait for an accidental occurrence of a state advantageous for detection of the identification object.

  Therefore, the present invention provides a reflection behavior information acquisition apparatus and method that can input reflection behavior information with higher identification accuracy using the reflection behavior of animals and humans.

  The invention according to claim 1 is a reflex behavior inducing means for giving a physical action to induce reflex behavior on an object, and an information acquisition instruction means for instructing the reflex behavior inducing means to cause the physical action. , At least one of a still image, a moving image, an infrared image, or a sound, vibration, or ultrasonic wave generated by the reflection behavior, which is obtained by photographing the reflection behavior of the object when the reflection behavior inducing means causes the physical action. Reflection result input means for inputting one as reflection behavior information, and reflection result presentation means for displaying, storing, or recognizing the identification object for reflecting behavior information input to the reflection result input means And an apparatus for acquiring reflex behavior information.

  According to a second aspect of the present invention, there is provided a reflex action inducing means for giving a physical action for inducing a reflex action to a human, and an information acquisition instruction means for instructing the reflex action inducing means to cause the physical action. When the reflex action inducing means causes the physical action, the human face that has performed the reflex action is photographed, and the reflex result input means for inputting the face image is input by the reflex result input means. A reflection behavior information acquisition device comprising: a reflection result presentation means for displaying, storing, or performing at least one of the face recognition.

  The invention according to claim 3 is characterized in that the reflection result input means has a camera arranged at a position where a human face can be photographed when the reflection action occurs. It is an acquisition device.

  According to a fourth aspect of the present invention, the reflection result input means changes the definition and captures the human face image only when instructed by the information acquisition instruction means, or zooms in on the human face. The reflex behavior information acquisition apparatus according to claim 1, wherein the reflex action information acquisition device is used.

  The invention according to claim 5 is characterized in that the information acquisition instruction means detects the presence or absence of the object or the person, and gives the instruction when the object or the person is detected. 2. A device for acquiring reflex behavior information according to 2.

  The invention according to claim 6 is characterized in that the information acquisition instruction means performs the instruction periodically or randomly, or performs the instruction by a human operation. This is an apparatus for acquiring reflex behavior information.

  The invention according to claim 7 is characterized in that the reflex behavior inducing means induces reflex behavior in the object or the human using at least one physical action of light or sound. 2. A device for acquiring reflex behavior information according to 2.

  According to an eighth aspect of the present invention, the information acquisition instruction means inputs the object or the human image, and the object is at least one of a change in luminance of the image, a motion vector, and a background difference amount. 6. The apparatus according to claim 5, wherein the presence / absence of the person is detected.

  The invention according to claim 9 is the apparatus for acquiring reflex behavior information according to claim 1 or 2, further comprising information storage means for storing the reflex behavior information or the face image.

  The invention according to claim 10 is the apparatus for acquiring reflex behavior information according to claim 9, wherein the information information storage means stores only when the object or the person is detected.

  According to an eleventh aspect of the present invention, a still image obtained by photographing the reflex behavior of the object when the reflex behavior inducing means causes a physical action to induce the reflex behavior on the object by an instruction from the information acquisition instruction means. A reflection result input step of inputting at least one of a moving image, an infrared image, or a sound, vibration, or ultrasonic wave generated by the reflection behavior as reflection behavior information; and displaying, storing, or storing the input reflection behavior information, or And a reflection result presentation step for performing at least one of recognition of the identification object.

  According to a twelfth aspect of the present invention, when the reflex action inducing means causes a physical action for inducing a reflex action on a human by an instruction from the information acquisition instruction means, the human face that has performed the reflex action is photographed. A reflection result input step for inputting the face image, and a reflection result presentation step for displaying, storing, or performing at least one of the face recognition. This is a method for acquiring behavior information.

  In the reflection behavior information acquisition device of the invention according to claims 1 and 11 of the present invention, a physical action is applied to an object to induce reflex behavior such as turning around or uttering a voice, The identification accuracy can be increased by effectively obtaining images and sounds advantageous for the discrimination of the object.

  In the reflection behavior information acquisition device according to the second and twelfth aspects of the present invention, a physical action such as light or sound is applied to a person to induce the reflection behavior, and the face image at that time is input. Thus, the human face image can be taken from the front.

(First embodiment)
With reference to FIGS. 1 and 2, a reflex action information acquisition apparatus (hereinafter simply referred to as an information acquisition apparatus) according to a first embodiment of the present invention will be described.

  The information acquisition device of this embodiment will be described as a monitoring device that is installed at a boarding gate in an airport and identifies a suspicious person.

(1) Configuration of Information Acquisition Device FIG. 1 is a schematic configuration diagram of an information acquisition device according to this embodiment.

  An image input unit 100 for inputting the video L100 of the identification object 104, an image display unit 101 for displaying the input video L101 to the monitor, an instruction input unit 102 for inputting an instruction for reflex behavior induction by the monitor, and a reflective behavior A reflex action inducing unit 103 that induces a reflex action by giving a physical action L103 to the identification object 104 by an instruction signal L102 for inducing the image, and what kind of reflex action the identification object 104 causes is input as an image. The reflection result input unit 105 and the reflection result presentation unit 106 that presents an image input by the reflection result input unit 105 to a monitor.

  The image input unit 100 is composed of, for example, a video camera, and converts the video L100 of the identification object 104 into a video signal L101 such as an NTSC video signal composed of 30 still image sequences per second.

  The image display unit 101 is composed of, for example, a video monitor, displays the input video L101 on the screen, and presents it to the monitor.

  The instruction input unit 102 is a means for instructing the acquisition of information. The instruction input unit 102 includes, for example, a push button switch, and outputs an instruction signal L102 when pressed when it is desired to induce a reflex action on the identification object 104.

  The image display unit 101 and the instruction input unit 102 are installed, for example, in a monitoring room provided in a separate room from the boarding gate at the airport.

  The reflex behavior inducing unit 103 is composed of, for example, a flashlight installed in the vicinity of the image input unit 100, and turns on the flashlight when receiving the instruction signal L102. The image input unit 100 and the reflex behavior inducing unit 103 are installed, for example, in a high place so that the entire boarding gate of the airport can be seen.

  The reflection result input unit 105 also serves as a video camera constituting the image input unit 100, and inputs the result of the physical action of the reflection behavior inducing unit 103 on the identification target as an image.

  The reflection result presentation unit 106 also serves as a video monitor that constitutes the image display unit 101, and presents the image input by the reflection action input unit 105 to the monitor.

(2) Processing Procedure The processing procedure when a person passes through the boarding gate at the airport is described below.

  For example, when the monitoring person confirms that there is a suspicious person by looking at the monitoring video displayed on the image display unit 101, when the button of the instruction input unit 102 is pressed to specify who the person is, the reflection action The flash light of the inducing unit 103 is turned on.

  Then, when the identification object 104 is a person, a reflection action of turning around the light source is caused. It is known that when living organisms are exposed to intense light such as flashlight, this light will be a key stimulus and cause a typical behavior of looking at the direction of the light source [Reference: “Brain and Behavior”, Kiyoshi Aoki Asakura Shobo, 1986, pp.12-41]. Therefore, in this specification, the “reflective behavior” refers to behavior that a human or an animal causes unintentionally by a key stimulus as described above.

  Since the reflection behavior inducing unit 103 is installed in the vicinity of the image input unit 100, the person of the identification object 104 reflects his face to the video camera of the reflection result input unit 105 in response to intense light. Therefore, the face of the person of the identification object 104 can be seen from the front as viewed from the video camera. The reason for this kind of reflex behavior is that if a person is walking at the boarding gate of the airport, they may walk downward or may not be maliciously facing the video camera like a criminal. Since the person's face cannot be recognized, the face image from the front is copied using the reflection behavior of the person as described above.

  The reflection result input unit 105 inputs the face image of the person of the identification object 104 facing the video camera and displays it on the video monitor of the reflection result presentation unit 106. Then, the monitoring person can easily specify who the person of the identification object 104 is.

  If the object 104 to be identified is an animal, it can be recognized as a person even if the animal's shadow resembles a human being because the face image of the front can be seen though a reflex behavior occurs. . Alternatively, even if the leaves appear to be moving due to the swaying of the leaves in the wind, the reflex behavior does not occur so that the person can be easily identified.

  FIG. 2 is a flowchart showing a processing procedure for performing monitoring work using the information acquisition apparatus of this embodiment.

  The image input unit 100 continues to input the identification target video L100, and the image display unit 101 continues to display the input signal L101 on the monitor screen. The monitor looks at the video displayed on the monitor screen and presses the push button of the instruction input unit 102 when he wants to clearly see the face of the person to be identified. The instruction input unit 102 checks whether or not the push button has been pressed, and if not, returns to step S200 (step 200). When the button is pressed, the instruction input unit 102 generates an instruction signal L102 (step S201), and the reflex action inducing unit 103 turns on the flashlight (S202). Then, the person to be identified turns in a reflective manner toward the light source of the flashlight, and this image is input by the reflection result input unit 105 and displayed on the reflection result presentation unit 106.

(3) Effects of First Embodiment According to the above description, in the information acquisition apparatus of this embodiment, even if the monitoring person cannot clearly see the face image of the person to be identified, the person to be identified is a coincidence. It is possible to improve the accuracy of person identification by causing the person to be identified to reflect in the direction of the camera and clearly see the image of the front face without waiting for the face to face the camera. it can.

(Modification of the first embodiment)
Hereinafter, modifications of the first embodiment will be described.

(1) Modification 1
Although the first embodiment is applied to an airport boarding gate, the present invention is not limited to this and can be applied to shops, streets, and the like.

  When the information acquisition apparatus according to the first embodiment is applied to monitor a customer who visits a store such as a bank or a convenience store, the image input unit 100 and the reflex behavior induction unit 103 are located near the store entrance / exit, near the register, and in the store. When the reflection result input unit 105 is installed, a face image of the criminal can be taken when a burglar or shoplifting occurs. In particular, when it is installed near the doorway outside the store at night, it can effectively induce reflex behavior.

  When the information acquisition apparatus of the first embodiment is applied to monitor street passers-by, the image input unit 100, the reflex action induction unit 103, and the reflex result 105 are installed at a high place where the road can be well looked at. , You can take a face-to-face image of the criminal when there is a placement, snatching, slipping, etc.

  In addition to boarding gates in airport monitoring, installing the image input unit 100, the reflex action inducing unit 103, and the reflex result 105 in the check-in counter, security check counter, and lobby can effectively confront the arranger and suspicious person. A face image can be taken.

(2) Modification example 2
In the first embodiment, the image input unit 100 also serves as the reflection result input unit 105. However, the present invention is not limited to this, and the reflection result input unit 105 may be configured by an independent video camera.

  Similarly, although the image display unit 101 also serves as the reflection result presentation unit 106, the reflection result presentation unit 106 may be configured with an independent video monitor.

(3) Modification 3
In the first embodiment, the input video L101 is displayed on the image display unit 101 and is not saved, but may be saved.

  In this case, a storage unit 105 composed of, for example, a VTR (Video Tape Recorder) is provided between the image input unit 100 and the image display unit 101, and the input video L101 captured by the image input unit 100 is always stored. That's fine.

  As a result, since the face image of the person to be identified can be reproduced later for detailed confirmation, the accuracy of person identification can be further increased. The stored video can also be used as evidence. Alternatively, the fact that the instruction input unit 102 has been pressed is stored in the storage unit 105 as an event, so that a scene in which a person to be identified has caused a reflex action can be efficiently searched from images taken for a long time. Also good.

(4) Modification 4
The reflex behavior inducing unit 103 of the first embodiment turns on the flashlight. However, the present invention is not limited to this, and any device may be used as long as it induces a reflex behavior that can improve the identification accuracy. For example, light, sound, electromagnetic waves, vibration, or wind may be used.

  In particular, when the flashlight is turned on in the outdoors at night, it is possible to effectively induce a reflective behavior that turns around the direction of the light in a reflective manner. In the daytime, the flashlight is inconspicuous, so the effect of nighttime is not expected. However, the flashlight may be brightened, for example, a colored flashlight such as red may be used, or sound may be generated simultaneously.

  In the case of electromagnetic waves, for example, by outputting an instruction to sound a dial tone using electromagnetic waves to a mobile phone held by the person of the identification object 104, the reflection of taking out the mobile phone to the person of the identification object 104 Trigger behavior. Alternatively, when a person of the identification object 104 is driving a car while listening to the radio, a reflection action that lowers the volume may be induced by generating noise in the radio using electromagnetic waves.

  Alternatively, when a discharge sound is generated, a reflex action that turns around the direction of the sound source or a reflex action that looks around the surroundings may be induced.

(5) Modification 5
The reflex behavior inducing unit 103 of the first embodiment turns on the flashlight when the supervisor presses the instruction input unit 102, but the present invention is not limited to this, and any timing can be used as long as an advantageous state for identification can be obtained. For example, it may be periodic, random, or when the average value of the luminance of the input video L101 is calculated and the change is large.

(6) Modification 6
Although the image input unit 100 of the first embodiment always captures video with the same resolution, it may be input with high definition only when the identification target is reflected.

  For example, if the image input unit 100 is always shooting at a resolution of QVGA, 10 seconds after the instruction input unit 102 is pressed using the fact that the reflex action occurs in a very short time after acting. Only high resolution VGA resolution may be input. Alternatively, if the storage unit 105 normally records in the triple mode, the storage unit 105 may store the standard mode only for 10 seconds after the instruction input unit 102 is pressed.

  Alternatively, in order to clearly capture an image that causes a reflection action, a spotlight may be turned on only for 10 seconds after the instruction input unit 102 is pressed to illuminate the identification target. Since the reflex behavior is induced regardless of the intention of the identification object, the reflex behavior is not hindered even if the spotlight is exposed during the reflex behavior.

  Alternatively, a place where the change is large in the screen may be specified, and the place may be photographed with zoom. Alternatively, the reflection behavior may be photographed by a high-definition photographing apparatus different from the image input unit 100 such as a high-definition video camera or a digital still camera.

  FIG. 3 is a diagram for explaining the input timing in the reflection behavior and the high-definition mode. The reflex behavior is delayed by the nerve transmission time after giving a key stimulus action, and this delay time is very short and there are few individual differences. On the other hand, the time required for the completion of the reflex behavior is largely dependent on the posture of the identification object and individual differences. Therefore, when the action 300 is applied, the input is immediately started in the high-definition mode, and the storage time in the high-definition mode is set so as to return to the normal mode after a sufficient time.

(7) Modification example 7
The reflection result input unit 105 of the first embodiment inputs a video image. However, the present invention is not limited to this, and any object can be used as long as it can identify an identification target. For example, sound, ultrasonic waves, still images, moving images, vibrations, and fingerprints may be input.

  When a strange voice is generated by electrical stimulation, a sound is input as a reflection result. For example, when the user is surprised and stepped on by electrical stimulation, vibration is input as a reflection result.

  Similarly, for example, when the finger is pressed strongly by applying electrical stimulation to the back of the finger, the fingerprint is input as a reflection result.

(8) Modification 8
In the first embodiment, the reflection behavior inducing unit 103 is installed in the vicinity of the image input unit 100, and the person of the identification object 104 reflects the video camera of the reflection result input unit 105 in response to intense light. It has a structure that faces the

  However, human reactions are not all the same, and if the light is bright, it may be in the opposite direction of the light. Therefore, in consideration thereof, the reflex action inducing unit 103 may be provided on the opposite side to the image input unit 100.

  Also, an image input unit 100 is prepared, of which one image input unit 100 is provided in the vicinity of the reflex action inducing unit 103 and the other one image input unit 100 is on the opposite side of the reflex action inducing unit 103. It may be provided. In this way, the face can be photographed from the front regardless of the human reaction.

(Second Embodiment)
An information acquisition apparatus according to a second embodiment of the present invention will be described with reference to FIGS.

  The information acquisition device of this embodiment will be described as a monitoring device that is installed at a boarding gate in an airport and automatically captures the face images of all passengers passing through the gate.

(1) Configuration of Information Acquisition Device FIG. 4 is a schematic configuration diagram for explaining the information acquisition device of the present embodiment.

  An image input unit 100 that inputs the video L100 of the identification target object 104, a detection unit 107 that detects the presence of the identification target object 104 from the input video L101 and outputs an instruction signal L102 for inducing reflex behavior, and an instruction signal L102 Is input, the reflex action inducing unit 103 that induces the reflex behavior by giving a physical action L103 to the identification target object 104, and the reflection result input unit 105 that inputs the result of the reflex action of the identification target object 104 as an image, And a presentation unit 106 for storing the video input by the reflection result input unit 105.

The image input unit 100, the reflex behavior inducing unit 103, and the reflex result input unit 105 are configured in the same manner as in the first embodiment, and the image input unit 100 also serves as the reflex result input unit 105.

  The detection unit 107 is a means for instructing acquisition of information. For example, the detection unit 107 is configured by a computer having a video capturing function. The detection unit 107 captures the input video L101, calculates the average of the luminance of all pixels, and this value changes rapidly. In this case, it is detected that the monitoring object 104 is photographed on the screen, and the instruction signal L102 is output.

  The presentation unit 106 includes a hard disk, and stores the image input by the reflection result input unit 105.

(2) Processing procedure A processing procedure when a person passes through the boarding gate of the airport will be described.

  The detection unit 107 inputs the image of the airport boarding gate from the image input unit 100 and constantly calculates the average value of the luminance value. If nothing passes in front of the camera, the detection unit 107 calculates the same value. When a person or object of a different color crosses the screen, the average luminance value changes. When the amount of change exceeds a preset threshold value, it is detected that the identification object 104 has passed in front of the camera. By using the amount of change in luminance value, for example, when monitoring in a place exposed to direct sunlight, even if the shape of the shadow changes due to a natural phenomenon that changes the position of the sun, this change is very slow. In addition, since the change amount of the luminance value detected by the detection unit 106 is small, the detection unit 106 does not need to detect the identification target. When the detection unit 107 detects the identification object 104, the detection unit 107 outputs an instruction signal L102. However, the instruction signal L102 is not output for, for example, 10 seconds after the instruction signal L102 is output so that the instruction signal L102 is not continuously output when the identification object 104 continues to move in the video.

  Similar to the first embodiment, the reflex behavior inducing unit 103 turns on the flashlight to induce reflex behavior to turn the identification target 104 in the direction of the video camera of the reflex result input unit 105. Then, the reflection result input unit 105 can input the image of the front face, not just the image of the identification target 104 passing through the gate. The input face image is stored in the hard disk of the presentation unit 106.

  FIG. 5 is a flowchart showing the processing procedure of the second embodiment.

  Note that the presentation unit 106 stores an image only when the detection unit 107 detects the identification object 104, and only the processing procedure of the other configuration will be described.

  The image input unit 100 inputs an image of the identification target 104 (step S500), and outputs a reflex action induction instruction signal L102 when the detection unit 107 detects that there is a high possibility that the identification target exists. It progresses to step S502, and when that is not right, it returns to step S500 (step S501). In step S502, the reflex action inducing unit 103 turns on the flashlight with respect to the identification object 104, causes the reflex action to turn around the camera installed near the camera, and returns to step S500 (step S502).

(3) Effect of Information Acquisition Device From the above description, the second embodiment of the information acquisition device of the present invention detects that there is a high possibility that the identification object 104 exists in the input image, and at this time, It has been shown that by automatically inducing a reflective action on the identification object 104, an image advantageous for identifying a person can be efficiently stored.

(Modification of the second embodiment)
Hereinafter, various modifications of the second embodiment will be described.

  The information acquisition device of the second embodiment is applied to an airport entrance gate, but is not limited to this, and it can be applied to a store, a street, a cash dispenser, an automobile speed control device, etc. as in the first embodiment. Applicable.

  Below, the case where it applies to a motor vehicle speed control apparatus is demonstrated.

  FIG. 6 shows an example of an appearance in which the information acquisition device of this embodiment is applied to an automobile speed control device.

  The reflection behavior inducing unit 103 and the reflection result input unit 105 are installed at a high place overlooking the road, and the image input unit 100 is composed of a speed measuring device and is embedded in the road. For example, the speed measuring device can divide the distance between two points and the time when the object is detected by embedding two devices in the road that can detect the presence of an object in the upper part by transmitting the radio wave upward. The vehicle speed is measured from the deviation, and the one that is already in widespread use is used. Although not shown in the drawing, when the speed of the vehicle input by the image input unit 103 exceeds a preset threshold value, the detection unit 107 outputs an instruction signal L102 for inducing reflex behavior to the reflex behavior induction unit. Then, the flashlight is turned on, and the reflection result input unit 105 stores an image when the user turns around the camera. Although not shown in the figure, the presentation unit 106 is configured by a hard disk, and stores the driver's face image captured by the reflection result input unit 105.

  According to this, it is possible to clearly store the driver's face image when the speed violation occurred.

  As described above, the image input unit 100 may be anything other than the image as long as it can detect the presence or absence of the identification object 104 such as electromagnetic waves, sound, and light.

(Third embodiment)
An information acquisition apparatus according to a third embodiment of the present invention will be described with reference to FIGS.

  The information acquisition device of this embodiment will be described as a monitoring device that is installed at a boarding gate in an airport and identifies a suspicious person whose face image is registered in advance by facial image recognition.

(1) Configuration of Information Acquisition Device FIG. 7 is a schematic configuration diagram of the information acquisition device of this embodiment.

  An image input unit 100 that inputs the video L100 of the identification target object 104, a detection unit 107 that detects the presence of the identification target object 104 from the input video L101 and outputs an instruction signal L102 for inducing reflex behavior, and an instruction signal L102 Is input, the reflex action inducing unit 103 that induces the reflex behavior by giving a physical action L103 to the identification target object 104, and the reflection result input unit 105 that inputs the result of the reflex action of the identification target object 104 as an image, And a recognition unit 108 for recognizing an image input by the reflection result input unit 105.

The image input unit 100, the reflex behavior inducing unit 103, and the reflex result input unit 105 are configured in the same manner as in the first embodiment, and the image input unit 100 also serves as the reflex result input unit 105.

  The detection unit 107 is configured by, for example, a computer having a video capturing function, captures the input video L101, calculates the average of the luminance of all pixels, and when this value changes abruptly, the monitoring object 104 is displayed on the screen. It is detected that the image is taken and the instruction signal L102 is output.

  The recognition unit 108 is configured by, for example, a computer having a video capturing function, captures an input video L101, and recognizes a human face image by performing a block matching process with a face image template prepared in advance. As a result of the recognition, a speaker that emits an alarm sound is provided so as to be able to notify the monitoring person of the abnormality when it matches the face image prepared in advance.

(3) Specific example Here, consider a case where a person whose face image is registered in advance passes through a boarding gate at an airport.

  The detection unit 107 inputs the image of the airport boarding gate from the image input unit 100 and constantly calculates the average value of the luminance value. If nothing passes in front of the camera, the detection unit 107 calculates the same value. When a person or object of a different color crosses the screen, the average luminance value changes. When the amount of change exceeds a preset threshold value, it is detected that the identification object 104 has passed in front of the camera. By using the amount of change in luminance value, for example, when monitoring in a place exposed to direct sunlight, even if the shape of the shadow changes due to a natural phenomenon that changes the position of the sun, this change is very slow. In addition, the detection unit 106 does not need to detect the identification object. When the detection unit 107 detects the identification object 104, the detection unit 107 outputs an instruction signal L102. However, the instruction signal L102 is not output for, for example, 10 seconds after the instruction signal L102 is output so that the instruction signal L102 is not continuously output when the identification object 104 continues to move in the video.

  Similar to the first embodiment, the reflective behavior inducing unit 103 turns on the flashlight to induce the reflective behavior to turn the identification target 104 in the direction of the video camera of the reflection result input unit 105.

  The recognizing unit 108 always inputs the video imaged by the reflection result input unit 105 and recognizes whether or not a face image registered in advance is included in the input video image.

  If the identification object 104 accidentally turns around the camera of the image input unit, the recognition unit 108 can recognize the identification object 104. In this case, however, the identification object 104 does not necessarily turn around the camera. The recognition accuracy is low because it is not limited. When the person of the identification object 104 is induced to reflect, a face image facing the video camera can be input, so the identification object 104 accidentally turns around the direction of the video camera of the reflection result input unit 105. Rather than waiting passively, the recognition unit 108 can accurately recognize the face.

  For example, even if a natural phenomenon in which a leaf fluctuates occurs and the detection unit 106 detects the identification object 104 and the reflex action inducing unit 103 is exposed to a flashlight, the leaf is recognized by the recognition unit 108 as not being a face. The recognition accuracy is not adversely affected. Alternatively, for example, even if the cat crosses the front of the camera, the detection unit 106 detects the identification object 104, and the reflex action induction unit 103 turns the camera toward the cat, the recognition unit 108 does not recognize the human face image. Only the recognition does not adversely affect the recognition accuracy.

(4) Effects of Information Acquisition Device According to the above description, the third embodiment of the information acquisition device of the present invention detects that there is a high possibility that the identification target object 104 is present in the input image. By automatically inducing a reflex behavior in the identification object 104, it is possible to improve the accuracy of person identification by face image recognition.

(Modification of the third embodiment)
Hereinafter, various modifications of the third embodiment will be described.

(1) Modification 1
The information acquisition device of the third embodiment is applied to an airport entrance gate, but is not limited to this, and it is also applicable to stores, streets, cash dispensers, face image recognition devices, etc., as in the first embodiment. Applicable.

  Below, the case where it applies to a face image recognition apparatus is demonstrated.

  FIG. 8 shows an example of the appearance applied to the face image recognition device of this modification.

  A video camera 801, a video monitor 802, and a flash light 803 are incorporated in the housing 800. The video camera 801 constitutes the image input unit 100 and the reflection result input unit 105, and the video monitor 802 always displays the video input by the video camera 801. The flashlight 803 constitutes the reflex behavior inducing unit 103.

  As a general problem when recognizing his / her face image with this face image recognition device, the user cannot see the image of the front face with the video camera 801 because he / she sees his / her image on the video monitor 802. , Recognition accuracy will be low. In particular, when the image of the front face is used for learning the face image, the reduction in recognition accuracy becomes significant. However, according to this modified example, when the user comes in front of the apparatus, the detection unit 107 detects the presence of the user and turns on the flashlight 803, so that the user can be turned toward the video camera 801. As a result, the accuracy of face image recognition can be improved, and there is no need to call the user's attention by voice or screen display.

  When applied to a cash dispenser, the reflection image input unit is usually attached to the upper part of the device, but the user turns down to operate while looking at the screen installed at the lower part. The front face image cannot be input. However, this problem can be solved by arranging the reflex behavior inducing part around the camera.

(2) Modification example 2
In the third embodiment, an alarm sound is generated when a face image is recognized. However, the present invention is not limited to this, and any presentation method can be used as long as it can transmit an abnormality to a monitor. For example, the recognition event may be stored, and the input video and time at that time may be stored.

(3) Modification 3
The recognition unit 108 of the third embodiment always performs recognition using the same recognition method. However, the recognition unit 108 is not limited thereto, and may be recognized in another mode when the identification target 104 is detected by the detection unit 107. . For example, when the identification object 104 is detected by the detection unit 107, the image is divided into nine parts, and the recognition unit 108 is notified to which area the identification object 104 is detected, and the recognition unit 108 recognizes the area. You may increase the weight of the candidate of the recognition result detected within.

  Thereby, even when there are a plurality of candidates, it can be recognized with high accuracy. Alternatively, the recognition unit 108 performs face image recognition only when the identification object 104 is detected, thereby reducing misrecognition of face image recognition due to background image noise.

  The present invention can be applied to an airport entrance gate, a store, a street, a human monitoring device at a cash dispenser, an automobile speed control device, and the like.

It is a schematic block diagram of the information acquisition apparatus which concerns on the 1st Embodiment of this invention. It is a flowchart which shows the process sequence which induces a reflective action to an identification target object. It is a figure for demonstrating the timing which inputs reflective action in high definition. It is a schematic block diagram of the information acquisition apparatus which concerns on the 2nd Embodiment of this invention. It is a flowchart which shows the process sequence which induces a reflective action automatically to an identification target object. It is an example of the external appearance which applied the 2nd Embodiment of this invention to the automobile speed control. It is a schematic block diagram of the information acquisition apparatus which concerns on the 3rd Embodiment of this invention. It is an example of the external appearance which applied the 3rd Embodiment of this invention to the face image recognition apparatus.

Explanation of symbols

100: Image input unit,
101: Image display unit,
102 ... Instruction input unit,
103 ... reflex action induction part,
104 ... identification object,
105 ... reflection result input section,
106 ... reflection result presentation unit,
L100: Identification target video,
L101: Input video,
L102 ... Instruction signal,
L103 ... action

Claims (12)

Reflex behavior inducing means for giving a physical action to induce reflex behavior on an object;
Information acquisition instruction means for instructing the reflex behavior inducing means to cause the physical action;
At least one of a still image, a moving image, an infrared image, or a sound, vibration, or ultrasonic wave generated by the reflection behavior, which is obtained by photographing the reflection behavior of the object when the reflection behavior inducing means causes the physical action. Reflection result input means for inputting one as reflection behavior information;
Reflection result presentation means for displaying, storing, or recognizing the identification target object for reflecting behavior information input to the reflection result input means;
A device for acquiring reflex behavior information, comprising: Reflex behavior inducing means for giving a physical action to induce reflex behavior to humans,
Information acquisition instruction means for instructing the reflex behavior inducing means to cause the physical action;
When the reflex action inducing means causes the physical action, the reflex result input means for photographing the human face that has performed the reflex action and inputting the face image;
A reflection result presenting means for displaying, storing, or performing at least one of the face recognition;
A device for acquiring reflex behavior information, comprising: The reflection behavior information acquisition device according to claim 2, wherein the reflection result input means has a camera disposed at a position where a human face can be photographed when the reflection behavior is caused. The reflection result input means changes the definition and shoots the human face image only when instructed by the information acquisition instruction means, or zooms in and shoots the human face. The apparatus for acquiring reflex behavior information according to claim 1 or 2. 3. The acquisition of reflex behavior information according to claim 1, wherein the information acquisition instruction unit detects the presence or absence of the object or the human and performs the instruction when the object or the human is detected. apparatus. The apparatus for acquiring reflex behavior information according to claim 1, wherein the information acquisition instruction unit performs the instruction periodically or randomly, or performs the instruction by a human operation. The acquisition of reflex behavior information according to claim 1, wherein the reflex behavior inducing means induces reflex behavior in the object or the human using at least one physical action of light or sound. apparatus. The information acquisition instruction means inputs the object or the human image and detects the presence or absence of the object or the human by at least one means among a change in luminance of the image, a motion vector, and a background difference amount. The apparatus for acquiring reflex behavior information according to claim 5. The apparatus according to claim 1, further comprising an information storage unit that stores the reflection behavior information or the face image. The apparatus for acquiring reflex behavior information according to claim 9, wherein the information information storage means stores only when the object or the person is detected. A still image, a moving image, an infrared image, or an image obtained by photographing the reflex behavior of the object when the reflex behavior inducer causes a physical action to induce the reflex behavior on the object by an instruction from the information acquisition instruction unit, or A reflection result input step for inputting at least one of sound, vibration, and ultrasonic wave generated by the reflection behavior as reflection behavior information;
A reflection result presentation step of displaying, storing, or recognizing the identification object, the input reflection behavior information;
A method for acquiring reflex behavior information, characterized by comprising: When the reflex action inducing means causes a physical action for inducing a reflex action on a human by an instruction from the information acquisition instruction means, the face of the person who performed the reflex action is photographed and the face image is input. Reflection result input step;
A reflection result presentation step of displaying, storing, or performing at least one of the face recognition, the input face image;
A method for acquiring reflex behavior information, characterized by comprising:

Images