WO2019192256A1 - 人脸识别方法及装置 - Google Patents

人脸识别方法及装置 Download PDF

Info

Publication number
WO2019192256A1
WO2019192256A1 PCT/CN2019/073829 CN2019073829W WO2019192256A1 WO 2019192256 A1 WO2019192256 A1 WO 2019192256A1 CN 2019073829 W CN2019073829 W CN 2019073829W WO 2019192256 A1 WO2019192256 A1 WO 2019192256A1
Authority
WO
WIPO (PCT)
Prior art keywords
angle
face
face recognition
recognition
configuration information
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/CN2019/073829
Other languages
English (en)
French (fr)
Inventor
曾岳伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Alibaba Group Holding Ltd
Original Assignee
Alibaba Group Holding Ltd
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 Alibaba Group Holding Ltd filed Critical Alibaba Group Holding Ltd
Priority to US16/959,642 priority Critical patent/US20210019543A1/en
Publication of WO2019192256A1 publication Critical patent/WO2019192256A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/445—Program loading or initiating
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/445—Program loading or initiating
    • G06F9/44505—Configuring for program initiating, e.g. using registry, configuration files
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/24—Aligning, centring, orientation detection or correction of the image
    • G06V10/242—Aligning, centring, orientation detection or correction of the image by image rotation, e.g. by 90 degrees
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06V40/172—Classification, e.g. identification

Definitions

  • One or more embodiments of the present disclosure relate to the field of computer technologies, and in particular, to a face recognition method and apparatus.
  • the display angle of the image captured by the camera through the camera may be incorrect, or the recognition angle of the face recognition algorithm may be incorrect, and some devices may not recognize the face.
  • the problem is that there is a problem that the face recognition rate of some devices is zero.
  • One or more embodiments of the present specification describe a face recognition method and apparatus that can improve the success rate of a face recognition method.
  • a method for recognizing a face including:
  • the zero-pass rate model configuration table is configured to store a device model of the device with a face recognition pass rate of zero and a corresponding angle configuration information. Relationship; the angle configuration information is determined according to data recorded by the client's buried point when capturing a user's face recognition behavior;
  • the client configures the device according to the angle configuration information, and identifies the face through the configured device.
  • a method for recognizing a face including:
  • the face recognition request includes a device model of the device where the client is located; the face recognition request is used to instruct the server to configure from a zero pass rate model according to the device model Find the device in the table;
  • the response result includes indicating that the device and the corresponding angle configuration information are found, configuring the device according to the angle configuration information;
  • the face is identified by the configured device.
  • a face recognition device including:
  • a receiving unit configured to receive a face recognition request of the client, where the face recognition request includes a device model of the device where the client is located;
  • a searching unit configured to search for the device from a zero-pass rate model configuration table according to the device model received by the receiving unit;
  • the zero-pass rate model configuration table is configured to store a face recognition pass rate of zero Corresponding relationship between the device model of the device and the angle configuration information;
  • the angle configuration information is determined according to data recorded when the client's buried point captures the user's face recognition behavior;
  • An obtaining unit configured to acquire angle configuration information corresponding to the device model if the searching unit finds the device
  • a sending unit configured to return, to the client, the angle configuration information acquired by the acquiring unit, so that the client configures the device according to the angle configuration information, and passes the configured device pair The face is identified.
  • a face recognition device including:
  • a sending unit configured to send a face recognition request to the server, where the face recognition request includes a device model of the device where the face recognition device is located; and the face recognition request is used to indicate that the server is configured according to the device model. Find the device from the zero pass rate model configuration table;
  • a receiving unit configured to receive a response result returned by the server
  • a configuration unit configured to: if the response result received by the receiving unit is used to indicate that the device and the corresponding angle configuration information are found, configure the device according to the angle configuration information;
  • the identification unit is configured to identify the face by the device configured by the configuration unit.
  • a face recognition method and apparatus receives a face recognition request from a client.
  • the device is looked up from the zero pass rate model configuration table according to the device model of the device in the face recognition request. If the device is found, the angle configuration information corresponding to the model of the device is obtained from the zero-pass rate model configuration table.
  • the angle configuration information is returned to the client, so that the client configures the device according to the angle configuration information, and identifies the face through the configured device. Thereby, the success rate of the face recognition method can be improved.
  • FIG. 1 is a schematic diagram of an application scenario of a face recognition method provided by the present specification
  • FIG. 2 is a flow chart of a method for generating a zero-pass rate model configuration table provided by the present specification
  • FIG. 3 is a flowchart of a face recognition method according to an embodiment of the present disclosure
  • FIG. 4 is a flowchart of a face recognition method according to another embodiment of the present specification.
  • FIG. 5 is a schematic diagram of a face recognition apparatus according to an embodiment of the present disclosure.
  • FIG. 6 is a schematic diagram of a face recognition device according to another embodiment of the present specification.
  • the client has a corresponding camera, and the camera can be built in the device where the client is located, or can be connected outside the device.
  • the device may be, for example, a mobile phone, a tablet computer, or the like, which may have a corresponding operating system, and the operating system may be an Android system.
  • the android system has a default view, which can also be called a built-in window of the device.
  • the above built-in window refers to the general view.
  • the device may also have a corresponding external window, which may also be referred to as a display window.
  • the display window may refer to an interface rendered by the camera.
  • the built-in window of the device is usually not reconfigurable and is controlled by the android system.
  • the display window for the device can be reconfigured. For example, the display rotation angle of the display window can be configured.
  • the device where the client is located may also have a corresponding sensor. Based on the data sensed by the sensor, the orientation of the device can be determined.
  • the placement direction includes: a normal direction and an inverted direction.
  • the device can also have a built-in face recognition algorithm, and the face recognition algorithm has a corresponding algorithm recognition angle. Specifically, the device recognizes a face through a built-in face recognition algorithm.
  • the server may pre-generate a zero-pass rate model configuration table for storing a device model of a device with a face recognition pass rate of zero and a correspondence relationship of angle configuration information.
  • the above device with a face recognition pass rate of zero can also be referred to as a zero pass rate model.
  • the corresponding angle configuration information can be obtained from the zero pass rate model configuration table. And return the angle configuration information to the client.
  • the client configures the device according to the angle configuration information, and identifies the face through the configured device.
  • the angle configuration information may include: displaying a rotation angle and an algorithm recognition angle.
  • the display rotation angle refers to the rotation angle of the display window of the device, which can be determined according to the rotation angle of the built-in window.
  • the rotation angle of the built-in window can be obtained by calling the Application Programming Interface (API) function provided by the Android system.
  • API Application Programming Interface
  • the method of determining the rotation angle is a conventional conventional technique, and will not be described herein.
  • the algorithm recognition angle refers to the angle used by the device's face recognition algorithm. In one implementation, it can be determined according to the following formula: abs (360 degrees - display rotation angle).
  • the process of configuring the device may be: configuring a display rotation angle of the display window of the device according to the display rotation angle; Identify the angle and configure the algorithm recognition angle of the device's face recognition algorithm.
  • FIG. 2 is a flow chart of a method for generating a zero-pass rate model configuration table provided by the present specification. As shown in FIG. 2, the method may include:
  • Step 210 Acquire a device with a face recognition pass rate of zero and an adjustable angle.
  • the face recognition pass rate here may include the pass rate and the account pass rate.
  • the number of times of passing the device and the rate of account passing may be counted based on a plurality of users. Specifically, the number of times of passing may be determined according to the total number of times and the number of successes of face recognition by the device among the plurality of users. For example, for 50 users, the total number of face recognitions by the device is 100 times, wherein the number of successes is 60 times, then the pass rate of the device is 60%.
  • the account pass rate can be determined based on the total number of users who perform face recognition through the device and the number of successful users.
  • the total number of users for face recognition by the device is 50, wherein the number of successfully identified users is 30, and the account pass rate of the device is 60%. It should be noted that when determining the account pass rate, when a user performs face recognition multiple times through the device identification, once the recognition is successful, the user is the user who successfully identifies.
  • the process of acquiring a device with a face recognition rate of zero may be: collecting a plurality of devices in advance; and acquiring, for each device of the plurality of devices, user behavior data corresponding to the device.
  • the user behavior data herein may be recorded by the client's buried point (a way of recording data) when capturing the user's face recognition behavior through the device and completing the face recognition behavior and/or abandoning the face recognition behavior.
  • the face recognition pass rate of each device is determined according to user behavior data corresponding to each device.
  • the user behavior data here may include information such as account identification, device model, and time.
  • the face recognition pass rate may refer to the pass rate and/or the account pass rate.
  • the total number of times of using the respective devices for face recognition and the number of successes can be counted according to the information such as the account identifier, the device model, and the time. Then for each device, the corresponding number of pass rates is calculated. Or, for each device, count the total number of users who perform face recognition through the device and the number of users that successfully identify, and then calculate the corresponding account pass rate. Thereafter, a device whose face recognition pass rate (number pass rate and/or account pass rate) is zero is selected from a plurality of devices.
  • the device of adjustable angle in step 210 may refer to a device having front and rear cameras and/or recorded in a white list.
  • the devices in the whitelist can be pre-collected by humans.
  • Step 220 Determine whether the device has a corresponding algorithm output value.
  • the algorithm output value herein may be output when the device recognizes a human face, and may include: face quality score, position, face coordinates, and the like. Specifically, the algorithm output value may be recorded by the burying point of the client when the capturing user recognizes the face through the device. It can be understood that if the algorithm recognizes that the angle is wrong, the face recognition algorithm will be invalid, and the device will not recognize the face. When the device does not recognize the face, there will be no corresponding algorithm output value.
  • Step 230 if yes, acquiring the current recognition angle of the device, and using the current recognition angle as an algorithm recognition angle.
  • the current recognition angle here may refer to an algorithm recognition angle used by the device when recognizing a face by a face recognition algorithm. It can also be recorded by the client's buried point. For example, it may be recorded by the burying point when the capturing user recognizes the face through the device.
  • Step 240 if not, acquiring the current recognition angle of the device, correcting the current recognition angle, and obtaining an algorithm recognition angle.
  • the current recognition angle can be corrected according to the following formula: abs (360 degrees - current recognition angle). After correction, get the correct algorithm to identify the angle.
  • Step 250 Determine whether the face displayed by the display window of the device is inverted.
  • the process of determining whether the face displayed by the display window of the device is inverted may be: acquiring sensor data of the device and displaying data of the face captured by the camera of the device, and the sensor data and the display data may be The client's burying point is recorded when the user's behavior is recognized by the device configured by default. Based on the sensor data and the display data, it is determined whether the face displayed by the display window of the device is inverted.
  • the sensor data described above may be, for example, three-dimensional coordinates: x, y, and z. Based on the three-dimensional coordinates, the direction in which the device is placed can be determined.
  • the placement direction includes: a normal direction and an inverted direction.
  • the above display data may refer to the point coordinates of the upper left corner of the face and the point coordinates of the lower right corner.
  • the placement direction of the device is the normal direction, if the point coordinate of the upper left corner is smaller than the point coordinate of the lower right corner, the face displayed by the display window of the device is not inverted; otherwise, the orientation is inverted.
  • the orientation of the device is in the inverted direction, the above determination process is not performed. That is, the embodiment of the present specification only cares about the case where the device is placed in the normal direction.
  • Step 260 if yes, obtain the current rotation angle of the display window, and correct the current rotation angle to obtain the display rotation angle.
  • the current rotation angle can also be recorded by the buried point of the client. For example, it may be recorded by the burying point when the capturing user recognizes the face through the device. Specifically, if the face is inverted, the current rotation angle is wrong and needs to be corrected.
  • the current rotation angle can be corrected according to the following formula: abs (360 degrees - current rotation angle). After correction, get the correct display rotation angle.
  • Step 270 if no, obtain the current rotation angle of the display window, and use the current rotation angle as the display rotation angle.
  • Step 280 Generate a zero-pass rate model configuration table according to the device model of the face recognition rate, the angle of the adjustable angle device, the algorithm recognition angle, and the display rotation angle.
  • a zero pass rate model configuration table can be generated.
  • the zero pass rate model configuration table does not record the angle configuration information of the part of the device.
  • the generated zero pass rate model configuration table can be as shown in Table 1.
  • X, Y, and Z are 90 degrees or 270 degrees.
  • NULL indicates that the angle configuration information of the device of this device model is empty.
  • the face After generating the above zero-pass rate model configuration table, the face can be recognized.
  • FIG. 3 is a flowchart of a face recognition method according to an embodiment of the present disclosure.
  • the execution body of the method may be a server. As shown in FIG. 3, the method may include:
  • Step 310 Receive a face recognition request of the client.
  • the face recognition request may include the device model of the device where the client is located.
  • the client may send the face recognition request to the server when the user clicks the “Blip Login” button on the login page of the Alipay client.
  • Step 320 Find the device from the zero pass rate model configuration table according to the device model.
  • the device model can be compared with each device model in Table 1. If the device model is consistent with any of the device models in Table 1, the device is found. It can be understood that, in the case of finding, the device of the above device model is described as a zero pass rate device. Otherwise, the device is not found.
  • Step 330 If the device is found, obtain angle configuration information corresponding to the model of the device.
  • the angle configuration information corresponding to the device model can be obtained: Y and 360-Y.
  • step 340 the angle configuration information is returned to the client.
  • the client After receiving the angle configuration information returned by the server, the client configures the device according to the angle configuration information, and identifies the face through the configured device.
  • the process of configuring the device may be: configuring the display rotation angle of the display window of the device as: Y, and configuring the algorithm recognition angle of the device face recognition algorithm to be 360-Y.
  • the server finds the device but the corresponding angle configuration information is empty, the client enables the default configuration of the device to identify the face.
  • the default configuration here can mean that the angle configuration information of the device is set by the android system.
  • the client gets the rotation angle of the device's built-in window.
  • the display rotation angle is determined according to the rotation angle of the built-in window.
  • the algorithm recognizes the angle.
  • Configure the device based on the determined display rotation angle and algorithm recognition angle.
  • the face is identified by the configured device.
  • the success rate of face recognition can be greatly improved by generating a zero-pass rate configuration table by the server and delivering the angle configuration information to the client.
  • FIG. 4 is a flowchart of a face recognition method according to another embodiment of the present specification.
  • the execution body of the method may be a client. As shown in FIG. 4, the method may include:
  • step 410 the client receives a face recognition command.
  • the above-mentioned face recognition command may be triggered by the user clicking the “Brush Face Login” button on the login page of the Alipay client.
  • step 420 it is determined whether the device where the client is located is a zero-pass rate model. If no, step 430-step 450 is performed; if yes, step 460 is performed.
  • the client may send a face recognition request to the server, and the face recognition request may include a device model of the device where the client is located.
  • the server looks up the device from the zero-pass rate model configuration table based on the device model. Return the response result to the client. If the response result includes information indicating that the device has been found, it is determined that the device is a zero pass rate model. Otherwise, make sure the device is not a zero pass rate model.
  • Step 430 Acquire a rotation angle of the built-in window of the device.
  • Step 440 Determine the angle configuration information of the device according to the rotation angle of the built-in window.
  • the angle configuration information may include: displaying a rotation angle and an algorithm recognition angle. Specifically, according to the rotation angle of the built-in window, the display rotation angle is determined, and the burying point is automatically set for the display rotation angle. It can be understood that the buried point here corresponds to the buried point in step 260.
  • the method for determining the above-described display rotation angle is a conventional conventional technique, and will not be described herein.
  • the algorithm recognizes the angle and automatically sets the burying point for the algorithm recognition angle. It can be understood that the buried point here corresponds to the buried point in step 230.
  • the algorithm recognizes the angle as follows: abs (360 degrees - display rotation angle).
  • Step 450 Configure the device according to the determined angle configuration information.
  • step 460 it is determined whether the corresponding angle configuration information can be obtained. If no, step 470 is performed; if yes, step 480 is performed.
  • the response result it can be determined whether the corresponding angle configuration information can be obtained. Specifically, if the response result further includes corresponding angle configuration information, it is determined that the corresponding angle configuration information can be obtained.
  • the angle configuration information may include: displaying a rotation angle and an algorithm recognition angle.
  • step 470 the device configured with the default configuration is used to identify the face.
  • the default configuration of the device for face recognition process can be: a, perform a screen capture operation. b. Identify the face of the screen by the face recognition algorithm built into the device. c. If the face is not recognized, repeat steps a and b above. If a face is recognized, step d is performed. In step d, a burying point is automatically set for the sensor data of the device and the display data of the face captured by the camera of the device. The client can then upload the sensor data and display data to the server.
  • the buried point here corresponds to the buried point in step 250.
  • Step 480 Configure the device according to the obtained angle configuration information.
  • Step 490 automatically setting a burying point for the user behavior data of the user's starting face recognition behavior.
  • the burying point captures the user's starting face recognition behavior, and then records the corresponding user behavior data. It can be understood that the buried point here corresponds to the buried point in step 210.
  • step 4100 the face is identified by the configured device.
  • step 4110 it is determined whether the face recognition is passed. If not, step 4120 is performed; if yes, step 4140 is performed.
  • step 4120 it is determined whether the face recognition is ended. If not, step 4100 is performed; if yes, step 4130 is performed.
  • Step 4130 automatically setting a burying point for the user behavior data of the user to abandon the face recognition behavior.
  • the burying point captures the user's abandonment face recognition behavior, and then records the corresponding user behavior data. It can be understood that the buried point here corresponds to the buried point in step 210.
  • Step 4140 automatically setting a burying point for the user behavior data of the user completing the face recognition behavior.
  • the burying point captures the user's completed face recognition behavior, and then records the corresponding user behavior data. It can be understood that the buried point here corresponds to the buried point in step 210.
  • the embedding point is automatically set to the algorithm output value of the user's face recognition behavior. That is, the burying point is automatically set for the effective face recognition algorithm.
  • the buried point captures the user's completed face recognition behavior, and records the algorithm output value corresponding to the face recognition algorithm.
  • the buried point here corresponds to the buried point in step 220.
  • Step 4150 ending face recognition.
  • the correct angle configuration information of the zero-pass rate model is calculated, thereby achieving the recognition of the face for all the android system devices. purpose.
  • the success rate of face recognition is greatly improved without affecting the efficiency of the face recognition method.
  • an embodiment of the present disclosure further provides a face recognition device, as shown in FIG. 5, the device includes:
  • the receiving unit 501 is configured to receive a face recognition request of the client, where the face recognition request includes a device model of the device where the client is located.
  • the searching unit 502 is configured to search for the device from the zero-pass rate model configuration table according to the device model received by the receiving unit 501; the zero-pass rate model configuration table is configured to store the device model of the device with a face recognition pass rate of zero. And the corresponding relationship of the angle configuration information; the angle configuration information is determined according to data recorded when the client's buried point captures the user's face recognition behavior;
  • the obtaining unit 503 is configured to obtain angle configuration information corresponding to the model of the device if the searching unit 502 finds the device.
  • the sending unit 504 is configured to return the angle configuration information acquired by the obtaining unit 503 to the client, so that the client configures the device according to the angle configuration information, and identifies the face through the configured device.
  • the angle configuration information may include: displaying a rotation angle and an algorithm recognition angle.
  • the apparatus may also include a generating unit 505.
  • the generating unit 505 can be used to:
  • the current recognition angle of the device is obtained, and the current recognition angle refers to an angle used by the device to recognize the face through the face recognition algorithm; the current recognition angle is used as an algorithm to recognize the angle.
  • the current recognition angle of the device is obtained, and the current recognition angle is corrected to obtain an algorithm recognition angle.
  • the current rotation angle of the display window is obtained, and the current rotation angle is used as the display rotation angle.
  • a zero-pass rate model configuration table is generated according to the device model of the face recognition rate, the angle of the adjustable angle device, the algorithm recognition angle, and the display rotation angle.
  • the generating unit 505 is specifically configured to:
  • the user behavior data corresponding to the device is obtained, and the user behavior data is obtained by the client's burying point in capturing the user's face recognition behavior through the device and completing the face recognition behavior and/or abandoning the face recognition. Recorded when behavior.
  • the face recognition pass rate of each device is determined according to user behavior data corresponding to each device.
  • a device with a face recognition pass rate of zero is selected from a plurality of devices.
  • the generating unit 505 is further specifically configured to:
  • the sensor data of the device and the display data of the face captured by the camera of the device are acquired, and the sensor data and the display data are recorded by the client's buried point when the captured user recognizes the face through the default configuration device.
  • the receiving unit 501 receives a face recognition request of the client.
  • the lookup unit 502 looks up the device from the zero pass rate model configuration table according to the device model. If the device is found, the obtaining unit 503 acquires angle configuration information corresponding to the model of the device.
  • the sending unit 504 returns the angle configuration information to the client, so that the client configures the device according to the angle configuration information, and identifies the face through the configured device. Thereby, the success rate of face recognition can be improved.
  • the face recognition device provided in the embodiment of the present specification may be a module or a unit of the server in FIG.
  • an embodiment of the present disclosure further provides a face recognition device, as shown in FIG. 6, the device includes:
  • the sending unit 601 is configured to send a face recognition request to the server, where the face recognition request includes a device model of the device where the face recognition device is located; the face recognition request is used to instruct the server to use the zero pass rate model configuration table according to the device model Find the device in .
  • the receiving unit 602 is configured to receive a response result returned by the server.
  • the configuration unit 603 is configured to: if the response result received by the receiving unit 602 includes information indicating that the device has been found and corresponding angle configuration information, configure the device according to the angle configuration information.
  • the identifying unit 604 is configured to identify the face by the device configured by the configuration unit 603.
  • the identifying unit 604 is further configured to: if the response result received by the receiving unit 602 includes only information indicating that the device has been found, identify the face by using the device configured by default.
  • the device may further include:
  • the obtaining unit 605 is configured to acquire a rotation angle of the built-in window of the device if the response result received by the receiving unit 602 includes information indicating that the device is not found.
  • the determining unit 606 is configured to determine the angle configuration information of the device according to the rotation angle of the built-in window acquired by the obtaining unit 605.
  • the configuration unit 603 is further configured to configure the device according to the angle configuration information determined by the determining unit 606.
  • the identification unit 604 is further configured to identify the face by the device configured by the configuration unit 603.
  • the transmitting unit 601 transmits a face recognition request to the server.
  • the receiving unit 602 receives the response result returned by the server. If the response result includes information indicating that the device has been found and corresponding angle configuration information, the configuration unit 603 configures the device according to the angle configuration information.
  • the identification unit 604 identifies the face through the configured device. Thereby, the success rate of face recognition can be improved.
  • the face recognition device provided in the embodiment of the present specification may be a module or a unit of the client in FIG. 1 .
  • the functions described in this specification can be implemented in hardware, software, firmware, or any combination thereof.
  • the functions may be stored in a computer readable medium or transmitted as one or more instructions or code on a computer readable medium.

Landscapes

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

Abstract

一种人脸识别方法及装置，在人脸识别方法中，接收客户端的人脸识别请求（S310）。根据设备型号，从零通过率机型配置表中查找该设备（S320）。若查找到该设备，则获取对应于该设备型号的角度配置信息（S330）。向客户端返回角度配置信息（S340），以使客户端根据角度配置信息，对该设备进行配置，并通过配置后的设备对人脸进行识别。

Description

人脸识别方法及装置 技术领域
本说明书一个或多个实施例涉及计算机技术领域，尤其涉及一种人脸识别方法及装置。
背景技术
传统技术中，在用户通过设备对人脸进行识别时，可能会由于设备通过摄像头所捕获图像的显示角度错误，或者人脸识别算法的识别角度不正确，而导致部分设备无法对人脸进行识别的问题，也即会出现部分设备的人脸识别通过率为零的问题。
发明内容
本说明书一个或多个实施例描述了一种人脸识别方法及装置，可以提高人脸识别方法的成功率。
第一方面，提供了一种人脸识别方法，包括：
接收客户端的人脸识别请求，所述人脸识别请求包括所述客户端所在设备的设备型号；
根据所述设备型号，从零通过率机型配置表中查找所述设备；所述零通过率机型配置表用于存储人脸识别通过率为零的设备的设备型号以及角度配置信息的对应关系；所述角度配置信息是根据所述客户端的埋点在捕获用户的人脸识别行为时所记录的数据确定的；
若查找到所述设备，则获取对应于所述设备型号的角度配置信息；
向所述客户端返回所述角度配置信息，以使所述客户端根据所述角度配置信息，对所述设备进行配置，并通过配置后的设备对人脸进行识别。
第二方面，提供了一种人脸识别方法，包括：
向服务器发送人脸识别请求，所述人脸识别请求包括所述客户端所在设备的设备型号；所述人脸识别请求用于指示所述服务器根据所述设备型号，从零通过率机型配置表中查找所述设备；
接收所述服务器返回的响应结果；
若所述响应结果包括用于表示已查找到所述设备和对应的角度配置信息，则根据所述角度配置信息，对所述设备进行配置；
通过配置后的设备对人脸进行识别。
第三方面，提供了一种人脸识别装置，包括：
接收单元，用于接收客户端的人脸识别请求，所述人脸识别请求包括所述客户端所在设备的设备型号；
查找单元，用于根据所述接收单元接收的所述设备型号，从零通过率机型配置表中查找所述设备；所述零通过率机型配置表用于存储人脸识别通过率为零的设备的设备型号以及角度配置信息的对应关系；所述角度配置信息是根据所述客户端的埋点在捕获用户的人脸识别行为时所记录的数据确定的；
获取单元，用于若所述查找单元查找到所述设备，则获取对应于所述设备型号的角度配置信息；
发送单元，用于向所述客户端返回所述获取单元获取的所述角度配置信息，以使所述客户端根据所述角度配置信息，对所述设备进行配置，并通过配置后的设备对人脸进行识别。
第四方面，提供了一种人脸识别装置，包括：
发送单元，用于向服务器发送人脸识别请求，所述人脸识别请求包括所述人脸识别装置所在设备的设备型号；所述人脸识别请求用于指示所述服务器根据所述设备型号，从零通过率机型配置表中查找所述设备；
接收单元，用于接收所述服务器返回的响应结果；
配置单元，用于若所述接收单元接收的所述响应结果包括用于表示已查找到所述设备和对应的角度配置信息，则根据所述角度配置信息，对所述设备进行配置；
识别单元，用于通过所述配置单元配置后的设备对人脸进行识别。
本说明书一个或多个实施例提供的人脸识别方法及装置，接收客户端的人脸识别请求。根据人脸识别请求中的设备的设备型号，从零通过率机型配置表中查找该设备。若查找到该设备，则从零通过率机型配置表中获取对应于该设备型号的角度配置信息。向客户端返回角度配置信息，以使客户端根据角度配置信息，对该设备进行配置，并通过 配置后的设备对人脸进行识别。由此，可以提高人脸识别方法的成功率。
附图说明
为了更清楚地说明本说明书实施例的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本说明书的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其它的附图。
图1为本说明书提供的人脸识别方法的应用场景示意图；
图2为本说明书提供的零通过率机型配置表的生成方法流程图；
图3为本说明书一个实施例提供的人脸识别方法流程图；
图4为本说明书另一个实施例提供的人脸识别方法流程图；
图5为本说明书一个实施例提供的人脸识别装置示意图；
图6为本说明书另一个实施例提供的人脸识别装置示意图。
具体实施方式
下面结合附图，对本说明书提供的方案进行描述。
本说明书提供的人脸识别方法可以应用于如图1所示的场景中，图1中，客户端具有对应的摄像头，该摄像头可以内置于客户端所在设备，也可以连接在该设备的外部。上述设备例如可以为手机、平板电脑等，其可以具有对应的操作系统，该操作系统可以为安卓(android)系统。android系统具有默认的窗口(view)，该窗口也可以称为设备的内置窗口。上述内置窗口指的是一般的视图。此外，该设备还可以具有对应的外部窗口，该外部窗口也可以称为显示窗口(display)。该显示窗口可以是指摄像头渲染的界面。对设备的内置窗口，通常不可以重新配置，其由android系统进行操控。对设备的显示窗口，则可以重新配置。如，可以配置显示窗口的显示旋转角度等。
需要说明的是，客户端所在的设备还可以具有对应的传感器。根据该传感器所感应的数据，可以确定设备的摆放方向。该摆放方向包括：正常方向和倒置方向。此外，该设备还可以内置人脸识别算法，该人脸识别算法具有对应的算法识别角度。具体地，设备通过内置的人脸识别算法来对人脸进行识别。
图1中，服务器可以预先生成零通过率机型配置表，该零通过率机型配置表用于存储人脸识别通过率为零的设备的设备型号以及角度配置信息的对应关系。上述人脸识别通过率为零的设备也可以称为零通过率机型。具体地，在对人脸进行识别时，若客户端所在的设备为零通过率机型，则可以从零通过率机型配置表中获取对应的角度配置信息。并将该角度配置信息返回给客户端。由客户端根据角度配置信息，对设备进行配置，并通过配置后的设备对人脸进行识别。
上述角度配置信息可以包括：显示旋转角度和算法识别角度。显示旋转角度是指设备的显示窗口的旋转角度，其可以根据内置窗口的旋转角度来确定。内置窗口的旋转角度则可以通过调用android系统提供的应用程序编程接口(Application Programming Interface,API)函数来获取。显示旋转角度的确定方法为传统常规技术，在此不复赘述。算法识别角度是指设备的人脸识别算法所使用的角度。在一种实现方式中，其可以根据如下公式来确定：abs(360度-显示旋转角度)。
可以理解的是，当角度配置信息可以包括：显示旋转角度和算法识别角度时，上述对设备进行配置的过程可以为：根据显示旋转角度，对设备的显示窗口的显示旋转角度进行配置；根据算法识别角度，对设备的人脸识别算法的算法识别角度进行配置。
从图1可以看出，为使零通过率机型也可以对人脸进行识别，则可以先生成零通过率机型配置表。图2为本说明书提供的零通过率机型配置表的生成方法流程图。如图2所示，该方法可以包括：
步骤210，获取人脸识别通过率为零、可调角度的设备。
此处的人脸识别通过率可以包括次数通过率和账户通过率。本说明书实施例中，可以基于多个用户，来统计设备的次数通过率和账户通过率。具体地，可以根据多个用户中通过该设备进行人脸识别的总次数以及成功次数，来确定次数通过率。举例来说，50个用户，通过该设备进行人脸识别的总次数为：100次，其中，成功次数为：60次，那么该设备的次数通过率为：60％。可以根据通过该设备进行人脸识别的总用户数以及识别成功的用户数，来确定账户通过率。举例来说，通过该设备进行人脸识别的总用户数为：50个，其中，识别成功的用户数为：30个，那么该设备的账户通过率为：60％。需要说明的是，在确定账户通过率时，对某个用户，当其通过设备识别多次进行人脸识别时，只要有一次识别成功，那么该用户即为识别成功的用户。
在一种实现方式中，获取人脸识别通过率为零的设备的过程可以为：预先收集多个 设备；对多个设备中的每个设备，获取该设备对应的用户行为数据。此处的用户行为数据可以是由客户端的埋点(一种记录数据的方式)在捕获用户通过该设备开始人脸识别行为以及完成人脸识别行为和/或放弃人脸识别行为时记录。根据各个设备对应的用户行为数据，确定各个设备的人脸识别通过率。此处的用户行为数据可以包括：账户标识、设备型号以及时间等信息。人脸识别通过率可以是指次数通过率和/或账户通过率。具体地，可以根据账户标识、设备型号以及时间等信息，统计使用各个设备进行人脸识别的总次数为以及成功次数。然后针对每个设备，计算对应的次数通过率。或者，针对每个设备，统计通过该设备进行人脸识别的总用户数以及识别成功的用户数，然后计算对应的账户通过率。之后，从多个设备中选取人脸识别通过率(次数通过率和/或账户通过率)为零的设备。
步骤210中的可调角度的设备可以是指有前、后置摄像头和/或记录在白名单中的设备。此处，白名单中的设备可以是人为预先收集的。
步骤220，确定设备是否有对应的算法输出值。
此处的算法输出值可以是在设备识别到人脸时输出，其可以包括：人脸质量分，位置，人脸坐标等信息。具体地，可以由客户端的埋点在捕获用户通过该设备识别到人脸时记录上述算法输出值。可以理解的是，若算法识别角度是错误的，那么人脸识别算法就会失效，从而设备就会识别不到人脸。当设备识别不到人脸时，就不会有对应的算法输出值。
综上，可以通过判断设备是否有对应的算法输出值，来确定当前所使用的算法识别角度是否正确。
步骤230，若是，则获取设备的当前识别角度，将当前识别角度作为算法识别角度。
此处的当前识别角度可以是指设备在通过人脸识别算法识别人脸时所使用的算法识别角度。其也可以由客户端的埋点记录。如，可以是由埋点在捕获用户通过设备识别人脸时记录。
步骤240，若否，则获取设备的当前识别角度，对当前识别角度进行更正，得到算法识别角度。
在一个例子中，可以根据如下公式来更正当前识别角度：abs(360度-当前识别角度)。更正之后，得到正确的算法识别角度。
步骤250，确定设备的显示窗口所显示的人脸是否倒置。
在一种实现方式中，确定设备的显示窗口所显示的人脸是否倒置的过程可以为：获取设备的传感器数据以及设备的摄像头所捕获的人脸的显示数据，传感器数据和显示数据可以是由客户端的埋点在捕获用户通过默认配置的设备识别到人脸的行为时记录。根据传感器数据以及显示数据，确定设备的显示窗口所显示的人脸是否倒置。
举例来说，上述传感器数据例如可以为三维坐标：x、y和z。根据该三维坐标，可以确定设备的摆放方向。该摆放方向包括：正常方向和倒置方向。上述显示数据可以是指人脸的左上角的点坐标和右下角的点坐标。具体地，当设备的摆放方向为正常方向时，如果左上角的点坐标小于右下角的点坐标，则设备的显示窗口所显示的人脸不倒置；否则倒置。当设备的摆放方向为倒置方向时，不执行上述确定过程。也即，本说明书实施例只关心设备的摆放方向为正常方向的情况。
步骤260，若是，则获取显示窗口的当前旋转角度，对当前旋转角度进行更正，得到显示旋转角度。
当前旋转角度也可以由客户端的埋点记录。如，可以是由埋点在捕获用户通过设备识别人脸时记录。具体地，如果人脸倒置，则说明当前旋转角度是错误的，需要对其更正。在一个例子中，可以根据如下公式来更正当前旋转角度：abs(360度-当前旋转角度)。更正之后，得到正确的显示旋转角度。
步骤270，若否，则获取显示窗口的当前旋转角度，将当前旋转角度作为显示旋转角度。
如果人脸不倒置，则说明当前旋转角度是正确的，不需要更正。
步骤280，根据人脸识别通过率为零、可调角度的设备的设备型号、算法识别角度以及显示旋转角度，生成零通过率机型配置表。
可以理解的是，针对人脸识别通过率为零、可调角度的设备，在确定对应的算法识别角度以及显示旋转角度之后，就可以生成零通过率机型配置表。
当然，在实际应用中，针对部分人脸识别通过率为零的设备，如果客户端的埋点未捕获或者捕获用户的人脸识别行为失败，则服务器无法确定该设备的角度配置信息。从而零通过率机型配置表不记录该部分设备的角度配置信息。
在一个例子中，生成的零通过率机型配置表可以如表1所示。
表1
零通过率机型 显示旋转角度 算法识别角度
X9S X 360-X
V9play Y 360-Y
R11s Plus NULL NULL
... ... ...
表1中，X、Y、Z为90度或者270度。“NULL”表示该设备型号的设备的角度配置信息为空。
在生成上述零通过率机型配置表之后，就可以对人脸进行识别。
图3为本说明书一个实施例提供的人脸识别方法流程图。所述方法的执行主体可以为服务器。如图3所示，所述方法可以包括：
步骤310，接收客户端的人脸识别请求。
该人脸识别请求可以包括客户端所在设备的设备型号。
以客户端为支付宝客户端为例来说，可以是在用户在支付宝客户端的登录页面点击“刷脸登录”的beta按钮时，客户端向服务器发送上述人脸识别请求。
步骤320，根据设备型号，从零通过率机型配置表中查找该设备。
如，可以将上述设备型号与表1中的各个设备型号进行比对，若与表1中的任一设备型号比对一致，则查找到该设备。可以理解的是，在查找到的情况下，说明上述设备型号的设备为零通过率机型的设备。否则，未查找到该设备。
步骤330，若查找到该设备，则获取对应于该设备型号的角度配置信息。
如前述例子，假设上述设备型号与表1中第2行的设备型号比对一致，则可以获取对应于该设备型号的角度配置信息：Y和360-Y。
步骤340，向客户端返回角度配置信息。
客户端在接收到服务器返回的角度配置信息之后，根据角度配置信息，对设备进行配置，并通过配置后的设备对人脸进行识别。如前述例子，对设备进行配置的过程可以为：将设备的显示窗口的显示旋转角度配置为：Y，将设备的人脸识别算法的算法识别 角度配置为：360-Y。
需要说明的是，如果服务器查找到该设备，但对应的角度配置信息为空，则客户端启用默认配置的设备对人脸进行识别。此处的默认配置可以是指该设备的角度配置信息由android系统设定。
此外，如果服务器未查找到该设备，则客户端获取该设备的内置窗口的旋转角度。根据内置窗口的旋转角度，确定显示旋转角度。之后，根据显示旋转角度，确定算法识别角度。根据确定的显示旋转角度和算法识别角度，对设备进行配置。通过配置后的设备对人脸进行识别。需要说明的是，此处的显示旋转角度和算法识别角度的确定过程为传统常规技术，在此不复赘述。
本实施例中，通过服务器生成零通过率配置表，并将角度配置信息下发至客户端的方式，可以大大提高人脸识别的成功率。
图4为本说明书另一个实施例提供的人脸识别方法流程图。所述方法的执行主体可以为客户端。如图4所示，所述方法可以包括：
步骤410，客户端接收人脸识别指令。
以客户端为支付宝客户端为例来说，上述人脸识别指令可以是由用户在支付宝客户端的登录页面点击“刷脸登录”的beta按钮触发。
步骤420，判断客户端所在设备是否是零通过率机型。若否，则执行步骤430-步骤450；若是，则执行步骤460。
具体地，客户端可以向服务器发送人脸识别请求，该人脸识别请求可以包括客户端所在设备的设备型号。服务器根据设备型号，从零通过率机型配置表中查找该设备。向客户端返回响应结果。如果该响应结果包括用于表示已查找到该设备的信息，则确定该设备是零通过率机型。否则，确定该设备不是零通过率机型。
步骤430，获取设备的内置窗口的旋转角度。
步骤440，根据内置窗口的旋转角度，确定设备的角度配置信息。
该角度配置信息可以包括：显示旋转角度和算法识别角度。具体地，根据内置窗口的旋转角度，确定显示旋转角度，并自动对显示旋转角度设置埋点。可以理解的是，此处的埋点与步骤260中的埋点相对应。上述显示旋转角度的确定方法为传统常规技术，在此不复赘述。之后，根据显示旋转角度，确定算法识别角度，并自动对算法识别角度 设置埋点。可以理解的是，此处的埋点与步骤230中的埋点相对应。在一种实现方式中，算法识别角度的确定方式如下：abs(360度-显示旋转角度)。
步骤450，根据确定的角度配置信息，对设备进行配置。
步骤460，判断是否可以获取到对应的角度配置信息。若否，则执行步骤470；若是，则执行步骤480。
此处，可以根据响应结果，判断是否可以获取到对应的角度配置信息。具体地，如果响应结果还包括对应的角度配置信息，则确定可以获取到对应的角度配置信息。该角度配置信息可以包括：显示旋转角度和算法识别角度。
步骤470，启用默认配置的设备对人脸进行识别。
此处，默认配置的设备对人脸识别的过程可以为：a，执行截屏操作。b，通过设备内置的人脸识别算法识别截屏中是否有人脸。c，如果未识别到人脸，则重复执行上述步骤a和步骤b。如果识别到人脸，则执行步骤d。步骤d，自动对设备的传感器数据以及设备的摄像头捕获的人脸的显示数据设置埋点。之后，客户端可以向服务器上传该传感器数据和显示数据。
可以理解的是，此处的埋点与步骤250中的埋点相对应。
步骤480，根据获取到的角度配置信息，对设备进行配置。
步骤490，自动对用户的开始人脸识别行为的用户行为数据设置埋点。
具体地，埋点捕获用户的开始人脸识别行为，之后记录对应的用户行为数据。可以理解的是，此处的埋点与步骤210中的埋点相对应。
步骤4100，通过配置后的设备对人脸进行识别。
步骤4110，判断是否通过人脸识别，若否，则执行步骤4120；若是，则执行步骤4140。
步骤4120，判断是否结束人脸识别，若否，则执行步骤4100；若是，则执行步骤4130。
步骤4130，自动对用户的放弃人脸识别行为的用户行为数据设置埋点。
具体地，埋点捕获用户的放弃人脸识别行为，之后记录对应的用户行为数据。可以理解的是，此处的埋点与步骤210中的埋点相对应。
步骤4140，自动对用户的完成人脸识别行为的用户行为数据设置埋点。
具体地，埋点捕获用户的完成人脸识别行为，之后记录对应的用户行为数据。可以理解的是，此处的埋点与步骤210中的埋点相对应。
此外，自动对用户的完成人脸识别行为的算法输出值设置埋点。也即自动对生效的人脸识别算法设置埋点。该埋点捕获用户的完成人脸识别行为，记录人脸识别算法对应的算法输出值。此处的埋点与步骤220中的埋点相对应。
步骤4150，结束人脸识别。
本实施例中，通过自动对客户端的数据设置埋点的方式，来计算出零通过率机型的正确的角度配置信息，由此达到了针对所有android系统的设备均可以对人脸进行识别的目的。此外，由于上述设置以及计算的过程是零人工参与的，从而在不会影响人脸识别方法效率的情况下，大大提高了人脸识别的成功率。
与上述人脸识别方法对应地，本说明书一个实施例还提供的一种人脸识别装置，如图5所示，该装置包括：
接收单元501，用于接收客户端的人脸识别请求，该人脸识别请求包括客户端所在设备的设备型号。
查找单元502，用于根据接收单元501接收的设备型号，从零通过率机型配置表中查找该设备；零通过率机型配置表用于存储人脸识别通过率为零的设备的设备型号以及角度配置信息的对应关系；角度配置信息是根据客户端的埋点在捕获用户的人脸识别行为时所记录的数据确定的；
获取单元503，用于若查找单元502查找到该设备，则获取对应于该设备型号的角度配置信息。
发送单元504，用于向客户端返回获取单元503获取的角度配置信息，以使客户端根据角度配置信息，对该设备进行配置，并通过配置后的设备对人脸进行识别。
可选地，上述角度配置信息可以包括：显示旋转角度和算法识别角度。该装置还可以包括：生成单元505。
生成单元505可以用于：
获取人脸识别通过率为零、可调角度的设备。
确定设备是否有对应的算法输出值，算法输出值是在设备识别到人脸时输出的。
若是，则获取设备的当前识别角度，当前识别角度是指设备在通过人脸识别算法识别人脸时所使用的角度；将当前识别角度作为算法识别角度。
若否，则获取设备的当前识别角度，对当前识别角度进行更正，得到算法识别角度。
确定设备的显示窗口所显示的人脸是否倒置。
若是，则获取显示窗口的当前旋转角度，对当前旋转角度进行更正，得到显示旋转角度。
若否，则获取显示窗口的当前旋转角度，将当前旋转角度作为显示旋转角度。
根据人脸识别通过率为零、可调角度的设备的设备型号、算法识别角度以及显示旋转角度，生成零通过率机型配置表。
可选地，生成单元505具体可以用于：
预先收集多个设备。
对多个设备中的每个设备，获取设备对应的用户行为数据，用户行为数据是由客户端的埋点在捕获用户通过设备开始人脸识别行为以及完成人脸识别行为和/或放弃人脸识别行为时记录。
根据各个设备对应的用户行为数据，确定各个设备的人脸识别通过率。
从多个设备中选取人脸识别通过率为零的设备。
可选地，生成单元505还可以具体用于：
获取设备的传感器数据以及设备的摄像头所捕获的人脸的显示数据，传感器数据和显示数据是由客户端的埋点在捕获用户通过默认配置的设备识别到人脸时记录。
根据传感器数据以及显示数据，确定设备的显示窗口所显示的人脸是否倒置。
本说明书上述实施例装置的各功能模块的功能，可以通过上述方法实施例的各步骤来实现，因此，本说明书一个实施例提供的装置的具体工作过程，在此不复赘述。
本说明书一个实施例提供的人脸识别装置，接收单元501接收客户端的人脸识别请求。查找单元502根据设备型号，从零通过率机型配置表中查找该设备。若查找到该设备，获取单元503获取对应于该设备型号的角度配置信息。发送单元504向客户端返回角度配置信息，以使客户端根据角度配置信息，对设备进行配置，并通过配置后的 设备对人脸进行识别。由此，可以提高人脸识别的成功率。
需要说明的是，本说明书实施例提供的人脸识别装置可以为图1中服务器的一个模块或者单元。
与上述人脸识别方法对应地，本说明书一个实施例还提供的一种人脸识别装置，如图6所示，该装置包括：
发送单元601，用于向服务器发送人脸识别请求，该人脸识别请求包括人脸识别装置所在设备的设备型号；人脸识别请求用于指示服务器根据设备型号，从零通过率机型配置表中查找该设备。
接收单元602，用于接收服务器返回的响应结果。
配置单元603，用于若接收单元602接收的响应结果包括用于表示已查找到该设备的信息和对应的角度配置信息，则根据角度配置信息，对该设备进行配置。
识别单元604，用于通过配置单元603配置后的设备对人脸进行识别。
可选地，识别单元604，还用于若接收单元602接收的响应结果只包括用于表示已查找到设备的信息，则通过默认配置的设备对人脸进行识别。
可选地，该装置还可以包括：
获取单元605，用于若接收单元602接收的响应结果包括用于表示未查找到设备的信息，则获取设备的内置窗口的旋转角度。
确定单元606，用于根据获取单元605获取的内置窗口的旋转角度，确定该设备的角度配置信息。
配置单元603，还用于根据确定单元606确定的角度配置信息，对该设备进行配置。
识别单元604，还用于通过配置单元603配置后的设备对人脸进行识别。
本说明书上述实施例装置的各功能模块的功能，可以通过上述方法实施例的各步骤来实现，因此，本说明书一个实施例提供的装置的具体工作过程，在此不复赘述。
本说明书一个实施例提供的人脸识别装置，发送单元601向服务器发送人脸识别请求。接收单元602接收服务器返回的响应结果。若响应结果包括用于表示已查找到设备的信息和对应的角度配置信息，配置单元603根据角度配置信息，对设备进行配置。 识别单元604通过配置后的设备对人脸进行识别。由此，可以提高人脸识别的成功率。
需要说明的是，本说明书实施例提供的人脸识别装置可以为图1中客户端的一个模块或者单元。
本领域技术人员应该可以意识到，在上述一个或多个示例中，本说明书所描述的功能可以用硬件、软件、固件或它们的任意组合来实现。当使用软件实现时，可以将这些功能存储在计算机可读介质中或者作为计算机可读介质上的一个或多个指令或代码进行传输。
以上所述的具体实施方式，对本说明书的目的、技术方案和有益效果进行了进一步详细说明，所应理解的是，以上所述仅为本说明书的具体实施方式而已，并不用于限定本说明书的保护范围，凡在本说明书的技术方案的基础之上，所做的任何修改、等同替换、改进等，均应包括在本说明书的保护范围之内。

Claims (14)

  1. 一种人脸识别方法，包括：
    接收客户端的人脸识别请求，所述人脸识别请求包括所述客户端所在设备的设备型号；
    根据所述设备型号，从零通过率机型配置表中查找所述设备；所述零通过率机型配置表用于存储人脸识别通过率为零的设备的设备型号以及角度配置信息的对应关系；所述角度配置信息是根据所述客户端的埋点在捕获用户的人脸识别行为时所记录的数据确定的；
    若查找到所述设备，则获取对应于所述设备型号的角度配置信息；
    向所述客户端返回所述角度配置信息，以使所述客户端根据所述角度配置信息，对所述设备进行配置，并通过配置后的设备对人脸进行识别。
  2. 根据权利要求1所述的方法，所述角度配置信息包括：显示旋转角度和算法识别角度；所述方法还包括：生成所述零通过率机型配置表的步骤，包括：
    获取人脸识别通过率为零、可调角度的设备；
    确定所述设备是否有对应的算法输出值，所述算法输出值是在所述设备识别到人脸时输出的；
    若是，则获取所述设备的当前识别角度，当前识别角度是指所述设备在通过人脸识别算法识别人脸时所使用的角度；将当前识别角度作为所述算法识别角度；
    若否，则获取所述设备的当前识别角度，对当前识别角度进行更正，得到所述算法识别角度；
    确定所述设备的显示窗口所显示的人脸是否倒置；
    若是，则获取所述显示窗口的当前旋转角度，对当前旋转角度进行更正，得到所述显示旋转角度；
    若否，则获取所述显示窗口的当前旋转角度，将当前旋转角度作为所述显示旋转角度；
    根据所述人脸识别通过率为零、可调角度的设备的设备型号、所述算法识别角度以及所述显示旋转角度，生成所述零通过率机型配置表。
  3. 根据权利要求2所述的方法，所述获取人脸识别通过率为零的设备，包括：
    预先收集多个设备；
    对所述多个设备中的每个设备，获取所述设备对应的用户行为数据，所述用户行为数据是由客户端的埋点在捕获用户通过所述设备开始人脸识别行为以及完成人脸识别 行为和/或放弃人脸识别行为时记录；
    根据各个设备对应的用户行为数据，确定所述各个设备的人脸识别通过率；
    从所述多个设备中选取人脸识别通过率为零的设备。
  4. 根据权利要求2或3所述的方法，所述确定所述设备的显示窗口所显示的人脸是否倒置，包括：
    获取所述设备的传感器数据以及所述设备的摄像头所捕获的人脸的显示数据，所述传感器数据和所述显示数据是由客户端的埋点在捕获所述设备通过默认配置的设备识别到人脸时记录；
    根据所述传感器数据以及所述显示数据，确定所述显示窗口所显示的人脸是否倒置。
  5. 一种人脸识别方法，包括：
    向服务器发送人脸识别请求，所述人脸识别请求包括所述客户端所在设备的设备型号；所述人脸识别请求用于指示所述服务器根据所述设备型号，从零通过率机型配置表中查找所述设备；
    接收所述服务器返回的响应结果；
    若所述响应结果包括用于表示已查找到所述设备和对应的角度配置信息，则根据所述角度配置信息，对所述设备进行配置；
    通过配置后的设备对人脸进行识别。
  6. 根据权利要求5所述的方法，还包括：
    若所述响应结果只包括用于表示已查找到所述设备的信息，则通过默认配置的设备对人脸进行识别。
  7. 根据权利要求5或6所述的方法，还包括：
    若所述响应结果包括用于表示未查找到所述设备的信息，则获取所述设备的内置窗口的旋转角度；
    根据所述内置窗口的旋转角度，确定所述设备的角度配置信息；
    根据确定的角度配置信息，对所述设备进行配置；
    通过配置后的设备对人脸进行识别。
  8. 一种人脸识别装置，包括：
    接收单元，用于接收客户端的人脸识别请求，所述人脸识别请求包括所述客户端所在设备的设备型号；
    查找单元，用于根据所述接收单元接收的所述设备型号，从零通过率机型配置表中查找所述设备；所述零通过率机型配置表用于存储人脸识别通过率为零的设备的设备型 号以及角度配置信息的对应关系；所述角度配置信息是根据所述客户端的埋点在捕获用户的人脸识别行为时所记录的数据确定的；
    获取单元，用于若所述查找单元查找到所述设备，则获取对应于所述设备型号的角度配置信息；
    发送单元，用于向所述客户端返回所述获取单元获取的所述角度配置信息，以使所述客户端根据所述角度配置信息，对所述设备进行配置，并通过配置后的设备对人脸进行识别。
  9. 根据权利要求8所述的装置，所述角度配置信息包括：显示旋转角度和算法识别角度；所述装置还包括：生成单元；
    所述生成单元用于：
    获取人脸识别通过率为零、可调角度的设备；
    确定所述设备是否有对应的算法输出值，所述算法输出值是在所述设备识别到人脸时输出的；
    若是，则获取所述设备的当前识别角度，当前识别角度是指所述设备在通过人脸识别算法识别人脸时所使用的角度；将当前识别角度作为所述算法识别角度；
    若否，则获取所述设备的当前识别角度，对当前识别角度进行更正，得到所述算法识别角度；
    确定所述设备的显示窗口所显示的人脸是否倒置；
    若是，则获取所述显示窗口的当前旋转角度，对当前旋转角度进行更正，得到所述显示旋转角度；
    若否，则获取所述显示窗口的当前旋转角度，将当前旋转角度作为所述显示旋转角度；
    根据所述人脸识别通过率为零、可调角度的设备的设备型号、所述算法识别角度以及所述显示旋转角度，生成所述零通过率机型配置表。
  10. 根据权利要求9所述的装置，所述生成单元具体用于：
    预先收集多个设备；
    对所述多个设备中的每个设备，获取所述设备对应的用户行为数据，所述用户行为数据是由客户端的埋点在捕获用户通过所述设备开始人脸识别行为以及完成人脸识别行为和/或放弃人脸识别行为时记录；
    根据各个设备对应的用户行为数据，确定所述各个设备的人脸识别通过率；
    从所述多个设备中选取人脸识别通过率为零的设备。
  11. 根据权利要求9或10所述的装置，所述生成单元还具体用于：
    获取所述设备的传感器数据以及所述设备的摄像头所捕获的人脸的显示数据，所述传感器数据和所述显示数据是由客户端的埋点在捕获用户通过默认配置的设备识别到人脸时记录；
    根据所述传感器数据以及所述显示数据，确定所述显示窗口所显示的人脸是否倒置。
  12. 一种人脸识别装置，包括：
    发送单元，用于向服务器发送人脸识别请求，所述人脸识别请求包括所述人脸识别装置所在设备的设备型号；所述人脸识别请求用于指示所述服务器根据所述设备型号，从零通过率机型配置表中查找所述设备；
    接收单元，用于接收所述服务器返回的响应结果；
    配置单元，用于若所述接收单元接收的所述响应结果包括用于表示已查找到所述设备和对应的角度配置信息，则根据所述角度配置信息，对所述设备进行配置；
    识别单元，用于通过所述配置单元配置后的设备对人脸进行识别。
  13. 根据权利要求12所述的装置，
    所述识别单元，还用于若所述接收单元接收的所述响应结果只包括用于表示已查找到所述设备的信息，则通过默认配置的设备对人脸进行识别。
  14. 根据权利要求12或13所述的装置，还包括：
    获取单元，用于若所述接收单元接收的所述响应结果包括用于表示未查找到所述设备的信息，则获取所述设备的内置窗口的旋转角度；
    确定单元，用于根据所述获取单元获取的所述内置窗口的旋转角度，确定所述设备的角度配置信息；
    所述配置单元，还用于根据所述确定单元确定的角度配置信息，对所述设备进行配置；
    所述识别单元，还用于通过所述配置单元配置后的设备对人脸进行识别。
PCT/CN2019/073829 2018-04-04 2019-01-30 人脸识别方法及装置 Ceased WO2019192256A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/959,642 US20210019543A1 (en) 2018-04-04 2019-01-30 Method and apparatus for face recognition

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810301816.6 2018-04-04
CN201810301816.6A CN110363048B (zh) 2018-04-04 2018-04-04 人脸识别方法及装置

Publications (1)

Publication Number Publication Date
WO2019192256A1 true WO2019192256A1 (zh) 2019-10-10

Family

ID=68099925

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/073829 Ceased WO2019192256A1 (zh) 2018-04-04 2019-01-30 人脸识别方法及装置

Country Status (4)

Country Link
US (1) US20210019543A1 (zh)
CN (1) CN110363048B (zh)
TW (1) TWI709914B (zh)
WO (1) WO2019192256A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113139809A (zh) * 2021-04-30 2021-07-20 中国银行股份有限公司 手机银行信息保护方法及装置

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070041644A1 (en) * 2005-08-17 2007-02-22 Samsung Electronics Co., Ltd. Apparatus and method for estimating a facial pose and a face recognition system using the method
CN202120280U (zh) * 2011-06-30 2012-01-18 汉王科技股份有限公司 一种可调节角度的人脸识别设备
CN102930257A (zh) * 2012-11-14 2013-02-13 汉王科技股份有限公司 人脸识别装置
CN104853096A (zh) * 2015-04-30 2015-08-19 广东欧珀移动通信有限公司 一种基于旋转摄像头的拍摄参数的确定方法及终端

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5262493B2 (ja) * 2008-09-17 2013-08-14 株式会社リコー 画像処理装置、画像処理方法、プログラムおよびコンピュータ読み取り可能な記憶媒体
KR101760258B1 (ko) * 2010-12-21 2017-07-21 삼성전자주식회사 얼굴 인식 장치 및 그 방법
CN104036237B (zh) * 2014-05-28 2017-10-10 中国人民解放军海军总医院 基于在线预测的旋转人脸的检测方法
KR101718244B1 (ko) * 2015-12-14 2017-03-20 이우균 얼굴 인식을 위한 광각 영상 처리 장치 및 방법
CN105760836A (zh) * 2016-02-17 2016-07-13 厦门美图之家科技有限公司 基于深度学习的多角度人脸对齐方法、系统及拍摄终端
CN107229892A (zh) * 2016-03-24 2017-10-03 阿里巴巴集团控股有限公司 一种基于人脸识别产品的识别调整方法和设备
CN107370713A (zh) * 2016-05-11 2017-11-21 中兴通讯股份有限公司 一种人脸识别的登录认证方法、装置、终端及服务器
CN107563337A (zh) * 2017-09-12 2018-01-09 广东欧珀移动通信有限公司 人脸识别的方法及相关产品
CN107832730A (zh) * 2017-11-23 2018-03-23 高域（北京）智能科技研究院有限公司 提高人脸识别准确率的方法和人脸识别系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070041644A1 (en) * 2005-08-17 2007-02-22 Samsung Electronics Co., Ltd. Apparatus and method for estimating a facial pose and a face recognition system using the method
CN202120280U (zh) * 2011-06-30 2012-01-18 汉王科技股份有限公司 一种可调节角度的人脸识别设备
CN102930257A (zh) * 2012-11-14 2013-02-13 汉王科技股份有限公司 人脸识别装置
CN104853096A (zh) * 2015-04-30 2015-08-19 广东欧珀移动通信有限公司 一种基于旋转摄像头的拍摄参数的确定方法及终端

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
"Upload a photo preview on the mobile side (solve a bug where the photo of the device such as iOS is rotated by 90 degrees)", 8 September 2016 (2016-09-08), pages 3 - 6, XP055642004, Retrieved from the Internet <URL:https://blog.csdn.net/hsany330/article/details/52471522> *

Also Published As

Publication number Publication date
US20210019543A1 (en) 2021-01-21
CN110363048B (zh) 2020-04-24
CN110363048A (zh) 2019-10-22
TW202001666A (zh) 2020-01-01
TWI709914B (zh) 2020-11-11

Similar Documents

Publication Publication Date Title
US11145130B2 (en) Method for automatically capturing data from non-networked production equipment
JP6199507B2 (ja) 掃除方法、装置、プログラム及び記録媒体
US9697415B2 (en) Recording medium, image processing method, and information terminal
JP4951498B2 (ja) 顔画像認識装置、顔画像認識方法、顔画像認識プログラムおよびそのプログラムを記録した記録媒体
WO2018228218A1 (zh) 身份识别方法、计算设备及存储介质
WO2019119976A1 (zh) 储物终端取件方法及装置和电子设备
US12374062B2 (en) Method for automatically capturing data from non-networked production equipment
CN113033266A (zh) 人员运动轨迹追踪方法、装置、系统及电子设备
CN108875506B (zh) 人脸形状点跟踪方法、装置和系统及存储介质
CN205540781U (zh) 身份的识别系统
CN111354029A (zh) 手势深度确定方法、装置、设备及存储介质
CN108596127B (zh) 一种指纹识别方法、身份验证方法及装置和身份核验机
JP6829348B1 (ja) 撮影制御方法、情報処理装置、プログラム、及び撮影システム
WO2019192256A1 (zh) 人脸识别方法及装置
JP2015011404A (ja) 動作認識処理装置
CN116092176A (zh) 用于检测手势的方法和系统，以及计算机可读存储媒体
CN111368675B (zh) 手势深度信息的处理方法、装置、设备及存储介质
WO2016004765A1 (zh) 用于进行多点视频通讯的方法、设备及系统
CN116363401B (zh) 图像匹配方法、装置、电子设备、芯片
CN108834076B (zh) 目标设备寻找方法、装置及设备
JP6922768B2 (ja) 情報処理装置
HK40016247A (zh) 人脸识别方法及装置
HK40016247B (zh) 人脸识别方法及装置
CN110312069A (zh) 拍摄过程中的对焦方法及装置
CN114511891A (zh) 人脸识别处理方法、装置、终端及可读存储介质

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: 19781811

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19781811

Country of ref document: EP

Kind code of ref document: A1