WO2017000218A1 - 活体检测方法及设备、计算机程序产品 - Google Patents

活体检测方法及设备、计算机程序产品 Download PDF

Info

Publication number
WO2017000218A1
WO2017000218A1 PCT/CN2015/082829 CN2015082829W WO2017000218A1 WO 2017000218 A1 WO2017000218 A1 WO 2017000218A1 CN 2015082829 W CN2015082829 W CN 2015082829W WO 2017000218 A1 WO2017000218 A1 WO 2017000218A1
Authority
WO
WIPO (PCT)
Prior art keywords
obstacle
display
face
condition
living body
Prior art date
Application number
PCT/CN2015/082829
Other languages
English (en)
French (fr)
Inventor
曹志敏
陈可卿
贾开
Original Assignee
北京旷视科技有限公司
北京小孔科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京旷视科技有限公司, 北京小孔科技有限公司 filed Critical 北京旷视科技有限公司
Priority to CN201580000355.3A priority Critical patent/CN105518714A/zh
Priority to PCT/CN2015/082829 priority patent/WO2017000218A1/zh
Publication of WO2017000218A1 publication Critical patent/WO2017000218A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/161—Detection; Localisation; Normalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/40—Spoof detection, e.g. liveness detection
    • G06V40/45—Detection of the body part being alive

Definitions

  • the present disclosure relates to the field of face recognition technology, and more particularly to a living body detection method and apparatus, and a computer program product.
  • face recognition systems are increasingly used in online scenarios requiring authentication in security, finance, and social security fields, such as online bank account opening, online transaction operation verification, unattended access control systems, and online social security. Online medical insurance, etc.
  • face recognition systems in addition to ensuring that the verifier's face similarity matches the database stored in the database, it is first necessary to verify that the verifier is a legitimate biological living organism. That is to say, the face recognition system needs to be able to prevent an attacker from using a photo, a video, a 3D face model, or a mask to attack.
  • Embodiments of the present disclosure provide a living body detecting method and apparatus, and a computer program product capable of controlling display of at least a part of a virtual object based on a face motion, the virtual object including a controlled object and a obstacle object,
  • the living body detection is determined to be successful in the case where the display condition of the obstacle object and/or the target condition of the controlled object and the controlled object and the obstacle object do not always meet.
  • a living body detecting method includes: detecting a face motion from a captured image; and controlling a display state of at least a portion of the virtual object displayed on the display screen according to the detected facial motion
  • the virtual object includes a controlled object and a barrier object; and a display condition of the obstacle object and/or a target condition of the controlled object is satisfied, and the controlled object and the obstacle object are not In the case of encounter, it is determined that the face in the captured image is a living face.
  • a living body detecting apparatus including: a face motion detecting device configured to detect a face motion from a captured image; and a virtual object control device configured to detect the detected The face motion controls a display state of at least a portion of the virtual object displayed on the display device, wherein the virtual object includes the controlled object and the obstacle object; and the living body determining device configured to satisfy the display condition of the obstacle object And/or the target condition of the controlled object, and the controlled object and the obstacle object do not meet each other, the face in the captured image is determined to be a living human face.
  • a living body detecting apparatus includes: one or more processors; one or more memories; computer program instructions stored in the memory, in which the computer program instructions are provided Performing the following steps when the processor is running: detecting a face motion from the captured image; controlling a display state of at least a portion of the virtual object displayed on the display device according to the detected face motion, wherein the virtual object includes The controlled object and the obstacle object; and in the case that the display condition of the obstacle object and/or the target condition of the controlled object is satisfied, and the controlled object and the obstacle object have not met at all times, the The face in the captured image is a living face.
  • a computer program product comprising one or more computer readable storage media having stored thereon computer program instructions, the computer program instructions being The computer runs the following steps: detecting a face motion from the captured image; controlling a display state of at least a portion of the virtual object displayed on the display screen according to the detected face motion, wherein the virtual object includes the controlled object and Obstacle object; and in a case where the display condition of the obstacle object and/or the target condition of the controlled object is satisfied, and the controlled object and the obstacle object do not always meet, determining the captured image
  • the face is a living face.
  • the living body detecting method and apparatus and the computer program product of the embodiments of the present disclosure by controlling the virtual object display based on the face motion and performing the living body detection according to the virtual object display, the photo and video can be effectively prevented without depending on the special hardware device. Attacks in various ways, such as 3D face models or masks, can reduce the cost of living body detection. Further, by identifying a plurality of action attributes in the face action, a plurality of state variables of the virtual object can be controlled, and the virtual object can be caused to change the display state in multiple aspects, for example, causing the virtual object to perform a complex predetermined action. Or causing the virtual object to achieve a display effect that is greatly different from the initial display effect. Therefore, the accuracy of the living body detection can be further improved, and further the application can be improved according to the present invention. The safety of the living body detection method and apparatus of the embodiment, and the application scenario of the computer program product.
  • FIG. 1 is a schematic block diagram of an electronic device for implementing a living body detecting method and apparatus of an embodiment of the present disclosure
  • FIG. 2 is a schematic flow chart of a living body detecting method according to an embodiment of the present disclosure
  • FIG. 3 is a schematic flowchart of a face motion detecting step in a living body detecting method according to an embodiment of the present disclosure
  • FIG. 4 is a schematic flowchart of a virtual object display control step in a living body detecting method according to an embodiment of the present disclosure
  • FIG. 5 is another schematic flowchart of a living body detecting method according to an embodiment of the present disclosure.
  • 6A-6B are examples of virtual objects displayed on a display screen in accordance with a first embodiment of the present disclosure
  • FIG. 7 is another schematic flowchart of a living body detecting method according to an embodiment of the present disclosure.
  • 8A and 8B are examples of virtual objects displayed on a display screen according to a second embodiment of the present disclosure.
  • FIG. 9 is another schematic flowchart of a living body detecting method according to an embodiment of the present disclosure.
  • FIG. 10 is an example of a virtual object displayed on a display screen according to a third embodiment of the present disclosure.
  • FIG. 11 is a schematic block diagram of a living body detecting apparatus according to an embodiment of the present disclosure.
  • FIG. 12 is a schematic block diagram of another living body detecting apparatus according to an embodiment of the present disclosure.
  • FIG. 13 is a schematic block diagram of a face motion detecting device in a living body detecting apparatus according to an embodiment of the present disclosure
  • FIG. 14 is a schematic block diagram of a virtual object control device in a living body detecting device according to an embodiment of the present disclosure.
  • electronic device 100 includes one or more processors 102, one or more storage devices 104, output devices 108, and image acquisition devices 110 that pass through bus system 112 and/or other forms of connection mechanisms. (not shown) interconnected. It should be noted that the components and structures of the electronic device 100 illustrated in FIG. 1 are merely exemplary and not limiting, and the electronic device 100 may have other components and structures as needed.
  • the processor 102 can be a central processing unit (CPU) or other form of processing unit with data processing capabilities and/or instruction execution capabilities, and can control other components in the electronic device 100 to perform desired functions.
  • CPU central processing unit
  • the processor 102 can be a central processing unit (CPU) or other form of processing unit with data processing capabilities and/or instruction execution capabilities, and can control other components in the electronic device 100 to perform desired functions.
  • the storage device 104 can include one or more computer program products, which can include various forms of computer readable storage media, such as volatile memory and/or nonvolatile memory.
  • the volatile memory may include, for example, a random access memory (RAM) and/or a cache or the like.
  • the nonvolatile memory may include, for example, a read only memory (ROM), a hard disk, a flash memory, or the like.
  • One or more computer program instructions can be stored on the computer readable storage medium, and the processor 102 can execute the program instructions to implement the functions (implemented by the processor) of the embodiments of the invention described below and/or Or other desired features.
  • Various applications and various data may also be stored in the computer readable storage medium, such as image data collected by the image capture device 110, and the like, and various data used and/or generated by the application.
  • the output device 108 may output various information (eg, images or sounds) to the outside (eg, a user), and may include one or more of a display, a speaker, and the like.
  • the image capture device 110 may take an image of a predetermined viewing range (eg, photos, videos, etc.) and store the captured images in the storage device 104 for use by other components.
  • a predetermined viewing range eg, photos, videos, etc.
  • the device 100 may be an electronic device integrated with a face image collection device disposed at a face image collection end, such as a smart phone, a tablet, a personal computer, a face recognition based identification device, and the like.
  • a face image collection device disposed at a face image collection end
  • the electronic device 100 can be deployed at an image acquisition end of an access control system, and can be, for example, a face recognition based identification device; in the field of financial applications, it can be deployed at a personal terminal, such as a smart phone. , tablets, personal computers, etc.
  • the output device 108 and the image capture device 110 of the exemplary electronic device 100 for implementing the living body detecting method and apparatus of the embodiments of the present disclosure may be deployed at a face image collecting end, and the processor in the electronic device 100 102 can be deployed on the server side (or cloud).
  • a face motion is detected from the captured image.
  • the other image capturing device of the image captures a grayscale or color image of a predetermined shooting range as a captured image, which may be a photo or a frame in the video.
  • the image capture device may be a camera of a smart phone, a camera of a tablet, a camera of a personal computer, or even a webcam.
  • step S210 The face motion detection in step S210 is described with reference to FIG.
  • a face key point is located in the captured image.
  • it may first be determined whether a captured face is included in the acquired image, and a face key point is located in the case where the face is detected.
  • the key points of the face are some key points on the face, such as the eyes, the corners of the eyes, the center of the eyes, the eyebrows, the highest point of the cheekbones, the nose, the tip of the nose, the nose, the mouth, the corners of the mouth, and the contour points of the face.
  • the series of face key points may include, but is not limited to, at least a portion of the above-described face key points.
  • machine learning algorithms such as Deep Learning or local feature-based regression algorithm
  • a picture can be taken based on the established face key point model. Face detection and face key point positioning in the image. For example, the position of the face key point can be iteratively optimized in the captured image, and finally the coordinate position of each face key point is obtained. As another example, a method based on cascade regression can be used to locate a face key in a captured image.
  • the positioning of face key points plays an important role in face motion recognition, however, it should be understood that the present disclosure is not limited by the specific face key point positioning method.
  • the face key point positioning in step S310 can be performed using an existing face detection and face key point localization algorithm.
  • the living body detecting method 100 of the embodiment of the present disclosure is not limited to the use of the existing face detection and face key point positioning algorithms for face key point positioning, and should cover the use of face detection and face key developed in the future. Point location algorithm for face key location.
  • image texture information is extracted from the captured image.
  • fine information of a face such as eyeball position information, mouth type information, micro-expression information, and the like, may be extracted according to pixel information in the captured image, such as brightness information of a pixel.
  • the image texture information extraction in step S320 can be performed using an existing image texture information extraction algorithm. It should be understood that the living body detecting method 100 of the embodiment of the present disclosure is not limited to performing image texture information extraction using an existing image texture information extraction algorithm, and should cover image texture information extraction using a future developed image texture information extraction algorithm.
  • steps S310 and S320 may alternatively be performed, or both may be performed. In the case where both of steps S310 and S320 are performed, they may be executed simultaneously or may be performed sequentially.
  • a value of the face action attribute is obtained based on the located face key point and/or the image texture information.
  • the facial motion attribute obtained based on the located face key points may include, for example, but is not limited to, degree of eye closure, degree of mouth opening, degree of face pitch, degree of face deflection, distance of face from camera, and the like.
  • the facial motion attribute obtained based on the image texture information may include, but is not limited to, a degree of left and right eye deflection, an eyeball vertical deflection degree, and the like.
  • the value of the face action attribute may be obtained based on the previous captured image of the current captured image and the current captured image; or the value of the face action attribute may be obtained based on the first captured image and the current captured image; Alternatively, the value of the face action attribute may be obtained based on the current captured image and the first few captured images of the currently captured image.
  • the value of the face action attribute may be obtained based on the located face key points by means of geometric learning, machine learning, or image processing.
  • multiple key points can be defined in one eye, such as 8-20 key points, for example, the inner corner of the left eye, outside The corner of the eye, the center point of the upper eyelid and the center point of the lower eyelid, and the inner corner of the right eye, the outer corner of the eye, the center point of the upper eyelid, and the center point of the lower eyelid.
  • the ratio of the inner and outer corner distances is taken as the first distance ratio X, and the degree of eye closure Y is determined based on the first distance ratio.
  • step S220 a display state of at least a portion of the virtual objects displayed on the display screen is controlled according to the detected face motion, wherein the virtual object includes the controlled object and the obstacle object.
  • the virtual object can include a first set of objects, and the first set of objects can include one or more objects. Updating the display of at least one of the first set of objects on the display screen according to the detected face motion.
  • the initial display position and/or initial display form of at least a portion of the first set of objects is predetermined or randomly determined. Specifically, for example, the motion state, display position, size, shape, color, and the like of the at least a portion of the object may be changed.
  • step S220 The operation of step S220 will be described with reference to FIG.
  • step S410 the value of the state parameter of at least a part of the virtual object is updated according to the value of the face action attribute.
  • a face action attribute can be mapped to a certain state parameter of the virtual object.
  • the user's eye degree of closure or degree of mouth opening may be mapped to the size of the virtual object, and the size of the virtual object may be updated according to the value of the user's degree of eye closure or degree of mouth opening.
  • the user's face pitch degree may be mapped to a vertical display position of the virtual object on the display screen, and the vertical display position of the virtual object on the display screen may be updated according to the value of the user's face pitch degree.
  • the ratio K1 of the degree of mouth opening in the current captured image and the degree of mouth opening in the previously captured first captured image may be calculated, and the ratio K1 of the degree of mouth opening is mapped to the size S of the virtual object.
  • the degree K2 in which the face position in the current captured image deviates from the initial center position can be calculated, and the face position is mapped to the position W of the virtual object.
  • the face action attribute may include at least one action attribute, the state parameter of the virtual object including at least one state parameter.
  • An action attribute may correspond to only one state parameter, or an action attribute may correspond to a plurality of state parameters in chronological order.
  • the mapping relationship between the face action attribute and the state parameter of the virtual object may be preset, or may be randomly determined when starting the living body detection method according to an embodiment of the present disclosure.
  • the living body detecting method according to an embodiment of the present disclosure may further include prompting a user with a mapping relationship between the face action attribute and a state parameter of the virtual object.
  • step S420 the virtual object is displayed on the display screen according to the updated value of the state parameter of the virtual object.
  • the virtual object may include a first group of objects, and the first group of objects are displayed on the display screen when the living body detecting method according to an embodiment of the present disclosure starts to be executed, and may be updated according to the face action attribute. Display of at least one of the first set of objects.
  • the virtual object may further include a second group of objects, wherein the second group of objects are not displayed on the display screen when the living body detecting method according to the embodiment of the present disclosure is started, and may be according to the first group of objects A situation is displayed to control whether at least one of the second set of objects is displayed.
  • the state parameter of at least one of the first group of objects may be a display position, a size, a shape, a color, a motion state, and the like, whereby the first group may be changed according to the value of the face action attribute.
  • the motion state, display position, size, shape, color, etc. of at least one object in the object may be a display position, a size, a shape, a color, a motion state, and the like, whereby the first group may be changed according to the value of the face action attribute.
  • the state parameter of each of the at least one object of the second group of objects may include at least a visible state, and may further include a display position, a size, a shape, a color, a motion state, and the like. Whether to display at least one of the second group of objects, that is, at least one of the second group of objects is in a visible state, may be controlled according to a display condition of at least one of the first group of objects.
  • step S230 it is determined whether the display condition of the obstacle object and/or the target condition of the controlled object are satisfied, and it is determined whether the controlled object and the obstacle object have not met at all times.
  • the display condition of the obstacle object is a condition related to the total display time of the obstacle object, and/or a condition related to the total number of the obstacle object, and/or a display with the obstacle object State related conditions.
  • the target condition of the controlled object is A condition related to the progress of the controlled object, and/or a condition related to the form or position of the controlled object.
  • the first group of objects includes a first object and a second object
  • the first object is a controlled object
  • the second object is a background object
  • the background object is a barrier object
  • the condition is that the total display time of the obstacle object reaches a predetermined timing time
  • the target condition of the controlled object is that the first object and the obstacle object do not meet each other.
  • the first group of objects further includes a third object
  • the third object is a target object of the controlled object
  • the display condition of the obstacle object is that the total display time of the obstacle object does not exceed a predetermined timing time
  • the target condition of the controlled object is that the display position of the controlled object coincides with the display position of the target object, and the display position of the controlled object and the target object have not exceeded the predetermined timing time.
  • the face in the captured image is determined to be a living human face.
  • the obstacle object is displayed under the condition that the total number of the obstacle objects reaches a predetermined number and the obstacle objects are all moved out of the display screen, wherein the total number of the obstacle objects reaches a predetermined number, and the obstacle objects are all removed.
  • the screen is displayed and the controlled object does not meet the obstacle object, the face in the captured image is determined to be a living human face.
  • the display condition of the obstacle object is that the total display time of the obstacle object does not exceed a predetermined timing time
  • the target condition of the controlled object is that the travel distance of the controlled object reaches a predetermined distance
  • the face is a living face.
  • step S240 In a case where the display condition of the obstacle object and/or the target condition of the controlled object are satisfied, and it is determined that the controlled object does not meet the obstacle object at all times, it is determined in step S240 in the captured image.
  • the face is a living face.
  • the determined image is determined in step S250. The face is not a living face.
  • the living body detecting method of the embodiment of the present disclosure by displaying various face motion parameters as state control parameters of the virtual object, displaying the virtual object on the display screen according to the face motion control, according to whether the displayed virtual object satisfies a predetermined condition To perform a living body test.
  • the virtual object includes a first group of objects, and the first group of objects are displayed on a display screen when starting to perform a living body detecting method according to an embodiment of the present disclosure, and the first group An object consists of one or more objects. Updating display of at least one of the first set of objects on a display screen according to the detected face motion, wherein the at least one of the first set of objects is a controlled object.
  • the initial display position and/or initial display form of at least a portion of the first set of objects is predetermined or randomly determined.
  • the first group of objects includes a first object that is a controlled object, the second object is a background object, and the background object is a barrier object,
  • the initial display position and/or initial display form of the first object and the obstacle object are random.
  • the obstacle object may be stationary or may be sporty. In the case where the obstacle object moves, its motion trajectory may be a straight line or a curve, and the obstacle object may move in a vertical direction, move in a horizontal direction, or move in any direction. Optionally, the motion trajectory and the motion direction of the obstacle object are also random.
  • the face action attribute includes a first action attribute
  • the state parameter of the first object includes a first state parameter of the first object
  • the first state parameter of the first object is a display of the first object Position
  • the state parameter of the second object includes a first state parameter of the second object
  • the first state parameter of the second object is a display position of the second object, according to the first action attribute
  • the value updates the value of the first state parameter of the first object and displays the first object on the display screen in accordance with the updated value of the first state parameter of the first object.
  • the predetermined condition may be that the first object does not meet the second object, or a distance between a display position of the first object and a display position of the second object exceeds a predetermined distance, the predetermined The distance may be determined according to a display size of the first object and a display size of the second object.
  • the predetermined condition may be that the first object does not meet the second object within a predetermined timing time, or between a display position of the first object and a display position of the second object The distance is more than the predetermined distance.
  • the display condition of the obstacle object is that the total display time of the obstacle object reaches a predetermined timing time
  • the target condition of the controlled object is that the first object and the obstacle object do not meet each other.
  • FIG. 5 shows an exemplary flow chart of a living body detecting method 500 according to a first embodiment of the present disclosure.
  • the display condition of the obstacle object is that the total display time of the obstacle object reaches a predetermined timing time.
  • a timer is initialized.
  • the timer can be initialized based on user input, or The timer is automatically initialized when a face is detected in the captured image, or the timer can be automatically initialized when a face predetermined motion is detected in the captured image. Further, after initializing the timer, the first group of objects are displayed on the display screen.
  • step S520 an image (first image) of a predetermined shooting range is acquired in real time as a captured image.
  • the other image capturing device of the image captures a grayscale or color image of a predetermined shooting range as a captured image, which may be a photo or a frame in the video.
  • Steps S530-S540 respectively correspond to steps S210-S220 in FIG. 2, and details are not described herein again.
  • step S540 a display state of at least a portion of the virtual objects displayed on the display screen is controlled according to the detected face motion, wherein the virtual object includes the controlled object and the obstacle object.
  • the predetermined timing time may be predetermined, and it is further determined whether the controlled object and the obstacle object have not met at all times.
  • a timeout flag may be generated when the timer exceeds the predetermined timing time, and whether the timer exceeds the predetermined timing time may be determined according to the timeout flag in step S550.
  • step S550 it is determined in step S560 that the living human face is detected, or it is determined in step S570 that the living human face is not detected, or the process returns to step S520.
  • step S550 In the case where it is determined in step S550 that the total display time of the obstacle object reaches the predetermined timing time and the controlled object does not coincide with the obstacle object, it is determined in step S560 that the living human face is detected.
  • step S570 it is determined in step S570 that the living human face is not detected.
  • step S550 if it is determined in step S550 that the total display time of the obstacle object has not reached the predetermined timing time and the controlled object does not coincide with the obstacle object, the flow returns to step S520.
  • the image (second image) of the predetermined shooting range is acquired as a captured image in real time, and steps S530-S550 are next performed.
  • the image acquired first is referred to as a first image
  • the image acquired thereafter is referred to as a second image.
  • the first image and the second image are maps within the same viewing range. Like, just the time of collection is different.
  • Steps S520-S550 shown in FIG. 5 are repeatedly executed until it is determined according to the determination result of step S550 that the total display time of the obstacle object reaches a predetermined timing time and the controlled object does not meet the obstacle object, thereby Step S560 determines that the living human face is detected; or until it is determined according to the determination result of step S550 that the controlled object meets the obstacle object, thereby determining that the living human face is not detected in step S570.
  • step S550 the determination as to whether or not the timer exceeds the predetermined timing time is performed in step S550 in FIG. 5, it should be understood that the present invention is not limited thereto, and the determination may be performed in any step of the living body detecting method according to an embodiment of the present disclosure.
  • a timeout flag is generated when the timer exceeds a predetermined timing time, and the timeout flag may directly trigger step S560 or S570 of the living body detecting method according to an embodiment of the present disclosure, that is, determine whether a living person is detected face.
  • FIG. 6A An example of the position of the first object A and the obstacle object D is shown in FIG. 6A.
  • the obstacle object D may continuously move on the display screen, and the moving direction of the obstacle object D may be random.
  • step S550 in a case where it is determined in step S550 that the timer exceeds the predetermined timing time and the first object has not met the obstacle object, it is determined in step S560.
  • the living human face is detected; if it is determined in step S550 that the timer does not exceed the predetermined timing time and the first object has not met the obstacle object, the process returns to step S520; on the other hand, in the step S550 determines that the timer does not exceed the predetermined timing time and the first object meets the obstacle object, and determines in step S570 that the living human face is not detected.
  • the first group of objects further includes a third object
  • the first object is a controlled object
  • the second object and the third object constitute a background object
  • the second object is a barrier object
  • the third object is a target object that is randomly displayed or randomly generated.
  • the state parameter of the third object may include a first state parameter of the third object, and the first state parameter of the third object is a display position of the third object.
  • the predetermined condition may be that the first object does not meet the second object and the first object coincides with the third object, or the display position of the first object and the second object The distance between the display positions exceeds a predetermined distance and the first object coincides with the third object, the predetermined distance may be determined according to a display size of the first object and a display size of the second object set.
  • the predetermined condition may be that the first object coincides with the third object and the first object does not meet the second object within a predetermined timing time, or is within a predetermined timing time The first object coincides with the third object and the distance between the display position of the first object and the display position of the second object always exceeds a predetermined distance.
  • a first object A, a second object (obstacle object) D, and a third object (target object) B are shown in FIG. 6B.
  • the obstacle object D may continuously move on the display screen, and the moving direction of the obstacle object D may be random, the first object A and the obstacle object D do not meet and the first object A and In the case where the target objects B are coincident, it is determined that the living body detection is successful.
  • the first object A does not meet the obstacle object D and the display position of the first object A coincides with the display position of the target object B within a predetermined timing time, it is determined that the living body detection is successful.
  • step S550 it is determined in step S550 whether the timer exceeds the predetermined timing time, and it is determined whether the display position of the controlled object coincides with the display position of the target object. And determining whether the controlled object does not meet the obstacle object.
  • step S550 it is determined that the timer exceeds the predetermined timing time, it is determined in step S570 that no living human face is detected; in step S550, it is determined that the timer does not exceed the predetermined timing time, If the first object coincides with the target object and the controlled object does not meet the obstacle object at all times, it is determined in step S560 that the living human face is detected; on the other hand, the timer is determined in step S550. If the predetermined time is not exceeded, the first object does not coincide with the target object, and the first object does not meet the obstacle object, the process returns to step S520.
  • the horizontal position and the vertical position of the first object A and the second object B are different.
  • the first action attribute may include the first child.
  • the action attribute and the second sub-action attribute, the first state parameter of the first object A may include a first sub-state parameter and a second sub-state parameter, the value of the first sub-state parameter being the first object A a horizontal position coordinate, the value of the second sub-state parameter is a vertical position coordinate of the first object A, and the first object A may be updated on the display screen according to a value of the first sub-action attribute
  • the horizontal position coordinates on the upper side, and the vertical position coordinates of the first object A on the display screen are updated according to the value of the second sub-action attribute.
  • the first action attribute may be defined as a position of the face in the captured image. And updating the display position of the first object A on the display screen according to the position coordinates of the face in the captured image.
  • the first sub-action attribute may be defined as a horizontal position of the face in the captured image and the second sub-action attribute is defined as a vertical position of the face in the captured image, which may be shot according to the face Horizontal position coordinates in the image to update the horizontal position coordinates of the first object A on the display screen, and update the first object A on the display screen according to the vertical position coordinates of the face in the captured image Vertical position coordinates on.
  • the first sub-action attribute may be defined as a degree of facial deflection and the second sub-action attribute may be defined as a degree of facial pitch, and then the first object A may be updated according to the value of the degree of facial deflection.
  • the horizontal position coordinates on the display screen, and the vertical position coordinates of the first object A on the display screen are updated according to the value of the degree of face pitch.
  • the virtual object includes a first group of objects and a second group of objects, the first group of objects being displayed on a display screen when starting to perform a living body detecting method according to an embodiment of the present disclosure
  • the first set of objects includes one or more objects that have not been displayed on the display screen and include one or more objects at the beginning of execution of the living body detection method according to an embodiment of the present disclosure.
  • the initial display position and/or initial display form of at least a portion of the first set of objects is predetermined or randomly determined.
  • the other at least one of the second group of objects is also displayed according to the display condition of the at least one object of the second group of objects.
  • at least one of the second set of objects may be randomly displayed.
  • An initial display position and/or an initial display form of at least a portion of the objects of the second set of objects is predetermined or randomly determined.
  • the first state parameter of each object in the first group of objects is a display position of the object
  • the first and second state parameters of each object in the second group of objects are respectively The display position and visual state of the object.
  • the first group of objects includes a first object and a second object
  • the second group of objects includes a plurality of objects
  • the first object is a controlled object
  • the second object and the Second
  • the group object is a background object
  • the background object is a barrier object
  • the initial display position and/or initial display form of the first object and the obstacle object are random.
  • the obstacle object moves
  • its motion trajectory may be a straight line or a curve
  • the obstacle object may move in a vertical direction, move in a horizontal direction, or move in any direction.
  • the motion trajectory and the motion direction of the obstacle object are also random.
  • the face action attribute includes a first action attribute
  • the state parameter of the first object includes a first state parameter of the first object
  • the first state parameter of the first object is a display of the first object Positioning, updating a value of the first state parameter of the first object according to the value of the first action attribute, and displaying the value on the display screen according to the updated value of the first state parameter of the first object Said the first object.
  • the predetermined condition may be that the first object does not meet the obstacle object within a predetermined time, the first object does not meet a predetermined number of obstacle objects, or is within a predetermined time The first object does not meet a predetermined number of obstacle objects.
  • FIG. 7 shows an exemplary flow chart of a living body detection method 700 in accordance with a second embodiment of the present disclosure.
  • the obstacle object is displayed under the condition that the total number of the obstacle objects reaches a predetermined number and the obstacle objects are all moved out of the display screen.
  • an image (first image) of a predetermined shooting range is acquired in real time as a captured image.
  • the other image capturing device of the image captures a grayscale or color image of a predetermined shooting range as a captured image, which may be a photo or a frame in the video.
  • Steps S720-S730 respectively correspond to steps S530-S540 in FIG. 5, and details are not described herein again.
  • step S740 It is determined in step S740 whether the controlled object and the obstacle object meet. In the case where it is determined in step S740 that the controlled object meets the obstacle object, it is determined in step S790 that the living human face is not detected.
  • step S750 it is determined in step S750 whether at least a part of the obstacle object moves out of the display screen. In the case where it is determined in step S750 that no obstacle object has moved out of the display screen, the flow returns to step S710.
  • Step S760 determines whether a predetermined number of obstacle objects have been displayed.
  • step S765 it is determined in step S765 whether the obstacle objects are all moved out of the display screen, and in a case where it is determined in step S765 that the obstacle objects are all moved out of the display screen, in step S780 It is determined that the living human face is detected; if it is determined in step S765 that the obstacle object has not all moved out of the display screen, the process returns to step S710.
  • step S760 In the case where it is determined in step S760 that a predetermined number of obstacle objects have not been displayed, at least one of the second group of objects is displayed, and at least one of the second group of objects and the obstacle object has not been removed yet in step S770 The objects of the display screen together become a new obstacle object, and then return to step S710.
  • the living body detecting methods described in FIGS. 5 and 7 may be combined as needed, in which case the predetermined condition may be that the first object does not meet a predetermined number of obstacle objects within a predetermined timing time.
  • the image (second image) of the predetermined shooting range is acquired as a captured image in real time, and steps S720-S740 are next performed.
  • the image acquired first is referred to as a first image
  • the image acquired thereafter is referred to as a second image. It should be understood that the first image and the second image are images within the same viewing range, only the time of acquisition is different.
  • Steps S710-S740 shown in Fig. 7 are repeatedly performed until it is determined in step S780 that the living human face is detected; or until it is determined in step S790 that the living human face is not detected.
  • FIG. 8A An example of the position of the first object A and the obstacle object D is shown in FIG. 8A.
  • the obstacle object D may continuously move on the display screen, and the moving direction of the obstacle object D may be random.
  • the living body detection is successful.
  • the living body detection is successful.
  • the living body detection is successful.
  • the living body detection is successful.
  • the first group of objects further includes a third object, the second object and a third object structure
  • the third object is a target object.
  • the predetermined condition may be that the first object does not meet the obstacle object and the first object coincides with the third object within a predetermined timing time.
  • the first object A, the second object (obstacle) D, and the third object (target object) B of the first group of objects, and the obstacle objects D1 and D2 of the second group of objects are shown in FIG. 8B.
  • the obstacle object may continuously move on the display screen, and the moving direction of the obstacle object D may be random, where the first object A and the obstacle object do not meet and the first object A and the object In the case where the target objects B overlap, it is determined that the living body detection is successful.
  • the first object A does not meet the obstacle object and the display position of the first object A coincides with the display position of the target object B within a predetermined timing time it is determined that the living body detection is successful.
  • step S740 it may be determined in step S740 that the first object A does not meet the currently displayed obstacle object, and in step S750, it may be determined that the currently displayed obstacle object moves out of the display screen, and may be determined to be displayed in step S760.
  • the number of obstacle objects has not reached the predetermined number, so a new obstacle object is displayed on the display screen at step S770, and the flow returns to step S710.
  • step S740 it may be determined that the first object A does not meet the currently displayed obstacle object, and in step S750, it may be determined that the currently displayed obstacle object moves out of the display screen, and in step S760, the number of obstacle objects that have been displayed may be determined. The predetermined number has been reached, and it can be determined in step S765 that all obstacle objects have been moved out of the display screen, so it is determined in step S780 that the living human face is detected.
  • the virtual object includes a first group of objects and a second group of objects, and the first group of objects are displayed on a display screen when starting to perform a living body detecting method according to an embodiment of the present disclosure, and
  • the first set of objects includes one or more objects that have not been displayed on the display screen and include one or more objects at the beginning of execution of the living body detection method according to an embodiment of the present disclosure.
  • the initial display position and/or initial display form of at least a portion of the first set of objects is predetermined or randomly determined.
  • Displaying the second group of objects according to a display condition of at least one of the first group of objects At least one object.
  • the other at least one of the second group of objects is also displayed according to the display condition of the at least one object of the second group of objects.
  • at least one of the second set of objects may be randomly displayed.
  • An initial display position and/or an initial display form of at least a portion of the objects of the second set of objects is predetermined or randomly determined.
  • the objects in the second group of objects are uncontrolled objects, that is, background objects, and the background objects are obstacle objects.
  • the first state parameter of each object in the first group of objects is a display position of the object
  • the first and second state parameters of each object in the second group of objects are respectively The display position and visual state of the object.
  • the predetermined condition may be that the first object does not meet the obstacle object within a predetermined time and the traveling distance of the controlled object reaches a predetermined distance, the first object and a predetermined number The obstacle objects do not meet and the traveled distance of the controlled object reaches a predetermined distance, or the first object does not meet a predetermined number of obstacle objects within a predetermined time and the traveled distance of the controlled object reaches a predetermined distance.
  • FIG. 9 shows an exemplary flow chart of a living body detection method 900 in accordance with a third embodiment of the present disclosure.
  • the display condition of the obstacle object is that the total display time of the obstacle object does not exceed a predetermined timing time
  • the target condition of the controlled object is that the travel distance of the controlled object reaches a predetermined distance.
  • the determination is performed.
  • the face in the captured image is a living face.
  • a timer is initialized.
  • the timer may be initialized according to user input, or the timer may be automatically initialized when a face is detected in the captured image, or may be automatically initialized when a predetermined action of the face is detected in the captured image. Further, after initializing the timer, the first group of objects are displayed on the display screen.
  • an image (first image) of a predetermined shooting range is acquired in real time as a captured image.
  • the other image capturing device of the image captures a grayscale or color image of a predetermined shooting range as a captured image, which may be a photo or a frame in the video.
  • Steps S930-S940 correspond to steps S530-S540 in FIG. 5, respectively, and details are not described herein.
  • step S950 It is determined in step S950 whether the controlled object and the obstacle object meet. In the case where it is determined in step S950 that the controlled object meets the obstacle object, it is determined in step 960 that the living human face is not detected.
  • step S950 If it is determined in step S950 that the controlled object does not meet the obstacle object, it is determined in step S970 whether the total display time of the obstacle object reaches a predetermined timing time, and the predetermined timing time may be predetermined. And determining whether the travel distance of the controlled object reaches a predetermined distance, the predetermined distance may be predetermined.
  • a timeout flag may be generated when the timer exceeds the predetermined timing time, and whether the timer exceeds the predetermined timing time may be determined according to the timeout flag in step S970.
  • step S970 it is determined in step S980 that the living face is detected, or it is determined in step S960 that the living face is not detected, or the process returns to step S920.
  • step S970 In the case where it is determined in step S970 that the total display time of the obstacle object reaches the predetermined timing time and the travel distance of the controlled object does not reach the predetermined distance, it is determined in step S980 that the living human face is not detected.
  • step S970 it is determined that the total display time of the obstacle object has not reached or just reached the predetermined timing time and the travel distance of the controlled object reaches the predetermined distance. It is determined in step S960 that the living human face is detected.
  • step S970 if it is determined in step S970 that the total display time of the obstacle object has not reached the predetermined timing time and the travel distance of the controlled object has not reached the predetermined distance, the flow returns to step S920.
  • the image (second image) of the predetermined shooting range is acquired as a captured image in real time, and steps S930-S950 are next performed.
  • the image acquired first is referred to as a first image
  • the image acquired thereafter is referred to as a second image. It should be understood that the first image and the second image are images within the same viewing range, only the time of acquisition is different.
  • Steps S920-S950 shown in Fig. 9 are repeatedly performed until it is determined in step S980 that the living human face is detected; or it is determined in step S960 that the living human face is not detected.
  • the living body detection methods described in FIGS. 7 and 9 may be combined as needed, in which case the predetermined condition may be that the first object does not meet a predetermined number of obstacle objects within a predetermined timing time and The travel distance of the first object exceeds a predetermined distance.
  • the first group of objects includes a first object and a second object, and may be rooted in step S940.
  • the display of the first object and the second object on the display screen is updated according to the detected face motion.
  • the obstacle object in the second group of objects is also displayed according to the display condition of the second object, and the new one of the second group of objects may also be displayed according to the display situation of the obstacle object in the second group of objects. Obstacle object.
  • the obstacle object in the second group of objects may also be randomly displayed. Specifically, the horizontal display position of the first object and the horizontal and vertical display positions of the obstacle object in the first and second groups of objects are updated according to the detected face motion.
  • a fixed number of obstacle objects are displayed on the display screen at any one time, and when any obstacle object disappears, the new obstacle object is displayed such that a fixed number of obstacle objects remain on the display screen.
  • the face action attribute may include a first action attribute and a second action attribute
  • the state parameter of the first object includes first and second state variables of the first object
  • the second state parameter is a travel parameter and a horizontal position of the first object, respectively, and the travel parameter may be a motion speed, a travel distance, or the like.
  • a value of a motion speed of the first object is updated according to a value of the first motion attribute
  • a first object is updated according to a value of the second motion attribute.
  • Horizontal position coordinates are subsequently, according to the value of the moving speed of the first object A, the distance between the first object A and the obstacle object D (which may include a horizontal distance and a vertical distance), and the level of the first object A
  • the position coordinates determine the display position of the obstacle object D and the first object A. For example, in a case where the target advancing direction of the first object is the road extending direction (the direction in which the road is narrowed as shown in FIG.
  • the vertical display position of the first object A remains unchanged, a value of a moving speed of the first object A and a vertical distance between the first object A and the obstacle object D, determining whether to continue displaying the obstacle object D and a display position of the obstacle object D, and Determining a display position of the first object A according to the horizontal position coordinates of the first object A.
  • the first object A may be a car
  • the obstacle object D may be a stone randomly generated on a road ahead of the car
  • the first action attribute may be a degree of face pitch
  • the second action The attribute may be a degree of facial deflection
  • the first state parameter and the second state parameter of the first object A may be the motion speed and the horizontal position of the first object, respectively.
  • the face normal state corresponds to the motion speed V0
  • the face 30 degree or 45 degree bottom view state corresponds to the highest motion speed VH
  • the face 30 degree or 45 degree top view state corresponds to the lowest motion speed VL, according to the face pitch degree
  • the value of the face determines the speed of movement of the first object.
  • the face front view state may correspond to the intermediate position P0
  • the face 30 degree or 45 degree left bias state may correspond to the left edge position PL
  • the face 30 degree or 45 degree right deviation state corresponds to the right edge position.
  • PR determines the horizontal position coordinate of the first object according to the value of the degree of face deflection (for example, the face deflection angle).
  • the state parameter of the first object further comprises a third state parameter of the first object, and the third state parameter may be a travel distance of the first object.
  • the third state parameter may be a travel distance of the first object.
  • the living body detecting device may be an electronic device integrated with a face image capturing device, such as a smart phone, a tablet computer, a personal computer, a face recognition based identification device, or the like.
  • the living body detecting apparatus may further include a separate face image collecting device and a detecting processing device, the detecting processing device may receive the captured image from the face image collecting device, and perform living body according to the received captured image Detection.
  • the detection processing device may be a server, a smart phone, a tablet computer, a personal computer, a face recognition based identification device, or the like.
  • the living body detecting apparatus Since the details of the various operations performed by the living body detecting apparatus are substantially the same as those of the living body detecting method described above with respect to FIGS. 2-4, in order to avoid repetition, only the living body detecting apparatus will be briefly described below, and the description will be omitted. A description of the same details.
  • the living body detecting apparatus 1100 includes a face motion detecting device 1110, a virtual object control device 1120, and a living body determining device 1130.
  • the face motion detecting device 1110, the virtual object control device 1120, and the living body determining device 1130 can be realized by the processor 102 shown in FIG. 1.
  • the living body detecting apparatus 1200 includes an image capturing device 1240, a face motion detecting device 1110, a virtual object control device 1120, a living body determining device 1130, a display device 1250, and a storage device 1260.
  • the image capture device 1240 can be as shown in FIG. 1
  • the image capturing device 110 is implemented, and the face motion detecting device 1110, the virtual object control device 1120, and the living body determining device 1130 can be implemented by the processor 102 shown in FIG. 1, and the display device 1250 can be the output device shown in FIG. 108 implementations, storage device 1260 can be implemented by storage device 104 shown in FIG.
  • the image capturing device 1240 in the living body detecting device 1200 or other image capturing device that can transmit an image to the living body detecting device 1100 or 1200 independently of the living body detecting device 1100 or 1200 can be used to acquire the gradation of the predetermined shooting range or
  • the color image is a captured image, and the captured image may be a photo or a frame in the video.
  • the image capture device may be a camera of a smart phone, a camera of a tablet, a camera of a personal computer, or even a webcam.
  • the face motion detecting device 1110 is configured to detect a face motion from the captured image.
  • the face motion detecting device 1110 may include a key point positioning device 1310, a texture information extracting device 1320, and an action attribute determining device 1330.
  • the keypoint locating device 1310 is configured to locate a human key point in the captured image. As an example, the key point locating device 1310 may first determine whether a captured face is included in the acquired image, and locate a face key point in the case where the face is detected. The details of the operation of the key point locating device 1310 are the same as those described in step S310, and details are not described herein again.
  • the texture information extracting means 1320 is configured to extract image texture information from the captured image.
  • the texture information extracting device 1320 may extract fine information of a face, such as eyeball position information, mouth shape information, micro-expression information, and the like, according to pixel information in the captured image, such as brightness information of a pixel.
  • the action attribute determining means 1330 obtains a value of the face action attribute based on the located face key point and/or the image texture information.
  • the facial motion attribute obtained based on the located face key points may include, for example, but is not limited to, degree of eye closure, degree of mouth opening, degree of face pitch, degree of face deflection, distance of face from camera, and the like.
  • the facial motion attribute obtained based on the image texture information may include, but is not limited to, a degree of left and right eye deflection, an eyeball vertical deflection degree, and the like.
  • the details of the operation of the action attribute determining means 1330 are the same as those described in the step S330, and details are not described herein again.
  • the virtual object control device 1120 is configured to control a display state of at least a portion of the virtual objects displayed on the display device 1250 according to the detected face motion, wherein the virtual object includes the controlled object and the obstacle object.
  • the virtual object can include a first set of objects, and the first set of objects can include one or more objects. Updating the display of at least one of the first set of objects on the display screen according to the detected face motion.
  • the initial display position and/or initial display form of at least a portion of the first set of objects is predetermined or randomly determined. Specifically, for example, the motion state, display position, size, shape, color, and the like of the at least a portion of the object may be changed.
  • the virtual object control device 1120 may include a face action mapping device 1410 and a virtual object presenting device 1420.
  • the face motion mapping device 1410 updates the value of the state parameter of the virtual object based on the value of the face action attribute.
  • a face action attribute can be mapped to a certain state parameter of the virtual object.
  • the user's eye degree of closure or degree of mouth opening may be mapped to the size of the virtual object, and the size of the virtual object may be updated according to the value of the user's degree of eye closure or degree of mouth opening.
  • the user's face pitch degree may be mapped to a vertical display position of the virtual object on the display screen, and the vertical display position of the virtual object on the display screen may be updated according to the value of the user's face pitch degree.
  • the mapping relationship between the face action attribute and the state parameter of the virtual object may be preset.
  • the face action attribute may include at least one action attribute, the state parameter of the virtual object including at least one state parameter.
  • a motion attribute may correspond to only one state parameter, or a motion attribute may correspond to a plurality of state parameters in chronological order.
  • the virtual object presentity device 1420 presents the virtual object according to the updated value of the state parameter of the virtual object.
  • the virtual object rendering device 1420 can update the display of at least one of the first set of objects.
  • the virtual object rendering device 1420 can also display a new virtual object, ie a virtual object in the second set of objects.
  • the virtual object rendering device 1420 can also update the display of at least one of the second set of objects.
  • the living body judging device 1130 is configured to determine whether it is determined whether the display condition of the obstacle object and/or the target condition of the controlled object are satisfied, and whether the controlled object and the obstacle object have not met at all times.
  • the display condition of the obstacle object is a condition related to the total display time of the obstacle object, and/or a condition related to the total number of the obstacle object, and/or a display with the obstacle object State related conditions.
  • the target condition of the controlled object is A condition related to the progress of the controlled object, and/or a condition related to the form or position of the controlled object.
  • the first group of objects includes a first object and a second object
  • the first object is a controlled object
  • the second object is a background object
  • the background object is a barrier object
  • the condition is that the total display time of the obstacle object reaches a predetermined timing time
  • the target condition of the controlled object is that the first object and the obstacle object do not meet each other.
  • the first group of objects further includes a third object
  • the third object is a target object of the controlled object
  • the display condition of the obstacle object is that the total display time of the obstacle object does not exceed a predetermined timing time
  • the target condition of the controlled object is that the display position of the controlled object coincides with the display position of the target object, and the display position of the controlled object and the target object have not exceeded the predetermined timing time.
  • the face in the captured image is determined to be a living human face.
  • the obstacle object is displayed under the condition that the total number of the obstacle objects reaches a predetermined number and the obstacle objects are all moved out of the display screen, wherein the total number of the obstacle objects reaches a predetermined number, and the obstacle objects are all removed.
  • the screen is displayed and the controlled object does not meet the obstacle object, the face in the captured image is determined to be a living human face.
  • the display condition of the obstacle object is that the total display time of the obstacle object does not exceed a predetermined timing time
  • the target condition of the controlled object is that the travel distance of the controlled object reaches a predetermined distance
  • the face is a living face.
  • the face action mapping device 1410 and the virtual object presentation device 1420 can perform various operations in the first to third embodiments described above, and details are not described herein again.
  • the living body detecting apparatuses 1100 and 1200 may further include a timer for timing a predetermined timing time.
  • the timer can also be implemented by the processor 102.
  • the timer may be initialized according to user input, or the timer may be automatically initialized when a face is detected in the captured image, or may be automatically initialized when a predetermined action of the face is detected in the captured image.
  • the biological determination device 1130 is configured to determine whether the display condition of the obstacle object is satisfied based on the timer.
  • the storage device 1260 is configured to store the captured image. In addition, the storage device 1260 is further configured to store a state parameter and a state parameter value of the virtual object. In addition, the storage device 1260 It is also used to store virtual objects presented by the virtual object presentation device 1420 and store background images or the like to be displayed on the display device 1250.
  • the storage device 1260 can store computer program instructions that, when executed by the processor 102, can implement a living body detection method in accordance with an embodiment of the present disclosure, and/or can implement an embodiment in accordance with the present disclosure.
  • a computer program product comprising a computer readable storage medium on which computer program instructions are stored.
  • the computer program instructions may implement a living body detecting method according to an embodiment of the present disclosure while being operated by a computer, and/or may implement a key point positioning device, a texture information extracting device, and an action in the living body detecting device according to an embodiment of the present disclosure.
  • the attribute determines all or part of the functionality of the device.
  • the living body detecting method and apparatus and the computer program product of the embodiments of the present disclosure by controlling the virtual object display based on the face motion and performing the living body detection according to the virtual object display, the photo and video can be effectively prevented without depending on the special hardware device. Attacks in various ways, such as 3D face models or masks, can reduce the cost of living body detection. Further, by identifying a plurality of action attributes in the face action, a plurality of state variables of the virtual object can be controlled, and the virtual object can be caused to change the display state in multiple aspects, for example, causing the virtual object to perform a complex predetermined action. Or causing the virtual object to achieve a display effect that is greatly different from the initial display effect. Therefore, the accuracy of the living body detection can be further improved, and further, the safety of applying the living body detecting method and apparatus according to the embodiment of the present invention and the application scenario of the computer program product can be improved.
  • the computer readable storage medium can be any combination of one or more computer readable storage media.
  • the computer readable storage medium may include, for example, a memory card of a smart phone, a storage component of a tablet, a hard disk of a personal computer, a random access memory (RAM), a read only memory (ROM), an erasable programmable read only memory. (EPROM), Portable Compact Disk Read Only Memory (CD-ROM), USB memory, or any combination of the above storage media.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Image Analysis (AREA)
  • Processing Or Creating Images (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

一种活体检测方法及设备、以及计算机程序产品，属于人脸识别技术领域。所述活体检测方法，包括：从拍摄图像中检测人脸动作；根据所检测的人脸动作控制在显示屏幕上显示的虚拟对象中至少一部分的显示状态，其中，所述虚拟对象包括被控对象和障碍对象；在满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。通过基于人脸动作控制虚拟对象显示并根据虚拟对象显示进行活体检测，可以有效地防范照片、视频、3D人脸模型或者面具等多种方式的攻击。

Description

活体检测方法及设备、计算机程序产品 技术领域
本公开涉及人脸识别技术领域，更具体地涉及一种活体检测方法及设备、以及计算机程序产品。
背景技术
当前，人脸识别系统越来越多地应用于安防、金融、社保领域需要身份验证的线上场景，如线上银行开户、线上交易操作验证、无人值守的门禁系统、线上社保办理、线上医保办理等。在这些高安全级别的应用领域，除了确保被验证者的人脸相似度符合数据库中存储的底库，首先需要验证被验证者是一个合法的生物活体。也就是说，人脸识别系统需要能够防范攻击者使用照片、视频、3D人脸模型、或者面具等方式进行攻击。
目前市场上的技术产品中还没有公认成熟的活体验证方案，已有的技术要么依赖特殊的硬件设备(诸如，红外相机、深度相机)，或者只能防范简单的静态照片攻击。
因此，需要既不依赖于特殊的硬件设备又能够有效地防范照片、视频、3D人脸模型或者面具等多种方式的攻击的人脸识别方式。
发明内容
鉴于上述问题而提出了本发明。本公开实施例提供了一种活体检测方法及设备、以及计算机程序产品，其能够基于人脸动作控制虚拟对象中至少一部分的显示，所述虚拟对象包括被控对象和障碍对象，在满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且所述被控对象与所述障碍对象一直不相遇的情况下确定活体检测成功。
根据本公开实施例的一个方面，提供了一种活体检测方法，包括：从拍摄图像中检测人脸动作；根据所检测的人脸动作控制在显示屏幕上显示的虚拟对象中至少一部分的显示状态，其中，所述虚拟对象包括被控对象和障碍对象；以及在满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
根据本公开实施例的另一方面，提供了一种活体检测设备，包括：人脸动作检测装置，被配置为从拍摄图像中检测人脸动作；虚拟对象控制装置，被配置为根据所检测的人脸动作控制在显示装置上显示的虚拟对象中至少一部分的显示状态，其中，所述虚拟对象包括被控对象和障碍对象；以及活体判断装置，被配置为在满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
根据本公开实施例的又一方面，提供了一种活体检测设备，包括：一个或多个处理器；一个或多个存储器；存储在所述存储器中的计算机程序指令，在所述计算机程序指令被所述处理器运行时执行以下步骤：从拍摄图像中检测人脸动作；根据所检测的人脸动作控制在显示装置上显示的虚拟对象中至少一部分的显示状态，其中，所述虚拟对象包括被控对象和障碍对象；以及在满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
根据本公开实施例的再一方面，提供了一种计算机程序产品，包括一个或多个计算机可读存储介质，所述计算机可读存储介质上存储了计算机程序指令，所述计算机程序指令在被计算机运行时执行以下步骤：从拍摄图像中检测人脸动作；根据所检测的人脸动作控制在显示屏幕上显示的虚拟对象中至少一部分的显示状态，其中，所述虚拟对象包括被控对象和障碍对象；以及在满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
根据本公开实施例的活体检测方法及设备、以及计算机程序产品，通过基于人脸动作控制虚拟对象显示并根据虚拟对象显示进行活体检测，可以不依赖于特殊的硬件设备来有效地防范照片、视频、3D人脸模型或者面具等多种方式的攻击，从而可以降低活体检测的成本。更进一步，通过识别人脸动作中的多个动作属性，可以控制虚拟对象的多个状态参量，可以使得所述虚拟对象在多个方面改变显示状态，例如使得所述虚拟对象执行复杂的预定动作、或者使得所述虚拟对象实现与初始显示效果有很大不同的显示效果。因此，可以进一步提高活体检测的准确度，并且进而可以提高应用根据本发明 实施例的活体检测方法及设备、以及计算机程序产品的应用场景的安全性。
附图说明
通过结合附图对本公开实施例进行更详细的描述，本公开的上述以及其它目的、特征和优势将变得更加明显。附图用来提供对本公开实施例的进一步理解，并且构成说明书的一部分，与本公开实施例一起用于解释本公开，并不构成对本公开的限制。在附图中，相同的参考标号通常代表相同部件或步骤。
图1是用于实现本公开实施例的活体检测方法和设备的电子设备的示意性框图；
图2是根据本公开实施例的活体检测方法的示意性流程图；
图3是根据本公开实施例的活体检测方法中的人脸动作检测步骤的示意性流程图；
图4是根据本公开实施例的活体检测方法中的虚拟对象显示控制步骤的示意性流程图；
图5是根据本公开实施例的活体检测方法的另一示意性流程图；
图6A－6B是根据本公开第一实施例的在显示屏幕上显示的虚拟对象的示例；
图7是根据本公开实施例的活体检测方法的另一示意性流程图；
图8A和图8B是根据本公开第二实施例的在显示屏幕上显示的虚拟对象的示例；
图9是根据本公开实施例的活体检测方法的另一示意性流程图；
图10是根据本公开第三实施例的在显示屏幕上显示的虚拟对象的示例；
图11是根据本公开实施例的活体检测设备的示意性框图；
图12是根据本公开实施例的另一活体检测设备的示意性框图；
图13是根据本公开实施例的活体检测设备中的人脸动作检测装置的示意性框图；以及
图14是根据本公开实施例的活体检测设备中的虚拟对象控制装置的示意性框图。
具体实施方式
为了使得本公开的目的、技术方案和优点更为明显，下面将参照附图详细描述根据本公开的示例实施例。显然，所描述的实施例仅仅是本公开的一部分实施例，而不是本公开的全部实施例，应理解，本公开不受这里描述的示例实施例的限制。基于本公开中描述的本公开实施例，本领域技术人员在没有付出创造性劳动的情况下所得到的所有其它实施例都应落入本公开的保护范围之内。
首先，参照图1来描述用于实现本公开实施例的活体检测方法和设备的示例性电子设备100。
如图1所示，电子设备100包括一个或多个处理器102、一个或多个存储装置104、输出装置108、以及图像采集装置110，这些组件通过总线系统112和/或其它形式的连接机构(未示出)互连。应当注意，图1所示的电子设备100的组件和结构只是示例性的，而非限制性的，根据需要，所述电子设备100也可以具有其他组件和结构。
所述处理器102可以是中央处理单元(CPU)或者具有数据处理能力和/或指令执行能力的其它形式的处理单元，并且可以控制所述电子设备100中的其它组件以执行期望的功能。
所述存储装置104可以包括一个或多个计算机程序产品，所述计算机程序产品可以包括各种形式的计算机可读存储介质，例如易失性存储器和/或非易失性存储器。所述易失性存储器例如可以包括随机存取存储器(RAM)和/或高速缓冲存储器(cache)等。所述非易失性存储器例如可以包括只读存储器(ROM)、硬盘、闪存等。在所述计算机可读存储介质上可以存储一个或多个计算机程序指令，处理器102可以运行所述程序指令，以实现下文所述的本发明实施例中(由处理器实现)的功能以及/或者其它期望的功能。在所述计算机可读存储介质中还可以存储各种应用程序和各种数据，例如所述图像采集装置110采集的图像数据等以及所述应用程序使用和/或产生的各种数据等。
所述输出装置108可以向外部(例如用户)输出各种信息(例如图像或声音)，并且可以包括显示器和扬声器等中的一个或多个。
所述图像采集装置110可以拍摄预定取景范围的图像(例如照片、视频等)，并且将所拍摄的图像存储在所述存储装置104中以供其它组件使用。
作为示例，用于实现本公开实施例的活体检测方法和设备的示例性电子 设备100可以是布置在人脸图像采集端的集成了人脸图像采集装置的电子设备，诸如智能手机、平板电脑、个人计算机、基于人脸识别的身份识别设备等。例如，在安防应用领域，所述电子设备100可以部署在门禁系统的图像采集端，并且可以例如为基于人脸识别的身份识别设备；在金融应用领域，可以部署在个人终端处，诸如智能电话、平板电脑、个人计算机等。
替代地，用于实现本公开实施例的活体检测方法和设备的示例性电子设备100的输出装置108和图像采集装置110可以部署在人脸图像采集端，而所述电子设备100中的处理器102可以部署在服务器端(或云端)。
下面，将参考图2来描述根据本公开实施例的人脸检测方法200。
在步骤S210，从拍摄图像中检测人脸动作。具体地，可以利用如图1所示的用于实现本公开实施例的人脸检测方法的电子设备100中的图像采集装置110或者独立于所述电子设备100的可以向所述电子设备100传送图像的其它图像采集装置，采集预定拍摄范围的灰度或彩色图像作为拍摄图像，所述拍摄图像可以是照片，也可以是视频中的一帧。所述图像采集设备可以是智能电话的摄像头、平板电脑的摄像头、个人计算机的摄像头、或者甚至可以是网络摄像头。
参考图3来描述步骤S210中的人脸动作检测。
在步骤S310，在所述拍摄图像中定位人脸关键点。作为示例，在该步骤中，可以首先确定所获取的图像中是否包含人脸，在检测到人脸的情况下定位出人脸关键点。
人脸关键点是脸部一些表征能力强的关键点，例如眼睛、眼角、眼睛中心、眉毛、颧骨最高点、鼻子、鼻尖、鼻翼、嘴巴、嘴角、以及脸部外轮廓点等。
作为示例，可以预先搜集大量的人脸图像，例如N张人脸图像，例如，N＝10000，人工地在每张人脸图像中标注出预定的一系列人脸关键点，所述预定的一系列人脸关键点可以包括但不限于上述人脸关键点中的至少一部分。根据每张人脸图像中各人脸关键点附近的形状特征，基于参数形状模型，利用机器学习算法(如深度学习(Deep Learning)，或者基于局部特征的回归算法(local feature-based regression algorithm))进行人脸关键点模型训练，从而得到人脸关键点模型。
具体地，在步骤S310中可以基于已经建立的人脸关键点模型来在拍摄图 像中进行人脸检测和人脸关键点定位。例如，可以在拍摄图像中迭代地优化人脸关键点的位置，最后得到各人脸关键点的坐标位置。再例如，可以采用基于级联回归的方法在拍摄图像中定位人脸关键点。
人脸关键点的定位在人脸动作识别中起着重要的作用，然而应了解本公开不受具体采用的人脸关键点定位方法的限制。可以采用已有的人脸检测和人脸关键点定位算法来执行步骤S310中的人脸关键点定位。应了解，本公开实施例的活体检测方法100不限于利用已有的人脸检测和人脸关键点定位算法来进行人脸关键点定位，而且应涵盖利用将来开发的人脸检测和人脸关键点定位算法来进行人脸关键点定位。
在步骤S320，从所述拍摄图像中提取图像纹理信息。作为示例，可以根据所述拍摄图像中的像素信息，例如像素点的亮度信息，提取人脸的精细信息，例如眼球位置信息、口型信息、微表情信息等等。可以采用已有的图像纹理信息提取算法来执行步骤S320中的图像纹理信息提取。应了解，本公开实施例的活体检测方法100不限于利用已有的图像纹理信息提取算法来进行图像纹理信息提取，而且应涵盖利用将来开发的图像纹理信息提取算法来进行图像纹理信息提取。
应了解，步骤S310和S320可以择一执行，或者可以两者均执行。在步骤S310和S320两者均执行的情况下，它们可以同步执行，或者可以先后执行。
在步骤S330，基于所定位的人脸关键点以及/或者所述图像纹理信息，获得人脸动作属性的值。基于所定位的人脸关键点获得的所述人脸动作属性可以例如包括但不限于眼睛睁闭程度、嘴巴张闭程度、人脸俯仰程度、人脸偏转程度、人脸与摄像头的距离等。基于所述图像纹理信息获得的所述人脸动作属性可以包括但不限于眼球左右偏转程度、眼球上下偏转程度等等。
可选地，可以基于当前拍摄图像的前一拍摄图像以及当前拍摄图像，来获得人脸动作属性的值；或者，可以基于首个拍摄图像以及当前拍摄图像，来获得人脸动作属性的值；或者，可以基于当前拍摄图像以及当前拍摄图像的前几个拍摄图像，来获得人脸动作属性的值。
可选地，可以通过几何学习、机器学习、或图像处理的方式来基于所定位的人脸关键点获得人脸动作属性的值。例如，对于眼睛睁闭程度，可以在眼睛一圈定义多个关键点，例如8－20个关键点，例如，左眼的内眼角、外 眼角、上眼皮中心点和下眼皮中心点，以及右眼的内眼角、外眼角、上眼皮中心点和下眼皮中心点。然后，通过在拍摄图像上定位这些关键点，确定这些关键点在拍摄图像上的坐标，计算左眼(右眼)的上眼皮中心和下眼皮中心之间的距离作为左眼(右眼)上下眼皮距离，计算左眼(右眼)的内眼角和外眼角之间的距离作为左眼(右眼)内外眼角距离，计算左眼(或右眼)上下眼皮距离与左眼(或右眼)内外眼角距离的比值作为第一距离比值X，根据该第一距离比值来确定眼睛睁闭程度Y。例如，可以设定第一距离比值X的阈值Xmax，并且规定：Y＝X/Xmax，由此来确定眼睛睁闭程度Y。Y越大，则表示用户眼睛睁得越大。
返回图2，在步骤S220，根据所检测的人脸动作控制在显示屏幕上显示的虚拟对象中至少一部分的显示状态，其中，所述虚拟对象包括被控对象和障碍对象。
作为示例，所述虚拟对象可以包括第一组对象，所述第一组对象可以包括一个或多个对象。根据所检测的人脸动作更新所述第一组对象中至少一个对象在显示屏幕上的显示。所述第一组对象中至少一部分对象的初始显示位置和/或初始显示形态是预先确定的或随机确定的。具体地，例如可以改变所述至少一部分对象的运动状态、显示位置、尺寸大小、形状、颜色等。
参考图4来描述步骤S220的操作。
在步骤S410，根据所述人脸动作属性的值来更新所述虚拟对象中至少一部分的状态参量的值。
具体地，可以将一种人脸动作属性映射为虚拟对象的某一状态参量。例如，可以将用户眼睛睁闭程度或嘴巴张闭程度映射为虚拟对象的尺寸，并且根据用户眼睛睁闭程度或嘴巴张闭程度的值来更新虚拟对象的尺寸大小。再例如，可以将用户人脸俯仰程度映射为虚拟对象在显示屏幕上的垂直显示位置，并且根据用户人脸俯仰程度的值来更新虚拟对象在显示屏幕上的垂直显示位置。
可选地，可以计算当前拍摄图像中的嘴巴张闭程度和之前保存的首个拍摄图像中的嘴巴张闭程度的比值K1，并且将嘴巴张闭程度的比值K1映射为虚拟对象的尺寸S。具体地，可以采用一次函数S＝a*K1+b来实现映射。此外，可选地，可以计算当前拍摄图像中人脸位置偏离初始居中位置的程度K2，并且将人脸位置映射为虚拟对象的位置W。具体地，可以采用一次函数W＝ c*K2+d来实现映射。
例如，所述人脸动作属性可以包括至少一个动作属性，所述虚拟对象的状态参量包括至少一个状态参量。一个动作属性可以仅与一个状态参量对应，或者一个动作属性可以按照时间顺序依次与多个状态参量对应。
可选地，人脸动作属性与虚拟对象的状态参量之间的映射关系可以是预先设定的，或者可以是在开始执行根据本公开实施例的活体检测方法时随机确定的。根据本公开实施例的活体检测方法还可以包括：将所述人脸动作属性与虚拟对象的状态参量之间的映射关系提示给用户。
在步骤S420，按照更新后的所述虚拟对象的状态参量的值，在所述显示屏幕上显示所述虚拟对象。
如前所述，所述虚拟对象可以包括第一组对象，在根据本公开实施例的活体检测方法开始执行时将所述第一组对象显示在显示屏幕上，可以根据人脸动作属性来更新所述第一组对象中至少一个对象的显示。此外，所述虚拟对象还可以包括第二组对象，在根据本公开实施例的活体检测方法开始执行时所述第二组对象均未在显示屏幕上显示，可以根据所述第一组对象的显示情况来控制是否显示所述第二组对象中的至少一个对象。
具体地，所述第一组对象中至少一个对象的状态参量可以为显示位置、尺寸大小、形状、颜色、运动状态等，由此可以根据所述人脸动作属性的值改变所述第一组对象中至少一个对象的运动状态、显示位置、尺寸大小、形状、颜色等。
可选地，所述第二组对象中至少一个对象每个的状态参量至少可以包括可视状态，并且还可以包括显示位置、尺寸大小、形状、颜色、运动状态等。可以根据所述第一组对象中至少一个对象的显示情况来控制是否显示所述第二组对象中至少一个对象，即所述第二组对象中至少一个对象是否处于可视状态。
返回图2，在步骤S230，判断是否满足所述障碍对象的显示条件和/或所述被控对象的目标条件，并且判断所述被控对象与所述障碍对象是否一直不相遇。
可选地，所述障碍对象的显示条件为与所述障碍对象的总显示时间有关的条件、以及/或者与所述障碍对象的总数量有关的条件、以及/或者与所述障碍对象的显示状态有关的条件。可选地，所述被控对象的目标条件为与所述 被控对象的行进情况有关的条件、以及/或者与所述被控对象的形态或位置有关的条件。
例如，所述第一组对象包括第一对象和第二对象，所述第一对象是被控对象，所述第二对象是背景对象，所述背景对象是障碍对象，所述障碍对象的显示条件为所述障碍对象的总显示时间达到预定定时时间，所述被控对象的目标条件为所述第一对象与所述障碍对象一直不相遇。
例如，所述第一组对象还包括第三对象，所述第三对象是所述被控对象的目标对象，所述障碍对象的显示条件为所述障碍对象的总显示时间不超过预定定时时间，所述被控对象的目标条件为所述被控对象的显示位置与所述目标对象的显示位置重合，在尚未超出所述预定定时时间、所述被控对象的显示位置与所述目标对象的显示位置重合、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
例如，所述障碍对象的显示条件为所述障碍对象的总数量达到预定数量以及所述障碍对象均移出显示屏幕，其中，在所述障碍对象的总数量达到预定数量、所述障碍对象均移出显示屏幕、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
例如，所述障碍对象的显示条件为所述障碍对象的总显示时间不超过预定定时时间，所述被控对象的目标条件为所述被控对象的行进距离达到预定距离，其中，在所述障碍对象的总显示时间不超过预定定时时间、所述被控对象的行进距离达到预定距离、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
在满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且判断所述被控对象与所述障碍对象一直不相遇的情况下，在步骤S240确定所述拍摄图像中的人脸为活体人脸。反之，在不满足所述障碍对象的显示条件和/或所述被控对象的目标条件、或者判断所述被控对象与所述障碍对象相遇的情况下，在步骤S250确定所述拍摄图像中的人脸不是活体人脸。
根据本公开实施例的活体检测方法，通过将各种人脸动作参数作为虚拟对象的状态控制参量，根据人脸动作控制在显示屏幕上显示虚拟对象，可以根据所显示的虚拟对象是否满足预定条件来进行活体检测。
下面，参考具体实施例来进一步描述根据本公开实施例的活体检测方法。
第一实施例
在该第一实施例中，所述虚拟对象包括第一组对象，在开始执行根据本公开实施例的活体检测方法时将所述第一组对象显示在显示屏幕上，并且所述第一组对象包括一个或多个对象。根据所检测的人脸动作更新所述第一组对象中至少一个对象在显示屏幕上的显示，其中，所述第一组对象中的所述至少一个对象为被控对象。所述第一组对象中至少一部分对象的初始显示位置和/或初始显示形态是预先确定的或随机确定的。
在下面的示例中，所述第一组对象包括第一对象和第二对象，所述第一对象为被控对象，所述第二对象为背景对象，所述背景对象为障碍对象，所述第一对象和所述障碍对象的初始显示位置和/或初始显示形态是随机的。所述障碍对象可以是静止的，或者可以是运动的。在所述障碍对象运动的情况下，其运动轨迹可以为直线或曲线，并且所述障碍对象可以沿垂直方向移动、沿水平方向移动、或者沿任意方向移动。可选地，所述障碍对象的运动轨迹和运动方向也是随机的。
所述人脸动作属性包括第一动作属性，所述第一对象的状态参量包括所述第一对象的第一状态参量，所述第一对象的第一状态参量为所述第一对象的显示位置，所述第二对象的状态参量包括所述第二对象的第一状态参量，所述第二对象的第一状态参量为所述第二对象的显示位置，根据所述第一动作属性的值更新所述第一对象的第一状态参量的值，并且按照更新后的所述第一对象的第一状态参量的值在所述显示屏幕上显示所述第一对象。
所述预定条件可以为：所述第一对象与所述第二对象不相遇，或者所述第一对象的显示位置与所述第二对象的显示位置之间的距离超过预定距离，所述预定距离可以根据所述第一对象的显示尺寸和所述第二对象的显示尺寸确定。可选地，所述预定条件可以为：在预定定时时间内所述第一对象与所述第二对象不相遇，或者所述第一对象的显示位置与所述第二对象的显示位置之间的距离超过预定距离。具体地，所述障碍对象的显示条件为所述障碍对象的总显示时间达到预定定时时间，所述被控对象的目标条件为所述第一对象与所述障碍对象一直不相遇。
图5示出了根据本公开第一实施例的活体检测方法500的示例性流程图。在图5中，所述障碍对象的显示条件为所述障碍对象的总显示时间达到预定定时时间。
在步骤S510，初始化定时器。可以根据用户输入初始化定时器，或者可 以在拍摄图像中检测到人脸时自动初始化定时器，或者可以在拍摄图像中检测到人脸预定动作时自动初始化定时器。此外，在初始化定时器后，将所述第一组对象显示在显示屏幕上。
在步骤S520，实时地采集预定拍摄范围的图像(第一图像)作为拍摄图像。具体地，可以利用如图1所示的用于实现本公开实施例的人脸检测方法的电子设备100中的图像采集装置110或者独立于所述电子设备100的可以向所述电子设备100传送图像的其它图像采集装置，采集预定拍摄范围的灰度或彩色图像作为拍摄图像，所述拍摄图像可以是照片，也可以是视频中的一帧。
步骤S530－S540分别与图2中的步骤S210－S220对应，在此不再进行赘述。
在步骤S540，根据所检测的人脸动作控制在显示屏幕上显示的虚拟对象中至少一部分的显示状态，其中，所述虚拟对象包括被控对象和障碍对象。
在步骤S550判断所述障碍对象的总显示时间是否达到预定定时时间，所述预定定时时间可以是预先确定的，并且还判断所述被控对象与所述障碍对象是否一直不相遇。可选地，在所述定时器超出所述预定定时时间时可以产生超时标志，在步骤S550中可以根据该超时标志判断定时器是否超出所述预定定时时间。
根据步骤S550的判断结果，可以在步骤S560确定检测到活体人脸、或者在步骤S570确定没有检测到活体人脸、或者返回步骤S520。
在步骤S550确定所述障碍对象的总显示时间达到预定定时时间且所述被控对象与所述障碍对象一直不相遇的情况下，在步骤S560确定检测到活体人脸。
在步骤S550确定所述被控对象与所述障碍对象相遇的情况下，在步骤S570确定没有检测到活体人脸。
另一方面，在步骤S550确定所述障碍对象的总显示时间没有达到预定定时时间且所述被控对象与所述障碍对象一直不相遇的情况下，返回步骤S520。
在返回步骤S520的情况下，实时地采集所述预定拍摄范围的图像(第二图像)作为拍摄图像，并且接下来执行步骤S530－S550。这里，为区分先后采集的所述预定拍摄范围的图像，将先采集的图像称为第一图像，将后采集的图像称为第二图像。应了解，第一图像和第二图像是相同取景范围内的图 像，仅仅是采集的时间不同。
如图5所示的步骤S520－S550重复执行，直至根据步骤S550的判断结果确定所述障碍对象的总显示时间达到预定定时时间且所述被控对象与所述障碍对象一直不相遇，从而在步骤S560确定检测到活体人脸；或者直至根据步骤S550的判断结果确定所述被控对象与所述障碍对象相遇，从而在步骤S570确定没有检测到活体人脸。
尽管在图5中在步骤S550中进行定时器是否超出预定定时时间的判断，应了解本发明不限于此，可以在根据本公开实施例的活体检测方法的任一步骤中执行该判断。此外，可选地，在所述定时器超出预定定时时间的情况下产生超时标志，该超时标志可以直接触发根据本公开实施例的活体检测方法的步骤S560或S570，即确定是否检测到活体人脸。
在图6A中示出了第一对象A以及障碍对象D的位置示例。所述障碍对象D可以在显示屏幕上不断移动，并且所述障碍对象D的移动方向可以是随机的。可选地，在所述障碍对象D移出显示屏幕之前所述第一对象A与所述障碍对象D一直不相遇的情况下，确定活体检测成功。
在应用图5所示的活体检测方法的情况下，在步骤S550确定所述定时器超出所述预定定时时间并且所述第一对象一直不与所述障碍对象相遇的情况下，在步骤S560确定检测到活体人脸；在步骤S550确定所述定时器没有超出所述预定定时时间并且所述第一对象一直不与所述障碍对象相遇的情况下，返回到步骤S520；另一方面，在步骤S550确定所述定时器没有超出所述预定定时时间并且所述第一对象与所述障碍对象相遇的情况下，在步骤S570确定没有检测到活体人脸。
可选地，所述第一组对象还包括第三对象，所述第一对象为被控对象，所述第二对象和第三对象构成背景对象，所述第二对象为障碍对象，所述第三对象是目标对象，所述障碍对象是随机显示的或随机产生的。所述第三对象的状态参量可以包括所述第三对象的第一状态参量，所述第三对象的第一状态参量为所述第三对象的显示位置。
所述预定条件可以为：所述第一对象与所述第二对象不相遇且所述第一对象与所述第三对象重合，或者所述第一对象的显示位置与所述第二对象的显示位置之间的距离超过预定距离且所述第一对象与所述第三对象重合，所述预定距离可以根据所述第一对象的显示尺寸和所述第二对象的显示尺寸确 定。可选地，所述预定条件可以为：在预定定时时间内所述第一对象与所述第三对象重合且所述第一对象与所述第二对象不相遇，或者在预定定时时间内所述第一对象与所述第三对象重合且所述第一对象的显示位置与所述第二对象的显示位置之间的距离一直超过预定距离。
在图6B中示出了第一对象A、第二对象(障碍对象)D以及第三对象(目标对象)B。所述障碍对象D可以在显示屏幕上不断移动，并且所述障碍对象D的移动方向可以是随机的，在所述第一对象A与所述障碍对象D不相遇且所述第一对象A与所述目标对象B重合的情况下，确定活体检测成功。优选地，在预定定时时间内所述第一对象A与所述障碍对象D不相遇且所述第一对象A的显示位置与所述目标对象B的显示位置重合的情况下，确定活体检测成功。
在应用图5所示的活体检测方法的情况下，在步骤S550判断所述定时器是否超出所述预定定时时间，判断所述被控对象的显示位置是否与所述目标对象的显示位置重合，并且判断所述被控对象是否与所述障碍对象一直不相遇。
具体地，在步骤S550确定所述定时器超出所述预定定时时间的情况下，在步骤S570确定没有检测到活体人脸；在步骤S550确定所述定时器没有超出所述预定定时时间、所述第一对象与所述目标对象重合、且所述被控对象与所述障碍对象一直不相遇的情况下，在步骤S560确定检测到活体人脸；另一方面，在步骤S550确定所述定时器没有超出所述预定定时时间、所述第一对象未与所述目标对象重合、并且所述第一对象与所述障碍对象一直不相遇的情况下，返回到步骤S520。
可选地，如图6A和6B所示，所述第一对象A和所述第二对象B的水平位置和垂直位置均不同，在此情况下，所述第一动作属性可以包括第一子动作属性和第二子动作属性，所述第一对象A的第一状态参量可以包括第一子状态参量和第二子状态参量，所述第一子状态参量的值为所述第一对象A的水平位置坐标，所述第二子状态参量的值为所述第一对象A的垂直位置坐标，可以根据所述第一子动作属性的值来更新所述第一对象A在所述显示屏幕上的水平位置坐标，并且根据所述第二子动作属性的值来更新所述第一对象A在所述显示屏幕上的垂直位置坐标。
例如，可以将所述第一动作属性定义为所述人脸在拍摄图像中的位置， 并且根据人脸在拍摄图像中的位置坐标来更新所述第一对象A在所述显示屏幕上的显示位置。在此情况下，可以将所述第一子动作属性定义为人脸在拍摄图像中的水平位置并且将所述第二子动作属性定义为人脸在拍摄图像中的垂直位置，可以根据人脸在拍摄图像中的水平位置坐标来更新所述第一对象A在所述显示屏幕上的水平位置坐标，并且根据人脸在拍摄图像中的垂直位置坐标来更新所述第一对象A在所述显示屏幕上的垂直位置坐标。
再例如，可以将所述第一子动作属性定义为人脸偏转程度并且可以将所述第二子动作属性定义为人脸俯仰程度，然后可以根据人脸偏转程度的值来更新所述第一对象A在所述显示屏幕上的水平位置坐标，并且根据人脸俯仰程度的值来更新所述第一对象A在所述显示屏幕上的垂直位置坐标。
第二实施例
在该第二实施例中，所述虚拟对象包括第一组对象和第二组对象，在开始执行根据本公开实施例的活体检测方法时将所述第一组对象显示在显示屏幕上，并且所述第一组对象包括一个或多个对象，在开始执行根据本公开实施例的活体检测方法时所述第二组对象尚未显示在显示屏幕上并且包括一个或多个对象。
根据所检测的人脸动作更新所述第一组对象中至少一个对象在显示屏幕上的显示，其中，所述第一组对象中的所述至少一个对象为被控对象。所述第一组对象中至少一部分对象的初始显示位置和/或初始显示形态是预先确定的或随机确定的。
根据所述第一组对象中至少一个对象的显示情况显示所述第二组对象中至少一个对象。可选地，还根据所述第二组对象中至少一个对象的显示情况显示所述第二组对象中其它至少一个对象。可选地，可以随机地显示所述第二组对象中的至少一个对象。所述第二组对象中至少一部分对象的初始显示位置和/或初始显示形态是预先确定的或随机确定的。
在该实施例中，所述第一组对象中每个对象的第一状态参量为该对象的显示位置，并且所述第二组对象中每个对象的第一和第二状态参量分别为该对象的显示位置和可视状态。
在本实施例中，所述第一组对象包括第一对象和第二对象，所述第二组对象包括多个对象，所述第一对象为被控对象，所述第二对象以及所述第二 组对象为背景对象，所述背景对象为障碍对象，所述第一对象和所述障碍对象的初始显示位置和/或初始显示形态是随机的。在所述障碍对象运动的情况下，其运动轨迹可以为直线或曲线，并且所述障碍对象可以沿垂直方向移动、沿水平方向移动、或者沿任意方向移动。可选地，所述障碍对象的运动轨迹和运动方向也是随机的。
所述人脸动作属性包括第一动作属性，所述第一对象的状态参量包括所述第一对象的第一状态参量，所述第一对象的第一状态参量为所述第一对象的显示位置，根据所述第一动作属性的值更新所述第一对象的第一状态参量的值，并且按照更新后的所述第一对象的第一状态参量的值在所述显示屏幕上显示所述第一对象。
可选地，所述预定条件可以为：在预定时间内所述第一对象与所述障碍对象均不相遇、所述第一对象与预定数量的障碍对象不相遇、或在预定时间内所述第一对象与预定数量的障碍对象不相遇。
图7示出了根据本公开第二实施例的活体检测方法700的示例性流程图。在图7中，所述障碍对象的显示条件为所述障碍对象的总数量达到预定数量以及所述障碍对象均移出显示屏幕。
在步骤S710，实时地采集预定拍摄范围的图像(第一图像)作为拍摄图像。具体地，可以利用如图1所示的用于实现本公开实施例的人脸检测方法的电子设备100中的图像采集装置110或者独立于所述电子设备100的可以向所述电子设备100传送图像的其它图像采集装置，采集预定拍摄范围的灰度或彩色图像作为拍摄图像，所述拍摄图像可以是照片，也可以是视频中的一帧。
步骤S720－S730分别与图5中的步骤S530－S540对应，在此不再进行赘述。
在步骤S740判断所述被控对象与所述障碍对象是否相遇。在步骤S740确定所述被控对象与所述障碍对象相遇的情况下，在步骤S790确定没有检测到活体人脸。
在步骤S740确定所述被控对象与所述障碍对象一直不相遇的情况下，在步骤S750判断所述障碍对象的至少一部分是否移出显示屏幕。在步骤S750确定没有障碍对象移出显示屏幕的情况下，返回步骤S710。
在步骤S750确定所述障碍对象的至少一部分移出显示屏幕的情况下，在 步骤S760判断是否已经显示了预定数量的障碍对象。
在步骤S760确定已经显示了预定数量的障碍对象的情况下，在步骤S765判断是否所述障碍对象均移出显示屏幕，并且在步骤S765确定所述障碍对象均移出显示屏幕的情况下，在步骤S780确定检测到活体人脸；在步骤S765确定所述障碍对象尚未全部移出显示屏幕的情况下，返回步骤S710。
在步骤S760确定尚未显示预定数量的障碍对象的情况下，在步骤S770显示所述第二组对象中的至少一个对象，所述第二组对象中的至少一个对象和所述障碍对象中尚未移出显示屏幕的对象一起作为新的障碍对象，然后返回步骤S710。
此外，根据需要可以将图5和图7所述的活体检测方法组合，在此情况下，所述预定条件可以为：在预定定时时间内所述第一对象与预定数量的障碍对象不相遇。
在返回步骤S710的情况下，实时地采集所述预定拍摄范围的图像(第二图像)作为拍摄图像，并且接下来执行步骤S720－S740。这里，为区分先后采集的所述预定拍摄范围的图像，将先采集的图像称为第一图像，将后采集的图像称为第二图像。应了解，第一图像和第二图像是相同取景范围内的图像，仅仅是采集的时间不同。
如图7所示的步骤S710－S740重复执行，直至在步骤S780确定检测到活体人脸；或者直至在步骤S790确定没有检测到活体人脸。
在图8A中示出了第一对象A以及障碍对象D的位置示例。所述障碍对象D可以在显示屏幕上不断移动，并且所述障碍对象D的移动方向可以是随机的。
在所述障碍对象D移动出所述显示屏幕时，显示所述第二组对象中的障碍对象D2，而在所述障碍对象D2移出所述显示屏幕时，显示所述第二组对象中的障碍对象D3。依此类推，直至达到预定定时时间，或者显示了预定数量的障碍对象。
可选地，在预定定时时间内所述第一对象A与所述障碍对象一直不相遇的情况下，确定活体检测成功。替换地，所述第一对象A与预定数量的障碍对象不相遇的情况下，确定活体检测成功。替换地，在预定定时时间内所述第一对象A与预定数量的障碍对象不相遇的情况下，确定活体检测成功。
可选地，所述第一组对象还包括第三对象，所述第二对象和第三对象构 成背景对象，所述第三对象是目标对象。所述预定条件可以为：在预定定时时间内所述第一对象与所述障碍对象一直不相遇且所述第一对象与所述第三对象重合。
在图8B中示出了第一组对象中的第一对象A、第二对象(障碍对象)D以及第三对象(目标对象)B、以及第二组对象中的障碍对象D1和D2。所述障碍对象可以在显示屏幕上不断移动，并且所述障碍对象D的移动方向可以是随机的，在所述第一对象A与所述障碍对象均不相遇且所述第一对象A与所述目标对象B重合的情况下，确定活体检测成功。优选地，在预定定时时间内所述第一对象A与所述障碍对象均不相遇且所述第一对象A的显示位置与所述目标对象B的显示位置重合的情况下，确定活体检测成功。
具体地，如图7所示，在步骤S740可以确定所述第一对象A与当前显示的障碍对象不相遇，在步骤S750可以确定当前显示的障碍对象移出显示屏幕，在步骤S760可以确定已经显示的障碍对象的数量尚未达到预定数量，因此在步骤S770在显示屏幕上显示新的障碍对象，并且返回步骤S710。
另一方面，在步骤S740可以确定所述第一对象A与当前显示的障碍对象不相遇，在步骤S750可以确定当前显示的障碍对象移出显示屏幕，在步骤S760可以确定已经显示的障碍对象的数量已经达到预定数量，在步骤S765可以确定所有障碍对象均已移出显示屏幕，因此在步骤S780确定检测到活体人脸。
第三实施例
在该第三实施例中，所述虚拟对象包括第一组对象和第二组对象，在开始执行根据本公开实施例的活体检测方法时将所述第一组对象显示在显示屏幕上，并且所述第一组对象包括一个或多个对象，在开始执行根据本公开实施例的活体检测方法时所述第二组对象尚未显示在显示屏幕上并且包括一个或多个对象。
根据所检测的人脸动作更新所述第一组对象中至少一个对象在显示屏幕上的显示，其中，所述第一组对象中的所述至少一个对象为被控对象。所述第一组对象中至少一部分对象的初始显示位置和/或初始显示形态是预先确定的或随机确定的。
根据所述第一组对象中至少一个对象的显示情况显示所述第二组对象中 至少一个对象。可选地，还根据所述第二组对象中至少一个对象的显示情况显示所述第二组对象中其它至少一个对象。可选地，可以随机地显示所述第二组对象中的至少一个对象。所述第二组对象中至少一部分对象的初始显示位置和/或初始显示形态是预先确定的或随机确定的。所述第二组对象中对象是非被控对象，即背景对象，所述背景对象为障碍对象。
在该实施例中，所述第一组对象中每个对象的第一状态参量为该对象的显示位置，并且所述第二组对象中每个对象的第一和第二状态参量分别为该对象的显示位置和可视状态。
可选地，所述预定条件可以为：在预定时间内所述第一对象与所述障碍对象均不相遇且所述被控对象的行进距离达到预定距离、所述第一对象与预定数量的障碍对象不相遇且所述被控对象的行进距离达到预定距离、或在预定时间内所述第一对象与预定数量的障碍对象不相遇且所述被控对象的行进距离达到预定距离。
图9示出了根据本公开第三实施例的活体检测方法900的示例性流程图。在图9中，所述障碍对象的显示条件为所述障碍对象的总显示时间不超过预定定时时间，所述被控对象的目标条件为所述被控对象的行进距离达到预定距离。具体地，在所述障碍对象的总显示时间不超过预定定时时间、所述被控对象的行进距离达到预定距离、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
在步骤S910，初始化定时器。可以根据用户输入初始化定时器，或者可以在拍摄图像中检测到人脸时自动初始化定时器，或者可以在拍摄图像中检测到人脸预定动作时自动初始化定时器。此外，在初始化定时器后，将所述第一组对象显示在显示屏幕上。
在步骤S920，实时地采集预定拍摄范围的图像(第一图像)作为拍摄图像。具体地，可以利用如图1所示的用于实现本公开实施例的人脸检测方法的电子设备100中的图像采集装置110或者独立于所述电子设备100的可以向所述电子设备100传送图像的其它图像采集装置，采集预定拍摄范围的灰度或彩色图像作为拍摄图像，所述拍摄图像可以是照片，也可以是视频中的一帧。
步骤S930－S940分别与图5中的步骤S530－S540对应，在此不再进行赘述。
在步骤S950判断所述被控对象与所述障碍对象是否相遇。在步骤S950确定所述被控对象与所述障碍对象相遇的情况下，在步骤960确定没有检测到活体人脸。
在步骤S950确定所述被控对象与所述障碍对象一直不相遇的情况下，在步骤S970判断所述障碍对象的总显示时间是否达到预定定时时间，所述预定定时时间可以是预先确定的，并且判断所述被控对象的行进距离是否达到预定距离，所述预定距离可以是预先确定的。
可选地，在所述定时器超出所述预定定时时间时可以产生超时标志，在步骤S970中可以根据该超时标志判断定时器是否超出所述预定定时时间。
根据步骤S970的判断结果，可以在步骤S980确定检测到活体人脸、或者在步骤S960确定没有检测到活体人脸、或者返回步骤S920。
在步骤S970确定所述障碍对象的总显示时间达到预定定时时间且所述被控对象的行进距离没有达到预定距离的情况下，在步骤S980确定没有检测到活体人脸。
在步骤S970确定所述障碍对象的总显示时间没有达到或刚达到预定定时时间且所述被控对象的行进距离达到预定距离的情况下，在步骤S960确定检测到活体人脸。
另一方面，在步骤S970确定所述障碍对象的总显示时间没有达到预定定时时间且所述被控对象的行进距离没有达到预定距离的情况下，返回步骤S920。
在返回步骤S920的情况下，实时地采集所述预定拍摄范围的图像(第二图像)作为拍摄图像，并且接下来执行步骤S930－S950。这里，为区分先后采集的所述预定拍摄范围的图像，将先采集的图像称为第一图像，将后采集的图像称为第二图像。应了解，第一图像和第二图像是相同取景范围内的图像，仅仅是采集的时间不同。
如图9所示的步骤S920－S950重复执行，直至在步骤S980确定检测到活体人脸；或者在步骤S960确定没有检测到活体人脸。
此外，根据需要可以将图7和图9所述的活体检测方法组合，在此情况下，所述预定条件可以为：在预定定时时间内所述第一对象与预定数量的障碍对象不相遇且所述第一对象的行进距离超出预定距离。
具体地，所述第一组对象包括第一对象和第二对象，在步骤S940可以根 据所检测的人脸动作更新所述第一对象和第二对象在显示屏幕上的显示。
在图10中示出了第一对象A和第二对象D的位置示例。具体地，所述第一对象的垂直显示位置固定，根据所检测的人脸动作更新所述第一对象的水平显示位置以及所述第二对象的水平和垂直显示位置。
此外，还根据所述第二对象的显示情况来显示所述第二组对象中的障碍对象，并且还可以根据第二组对象中障碍对象的显示情况来显示所述第二组对象中新的障碍对象。可选地，还可以随机地显示第二组对象中的障碍对象。具体地，根据所检测的人脸动作更新所述第一对象的水平显示位置以及所述第一和二组对象中障碍对象的水平和垂直显示位置。
可选地，任一时刻在显示屏幕上显示固定数量的障碍对象，在任一障碍对象消失时，显示新障碍对象，使得在显示屏幕上保持存在固定数量的障碍对象。
所述人脸动作属性可以包括第一动作属性和第二动作属性，所述第一对象的状态参量包括所述第一对象的第一和第二状态参量，所述第一对象的第一和第二状态参量分别为所述第一对象的行进参量和水平位置，所述行进参量可以为运动速度、行进距离等。
例如，在所述行进参量为运动速度的情况下，首先，根据所述第一动作属性的值更新第一对象的运动速度的值，并且根据所述第二动作属性的值更新第一对象的水平位置坐标。其次，根据所述第一对象A的运动速度的值、所述第一对象A与所述障碍对象D之间的距离(可以包括水平距离和垂直距离)、以及所述第一对象A的水平位置坐标，确定所述障碍对象D和所述第一对象A的显示位置。例如，在所述第一对象的目标前进方向为道路延伸方向(如图10中道路变窄的方向)、以及所述第一对象A的垂直显示位置保持不变的情况下，可以根据所述第一对象A的运动速度的值以及所述第一对象A与所述障碍对象D之间的垂直距离，确定是否继续显示所述障碍对象D、以及所述障碍对象D的显示位置，并且可以根据所述第一对象A的水平位置坐标确定所述第一对象A的显示位置。
具体地，例如，所述第一对象A可以为汽车，所述障碍对象D可以是在汽车前进的道路上随机产生的石头，所述第一动作属性可以为人脸俯仰程度，所述第二动作属性可以为人脸偏转程度，所述第一对象A的第一状态参量和第二状态参量可以分别为所述第一对象的运动速度和水平位置。例如，可以 将人脸平视状态对应于运动速度V0，将人脸30度或45度仰视状态对应于最高运动速度VH，将人脸30度或45度俯视状态对应于最低运动速度VL，根据人脸俯仰程度的值(例如，人脸俯仰角度)确定第一对象的运动速度。例如，可以将人脸正视状态对应于中间位置P0，将人脸30度或45度左偏状态对应于左侧边缘位置PL，将人脸30度或45度右偏状态对应于右侧边缘位置PR，根据人脸偏转程度的值(例如，人脸偏转角度)确定第一对象的水平位置坐标。
此外，所述第一对象的状态参量还包括所述第一对象的第三状态参量，所述第三状态参量可以为所述第一对象的行进距离。在此情况下，在所述第一对象与障碍对象不相遇并且所述第一对象在预定时间内的行进距离达到预设距离值的情况下，确定活体检测成功。
上面已经在第一到第三实施例中描述了根据本公开实施例的活体检测方法的具体实现方式，应了解，可以根据需要组合第一到第三实施例中的各种具体操作。
接下来，将参考图11和图12来描述根据本公开实施例的活体检测设备。所述活体检测设备可以是集成了人脸图像采集装置的电子设备，诸如智能手机、平板电脑、个人计算机、基于人脸识别的身份识别设备等。替代地，所述活体检测设备还可以包括分离的人脸图像采集装置和检测处理装置，所述检测处理装置可以从所述人脸图像采集装置接收拍摄图像，并且依据所接收的拍摄图像进行活体检测。所述检测处理装置可以为服务器、智能手机、平板电脑、个人计算机、基于人脸识别的身份识别设备等。
由于该活体检测设备执行各个操作的细节与上文中针对图2－4描述的活体检测方法的细节基本相同，因此为了避免重复，在下文中仅对所述活体检测设备进行简要的描述，而省略对相同细节的描述。
如图11所示，根据本公开实施例的活体检测设备1100包括人脸动作检测装置1110、虚拟对象控制装置1120、以及活体判断装置1130。人脸动作检测装置1110、虚拟对象控制装置1120、以及活体判断装置1130可以由图1所示的处理器102实现。
如图12所示，根据本公开实施例的活体检测设备1200包括图像采集装置1240、人脸动作检测装置1110、虚拟对象控制装置1120、活体判断装置1130、显示装置1250以及存储装置1260。图像采集装置1240可以由图1所 示的图像采集装置110实现，人脸动作检测装置1110、虚拟对象控制装置1120、以及活体判断装置1130可以由图1所示的处理器102实现，显示装置1250可以由图1所示的输出装置108实现，存储装置1260可以由图1所示的存储装置104实现。
可以利用活体检测设备1200中的图像采集装置1240或者独立于所述活体检测设备1100或1200的可以向所述活体检测设备1100或1200传送图像的其它图像采集装置，采集预定拍摄范围的灰度或彩色图像作为拍摄图像，所述拍摄图像可以是照片，也可以是视频中的一帧。所述图像采集设备可以是智能电话的摄像头、平板电脑的摄像头、个人计算机的摄像头、或者甚至可以是网络摄像头。
人脸动作检测装置1110被配置为从拍摄图像中检测人脸动作。
如图13所示，人脸动作检测装置1110可以包括关键点定位装置1310、纹理信息提取装置1320、以及动作属性确定装置1330。
所述关键点定位装置1310被配置为在所述拍摄图像中定位人脸关键点。作为示例，所述关键点定位装置1310可以首先确定所获取的图像中是否包含人脸，在检测到人脸的情况下定位出人脸关键点。所述关键点定位装置1310操作的细节与步骤S310中描述的细节相同，在此不再赘述。
所述纹理信息提取装置1320被配置为从所述拍摄图像中提取图像纹理信息。作为示例，所述纹理信息提取装置1320可以根据所述拍摄图像中的像素信息，例如像素点的亮度信息，提取人脸的精细信息，例如眼球位置信息、口型信息、微表情信息等等。
所述动作属性确定装置1330基于所定位的人脸关键点以及/或者所述图像纹理信息，获得人脸动作属性的值。基于所定位的人脸关键点获得的所述人脸动作属性可以例如包括但不限于眼睛睁闭程度、嘴巴张闭程度、人脸俯仰程度、人脸偏转程度、人脸与摄像头的距离等。基于所述图像纹理信息获得的所述人脸动作属性可以包括但不限于眼球左右偏转程度、眼球上下偏转程度等等。所述动作属性确定装置1330操作的细节与步骤S330中描述的细节相同，在此不再赘述。
所述虚拟对象控制装置1120被配置为根据所检测的人脸动作控制在所述显示装置1250上显示的虚拟对象中至少一部分的显示状态，其中，所述虚拟对象包括被控对象和障碍对象。
作为示例，所述虚拟对象可以包括第一组对象，所述第一组对象可以包括一个或多个对象。根据所检测的人脸动作更新所述第一组对象中至少一个对象在显示屏幕上的显示。所述第一组对象中至少一部分对象的初始显示位置和/或初始显示形态是预先确定的或随机确定的。具体地，例如可以改变所述至少一部分对象的运动状态、显示位置、尺寸大小、形状、颜色等。
如图14所示，所述虚拟对象控制装置1120可以包括人脸动作映射装置1410、以及虚拟对象呈现装置1420。
所述人脸动作映射装置1410根据所述人脸动作属性的值来更新所述虚拟对象的状态参量的值。
具体地，可以将一种人脸动作属性映射为虚拟对象的某一状态参量。例如，可以将用户眼睛睁闭程度或嘴巴张闭程度映射为虚拟对象的尺寸，并且根据用户眼睛睁闭程度或嘴巴张闭程度的值来更新虚拟对象的尺寸大小。再例如，可以将用户人脸俯仰程度映射为虚拟对象在显示屏幕上的垂直显示位置，并且根据用户人脸俯仰程度的值来更新虚拟对象在显示屏幕上的垂直显示位置。可选地，人脸动作属性与虚拟对象的状态参量之间的映射关系可以是预先设定的。
例如，所述人脸动作属性可以包括至少一个动作属性，所述虚拟对象的状态参量包括至少一个状态参量。一个运动属性可以仅与一个状态参量对应，或者一个运动属性可以按照时间顺序依次与多个状态参量对应。
所述虚拟对象呈现装置1420按照更新后的所述虚拟对象的状态参量的值呈现所述虚拟对象。
具体地，所述虚拟对象呈现装置1420可以更新第一组对象中至少一个对象的显示。有利地，所述虚拟对象呈现装置1420还可以显示新的虚拟对象，即第二组对象中的虚拟对象。有利地，所述虚拟对象呈现装置1420还可以更新第二组对象中至少一个对象的显示。
所述活体判断装置1130被配置为判断判断是否满足所述障碍对象的显示条件和/或所述被控对象的目标条件，并且判断所述被控对象与所述障碍对象是否一直不相遇。
可选地，所述障碍对象的显示条件为与所述障碍对象的总显示时间有关的条件、以及/或者与所述障碍对象的总数量有关的条件、以及/或者与所述障碍对象的显示状态有关的条件。可选地，所述被控对象的目标条件为与所述 被控对象的行进情况有关的条件、以及/或者与所述被控对象的形态或位置有关的条件。
例如，所述第一组对象包括第一对象和第二对象，所述第一对象是被控对象，所述第二对象是背景对象，所述背景对象是障碍对象，所述障碍对象的显示条件为所述障碍对象的总显示时间达到预定定时时间，所述被控对象的目标条件为所述第一对象与所述障碍对象一直不相遇。
例如，所述第一组对象还包括第三对象，所述第三对象是所述被控对象的目标对象，所述障碍对象的显示条件为所述障碍对象的总显示时间不超过预定定时时间，所述被控对象的目标条件为所述被控对象的显示位置与所述目标对象的显示位置重合，在尚未超出所述预定定时时间、所述被控对象的显示位置与所述目标对象的显示位置重合、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
例如，所述障碍对象的显示条件为所述障碍对象的总数量达到预定数量以及所述障碍对象均移出显示屏幕，其中，在所述障碍对象的总数量达到预定数量、所述障碍对象均移出显示屏幕、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
例如，所述障碍对象的显示条件为所述障碍对象的总显示时间不超过预定定时时间，所述被控对象的目标条件为所述被控对象的行进距离达到预定距离，其中，在所述障碍对象的总显示时间不超过预定定时时间、所述被控对象的行进距离达到预定距离、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
所述人脸动作映射装置1410以及所述虚拟对象呈现装置1420可以执行上述第一到第三实施例中的各种操作，在此不再赘述。
此外，根据本公开实施例的活体检测装置1100和1200还可以包括定时器，用于对预定定时时间进行计时。所述定时器也可以由处理器102实现。可以根据用户输入初始化定时器，或者可以在拍摄图像中检测到人脸时自动初始化定时器，或者可以在拍摄图像中检测到人脸预定动作时自动初始化定时器。在此情况下，所述活体判断装置1130被配置为基于所述定时器判断是否满足所述障碍对象的显示条件。
所述存储装置1260用于存储所述拍摄图像。此外，所述存储装置1260还用于存储所述虚拟对象的状态参量及状态参量值。此外，所述存储装置1260 还用于存储所述虚拟对象呈现装置1420所呈现的虚拟对象并且存储要在显示装置1250上显示的背景图像等。
此外，所述存储装置1260可以存储计算机程序指令，所述计算机程序指令在被所述处理器102运行时可以实现根据本公开实施例的活体检测方法，并且/或者可以实现根据本公开实施例的活体检测设备中的关键点定位装置1310、纹理信息提取装置1320、以及动作属性确定装置1330。
此外，根据本公开实施例，还提供了一种计算机程序产品，其包括计算机可读存储介质，在所述计算机可读存储介质上存储了计算机程序指令。所述计算机程序指令在被计算机运行时可以实现根据本公开实施例的活体检测方法，并且/或者可以实现根据本公开实施例的活体检测设备中的关键点定位装置、纹理信息提取装置、以及动作属性确定装置的全部或部分功能。
根据本公开实施例的活体检测方法及设备、以及计算机程序产品，通过基于人脸动作控制虚拟对象显示并根据虚拟对象显示进行活体检测，可以不依赖于特殊的硬件设备来有效地防范照片、视频、3D人脸模型或者面具等多种方式的攻击，从而可以降低活体检测的成本。更进一步，通过识别人脸动作中的多个动作属性，可以控制虚拟对象的多个状态参量，可以使得所述虚拟对象在多个方面改变显示状态，例如使得所述虚拟对象执行复杂的预定动作、或者使得所述虚拟对象实现与初始显示效果有很大不同的显示效果。因此，可以进一步提高活体检测的准确度，并且进而可以提高应用根据本发明实施例的活体检测方法及设备、以及计算机程序产品的应用场景的安全性。
所述计算机可读存储介质可以是一个或多个计算机可读存储介质的任意组合。所述计算机可读存储介质例如可以包括智能电话的存储卡、平板电脑的存储部件、个人计算机的硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦除可编程只读存储器(EPROM)、便携式紧致盘只读存储器(CD-ROM)、USB存储器、或者上述存储介质的任意组合。
在上面详细描述的本发明的示例实施例仅仅是说明性的，而不是限制性的。本领域技术人员应该理解，在不脱离本发明的原理和精神的情况下，可对这些实施例进行各种修改，组合或子组合，并且这样的修改应落入本发明的范围内。

Claims (20)

  1. 一种活体检测方法，包括：
    从拍摄图像中检测人脸动作；
    根据所检测的人脸动作控制在显示屏幕上显示的虚拟对象中至少一部分的显示状态，其中，所述虚拟对象包括被控对象和障碍对象；
    在满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
  2. 如权利要求1所述的活体检测方法，还包括：
    实时地采集预定拍摄范围的第一图像作为所述拍摄图像；
    其中，所述活体检测方法还包括：在尚未满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且所述被控对象与所述障碍对象一直不相遇的情况下，实时地采集所述预定拍摄范围的第二图像作为所述拍摄图像。
  3. 如权利要求1所述的活体检测方法，其中，
    所述障碍对象的显示条件为与所述障碍对象的总显示时间有关的条件、以及/或者与所述障碍对象的总数量有关的条件、以及/或者与所述障碍对象的显示状态有关的条件，
    所述被控对象的目标条件为与所述被控对象的行进情况有关的条件、以及/或者与所述被控对象的形态或位置有关的条件。
  4. 如权利要求3所述的活体检测方法，其中，所述障碍对象包括一个或多个对象，所述障碍对象的所述一个或多个对象在所述显示屏幕上移动，
    所述活体检测方法还包括：
    在尚未满足所述障碍对象的显示条件的情况下，在所述障碍对象中至少一部分对象移出显示屏幕时，显示新的至少一个对象，其中，所述新的至少一个对象和所述障碍对象中尚未移出显示屏幕的对象一起作为所述障碍对象，所述新的至少一个对象的显示位置是随机确定的。
  5. 如权利要求3所述的活体检测方法，其中，所述虚拟对象还包括所述被控对象的目标对象，
    所述障碍对象的显示条件为所述障碍对象的总显示时间不超过预定定时时间，所述被控对象的目标条件为所述被控对象的显示位置与所述目标对象 的显示位置重合，
    其中，在尚未超出所述预定定时时间、所述被控对象的显示位置与所述目标对象的显示位置重合、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
  6. 如权利要求3所述的活体检测方法，其中，所述障碍对象的显示条件为所述障碍对象的总数量达到预定数量以及所述障碍对象均移出显示屏幕，
    其中，在所述障碍对象的总数量达到预定数量、所述障碍对象均移出显示屏幕、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
  7. 如权利要求3所述的活体检测方法，其中，
    所述障碍对象的显示条件为所述障碍对象的总显示时间不超过预定定时时间，所述被控对象的目标条件为所述被控对象的行进距离达到预定距离，
    其中，在所述障碍对象的总显示时间不超过预定定时时间、所述被控对象的行进距离达到预定距离、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
  8. 如权利要求1所述的活体检测方法，其中，从拍摄图像中检测人脸动作包括：
    在所述拍摄图像中定位人脸关键点，以及/或者从所述拍摄图像中提取图像纹理信息；以及
    基于所定位的人脸关键点和/或所提取的图像纹理信息，获得人脸动作属性的值。
  9. 如权利要求8所述的活体检测方法，其中，根据所检测的人脸动作控制在显示屏幕上显示的虚拟对象中至少一部分的显示状态包括：
    根据所检测的人脸动作的人脸动作属性的值来更新所述虚拟对象中至少一部分的状态参量的值；以及
    按照更新后的所述虚拟对象的状态参量的值，在所述显示屏幕上显示所述虚拟对象。
  10. 如权利要求8或9所述的活体检测方法，其中，所述人脸动作属性包括以下至少一项：眼睛睁闭程度、嘴巴张闭程度、人脸俯仰程度、人脸偏转程度、人脸与摄像头的距离、眼球左右转动程度、眼球上下转动程度。
  11. 一种活体检测设备，包括：
    一个或多个处理器；
    一个或多个存储器；以及
    存储在所述存储器中的计算机程序指令，在所述计算机程序指令被所述处理器运行时执行以下步骤：
    从拍摄图像中检测人脸动作；
    根据所检测的人脸动作控制在显示装置上显示的虚拟对象中至少一部分的显示状态，其中，所述虚拟对象包括被控对象和障碍对象；以及
    在满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
  12. 如权利要求11所述的活体检测设备，还包括：
    图像采集装置，用于实时地采集预定拍摄范围的第一图像作为所述拍摄图像；以及
    所述显示装置，
    其中，在尚未满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且所述被控对象与所述障碍对象一直不相遇的情况下，实时地采集所述预定拍摄范围的第二图像作为所述拍摄图像。
  13. 如权利要求11所述的活体检测设备，其中，
    所述障碍对象的显示条件为与所述障碍对象的总显示时间有关的条件、以及/或者与所述障碍对象的总数量有关的条件、以及/或者与所述障碍对象的显示状态有关的条件，
    所述被控对象的目标条件为与所述被控对象的行进情况有关的条件、以及/或者与所述被控对象的形态或位置有关的条件。
  14. 如权利要求13所述的活体检测设备，其中，所述障碍对象包括一个或多个对象，所述障碍对象的所述一个或多个对象在所述显示装置上移动，
    所述活体检测方法还包括：
    在尚未满足所述障碍对象的显示条件的情况下，在所述障碍对象中至少一部分对象移出显示装置时，显示新的至少一个对象，其中，所述新的至少一个对象和所述障碍对象中尚未移出显示装置的对象一起作为所述障碍对象，所述新的至少一个对象的显示位置是随机确定的。
  15. 如权利要求13所述的活体检测设备，其中，在所述计算机程序指令 被所述处理器运行时执行以下步骤：初始化定时器。
  16. 如权利要求15所述的活体检测设备，其中，所述虚拟对象还包括所述被控对象的目标对象，所述障碍对象的显示条件为所述障碍对象的总显示时间不超过预定定时时间，所述被控对象的目标条件为所述被控对象的显示位置与所述目标对象的显示位置重合，
    其中，在所述定时器未超出所述预定定时时间、所述被控对象的显示位置与所述目标对象的显示位置重合、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
  17. 如权利要求13所述的活体检测设备，其中，
    所述障碍对象的显示条件为所述障碍对象的总数量达到预定数量以及所述障碍对象均移出显示装置，
    其中，在所述障碍对象的总数量达到预定数量、所述障碍对象均移出显示装置、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
  18. 如权利要求15所述的活体检测设备，其中，所述障碍对象的显示条件为所述障碍对象的总显示时间不超过预定定时时间，所述被控对象的目标条件为所述被控对象的行进距离达到预定距离，
    其中，在所述定时器不超过预定定时时间、所述被控对象的行进距离达到预定距离、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
  19. 一种计算机程序产品，包括一个或多个计算机可读存储介质，所述计算机可读存储介质上存储了计算机程序指令，所述计算机程序指令在被计算机运行时执行以下步骤：
    从拍摄图像中检测人脸动作；
    根据所检测的人脸动作控制在显示装置上显示的虚拟对象中至少一部分的显示状态，其中，所述虚拟对象包括被控对象和障碍对象；
    在满足所述障碍对象的显示条件和/或所述被控对象的目标条件、且所述被控对象与所述障碍对象一直不相遇的情况下，确定所述拍摄图像中的人脸为活体人脸。
  20. 如权利要求19所述的计算机程序产品，其中，
    所述障碍对象的显示条件为与所述障碍对象的总显示时间有关的条件、 以及/或者与所述障碍对象的总数量有关的条件、以及/或者与所述障碍对象的显示状态有关的条件，
    所述被控对象的目标条件为与所述被控对象的行进情况有关的条件、以及/或者与所述被控对象的形态或位置有关的条件。
PCT/CN2015/082829 2015-06-30 2015-06-30 活体检测方法及设备、计算机程序产品 WO2017000218A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201580000355.3A CN105518714A (zh) 2015-06-30 2015-06-30 活体检测方法及设备、计算机程序产品
PCT/CN2015/082829 WO2017000218A1 (zh) 2015-06-30 2015-06-30 活体检测方法及设备、计算机程序产品

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2015/082829 WO2017000218A1 (zh) 2015-06-30 2015-06-30 活体检测方法及设备、计算机程序产品

Publications (1)

Publication Number Publication Date
WO2017000218A1 true WO2017000218A1 (zh) 2017-01-05

Family

ID=55725028

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/082829 WO2017000218A1 (zh) 2015-06-30 2015-06-30 活体检测方法及设备、计算机程序产品

Country Status (2)

Country Link
CN (1) CN105518714A (zh)
WO (1) WO2017000218A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110991235A (zh) * 2019-10-29 2020-04-10 北京海益同展信息科技有限公司 一种状态监测方法、装置、电子设备及存储介质
CN111274846A (zh) * 2018-12-04 2020-06-12 北京嘀嘀无限科技发展有限公司 一种张闭动作识别的方法和系统
CN111860343A (zh) * 2020-07-22 2020-10-30 杭州海康威视数字技术股份有限公司 确定人脸对比结果的方法及装置
CN112016482A (zh) * 2020-08-31 2020-12-01 成都新潮传媒集团有限公司 一种判别虚假人脸的方法、装置及计算机设备
CN113709564A (zh) * 2020-05-21 2021-11-26 云米互联科技(广东)有限公司 基于5g电视的预警方法、5g电视及可读存储介质
CN116110111A (zh) * 2023-03-23 2023-05-12 平安银行股份有限公司 人脸识别方法、电子设备及存储介质

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106154944A (zh) * 2016-06-30 2016-11-23 联想(北京)有限公司 一种数据处理方法及电子设备
CN106251396B (zh) 2016-07-29 2021-08-13 迈吉客科技（北京）有限公司 三维模型的实时控制方法和系统
CN108629260B (zh) * 2017-03-17 2022-02-08 北京旷视科技有限公司 活体验证方法和装置及存储介质
CN109271978A (zh) * 2018-11-23 2019-01-25 四川长虹电器股份有限公司 人脸识别防欺骗方法
CN111428570A (zh) * 2020-02-27 2020-07-17 深圳壹账通智能科技有限公司 非活体人脸的检测方法、装置、计算机设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101159016A (zh) * 2007-11-26 2008-04-09 清华大学 一种基于人脸生理性运动的活体检测方法及系统
US20100008552A1 (en) * 2008-07-14 2010-01-14 Union Community Co., Ltd. Apparatus for distinguishing living body fingerprint and method thereof
CN103020600A (zh) * 2012-12-12 2013-04-03 紫光股份有限公司 一种用于人脸识别系统的活体人脸鉴别装置
CN104166835A (zh) * 2013-05-17 2014-11-26 诺基亚公司 用于识别活体用户的方法和装置

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103778360A (zh) * 2012-10-26 2014-05-07 华为技术有限公司 一种基于动作分析的人脸解锁的方法和装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101159016A (zh) * 2007-11-26 2008-04-09 清华大学 一种基于人脸生理性运动的活体检测方法及系统
US20100008552A1 (en) * 2008-07-14 2010-01-14 Union Community Co., Ltd. Apparatus for distinguishing living body fingerprint and method thereof
CN103020600A (zh) * 2012-12-12 2013-04-03 紫光股份有限公司 一种用于人脸识别系统的活体人脸鉴别装置
CN104166835A (zh) * 2013-05-17 2014-11-26 诺基亚公司 用于识别活体用户的方法和装置

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111274846A (zh) * 2018-12-04 2020-06-12 北京嘀嘀无限科技发展有限公司 一种张闭动作识别的方法和系统
CN111274846B (zh) * 2018-12-04 2023-09-19 北京嘀嘀无限科技发展有限公司 一种张闭动作识别的方法和系统
CN110991235A (zh) * 2019-10-29 2020-04-10 北京海益同展信息科技有限公司 一种状态监测方法、装置、电子设备及存储介质
CN110991235B (zh) * 2019-10-29 2023-09-01 京东科技信息技术有限公司 一种状态监测方法、装置、电子设备及存储介质
CN113709564A (zh) * 2020-05-21 2021-11-26 云米互联科技(广东)有限公司 基于5g电视的预警方法、5g电视及可读存储介质
CN113709564B (zh) * 2020-05-21 2023-12-12 云米互联科技(广东)有限公司 基于5g电视的预警方法、5g电视及可读存储介质
CN111860343A (zh) * 2020-07-22 2020-10-30 杭州海康威视数字技术股份有限公司 确定人脸对比结果的方法及装置
CN111860343B (zh) * 2020-07-22 2023-04-28 杭州海康威视数字技术股份有限公司 确定人脸对比结果的方法及装置
CN112016482A (zh) * 2020-08-31 2020-12-01 成都新潮传媒集团有限公司 一种判别虚假人脸的方法、装置及计算机设备
CN116110111A (zh) * 2023-03-23 2023-05-12 平安银行股份有限公司 人脸识别方法、电子设备及存储介质
CN116110111B (zh) * 2023-03-23 2023-09-08 平安银行股份有限公司 人脸识别方法、电子设备及存储介质

Also Published As

Publication number Publication date
CN105518714A (zh) 2016-04-20

Similar Documents

Publication Publication Date Title
WO2017000213A1 (zh) 活体检测方法及设备、计算机程序产品
WO2017000218A1 (zh) 活体检测方法及设备、计算机程序产品
US10339402B2 (en) Method and apparatus for liveness detection
US10546183B2 (en) Liveness detection
TWI751161B (zh) 終端設備、智慧型手機、基於臉部識別的認證方法和系統
US9985963B2 (en) Method and system for authenticating liveness face, and computer program product thereof
EP3332403B1 (en) Liveness detection
CN105612533B (zh) 活体检测方法、活体检测系统以及计算机程序产品
CN107209849B (zh) 眼睛跟踪
CN105184246B (zh) 活体检测方法和活体检测系统
US9607138B1 (en) User authentication and verification through video analysis
WO2016172872A1 (zh) 用于验证活体人脸的方法、设备和计算机程序产品
CN105740688B (zh) 解锁的方法和装置
CN109684951A (zh) 人脸识别方法、底库录入方法、装置及电子设备
CN108875468B (zh) 活体检测方法、活体检测系统以及存储介质
US10846514B2 (en) Processing images from an electronic mirror
WO2017000217A1 (zh) 活体检测方法及设备、计算机程序产品
CN109858464B (zh) 底库数据处理方法、人脸识别方法、装置和电子设备
KR101640014B1 (ko) 허위 안면 이미지 분류가 가능한 홍채 인식 장치
WO2018103416A1 (zh) 用于人脸图像的检测方法和装置
CN105242888B (zh) 一种系统控制方法及电子设备
TWI498857B (zh) 瞌睡提醒裝置
US20230306792A1 (en) Spoof Detection Based on Challenge Response Analysis
JP4659722B2 (ja) 人体特定領域抽出・判定装置、人体特定領域抽出・判定方法、人体特定領域抽出・判定プログラム
Cai et al. Gaze estimation driven solution for interacting children with ASD

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15896749

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 08/05/2018)

122 Ep: pct application non-entry in european phase

Ref document number: 15896749

Country of ref document: EP

Kind code of ref document: A1