WO2019033576A1 - 人脸姿态检测方法、装置及存储介质 - Google Patents

人脸姿态检测方法、装置及存储介质 Download PDF

Info

Publication number
WO2019033576A1
WO2019033576A1 PCT/CN2017/108803 CN2017108803W WO2019033576A1 WO 2019033576 A1 WO2019033576 A1 WO 2019033576A1 CN 2017108803 W CN2017108803 W CN 2017108803W WO 2019033576 A1 WO2019033576 A1 WO 2019033576A1
Authority
WO
WIPO (PCT)
Prior art keywords
face
feature point
angle
distance
axis
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/CN2017/108803
Other languages
English (en)
French (fr)
Inventor
陈林
张国辉
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.)
Ping An Technology Shenzhen Co Ltd
Original Assignee
Ping An Technology Shenzhen Co 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 Ping An Technology Shenzhen Co Ltd filed Critical Ping An Technology Shenzhen Co Ltd
Priority to JP2019556821A priority Critical patent/JP6815707B2/ja
Priority to SG11202001403XA priority patent/SG11202001403XA/en
Priority to US16/632,371 priority patent/US11062124B2/en
Publication of WO2019033576A1 publication Critical patent/WO2019033576A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • 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
    • G06V40/165—Detection; Localisation; Normalisation using facial parts and geometric relationships
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30196—Human being; Person
    • G06T2207/30201—Face
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30244—Camera pose

Definitions

  • the present application relates to the field of computer vision processing technologies, and in particular, to a face gesture detection method, apparatus, and storage medium.
  • Computer vision processing technology began in the early 1960s and includes multiple research directions such as detection, tracking, and recognition. With the rapid development of computer and network technology, computer vision processing technology has been widely used in many important fields such as industrial inspection, security monitoring, clinical diagnosis and treatment, transportation, entertainment, and military. Among them, the computer face gesture detection technology can be used for intelligent systems such as assisted driving and human-computer interaction, and has broad application prospects.
  • the existing face attitude detection system mostly uses depth learning to directly train face sample data of different poses, and then uses the trained model to judge the face pose, but this detection method requires a large number of samples and accuracy. Poor, and can only detect the trained posture, can not detect all face gestures.
  • the present invention provides a face gesture detection method, device and storage medium, the main purpose of which is to calculate face pose information according to coordinate values of key feature points of a human face, improve the speed and accuracy of face gesture detection, and realize arbitrary angles. Face gesture detection.
  • the present application provides a method for detecting a face gesture, the method comprising:
  • Face feature point extraction step extracting N face feature points from the face image by a face detection algorithm
  • a key feature point extraction step extracting, from the N facial feature points, a first feature point near the right ear, a second feature point near the right ear, and a third feature near the lower ear Pointing, near the fourth feature point above the left ear, and extracting the nose point feature point as the fifth feature point;
  • a first calculating step forming the fifth feature point and the first feature point into a first vector, the fifth feature point and the fourth feature point forming a second vector, and calculating a first angle between the first vector and the second vector, Determining, according to a change of the first angle and the initial angle, a rotation angle of the face in the face image about the X axis;
  • a second calculating step forming the first feature point and the fourth feature point into a third vector, and extracting a line L1 parallel to the X axis from the fourth feature point, and calculating a second angle between the third vector and the straight line L1, Determining, according to the second angle, a rotation angle of the face in the face image about the Y axis;
  • a third calculating step the distance between the second feature point and the fifth feature point is recorded as a first distance, and the distance between the third feature point and the fifth feature point is recorded as a second distance, and the first distance and the second distance are calculated of The ratio determines a rotation direction of the face in the face image about the Z axis according to a ratio of the first distance to the second distance.
  • the present application further provides a computing device including an image capturing device, a memory, and a processor, wherein the memory includes a face gesture detecting program, and the face pose detecting program is executed by the processor:
  • Face feature point extraction step extracting N face feature points from the face image by a face detection algorithm
  • a key feature point extraction step extracting, from the N facial feature points, a first feature point near the right ear, a second feature point near the right ear, and a third feature near the lower ear Pointing, near the fourth feature point above the left ear, and extracting the nose point feature point as the fifth feature point;
  • a first calculating step forming the fifth feature point and the first feature point into a first vector, the fifth feature point and the fourth feature point forming a second vector, and calculating a first angle between the first vector and the second vector, Determining, according to a change of the first angle and the initial angle, a rotation angle of the face in the face image about the X axis;
  • a second calculating step forming the first feature point and the fourth feature point into a third vector, and extracting a line L1 parallel to the X axis from the fourth feature point, and calculating a second angle between the third vector and the straight line L1, Determining, according to the second angle, a rotation angle of the face in the face image about the Y axis;
  • a third calculating step the distance between the second feature point and the fifth feature point is recorded as a first distance, and the distance between the third feature point and the fifth feature point is recorded as a second distance, and the first distance and the second distance are calculated The ratio of the first distance to the second distance determines the direction of rotation of the face in the face image about the Z axis.
  • the present application further provides a computer readable storage medium including a face gesture detection program.
  • a face pose detection program When the face pose detection program is executed by the processor, any step in the face pose detection method as described above is implemented.
  • the face gesture detection method, device and computer readable storage medium proposed by the present application, by extracting key feature points of a human face, calculating posture information such as a face rotation angle according to coordinate values of the key feature points, and improving face gesture detection Speed and accuracy for face detection at any angle.
  • FIG. 1 is a schematic diagram of an operating environment of a preferred embodiment of a computing device of the present application
  • FIG. 2 is a functional block diagram of the face pose detection program of FIG. 1;
  • FIG. 3 is a schematic diagram of a face feature point extracted by the extraction module of FIG. 2;
  • FIG. 4 is a flow chart of a preferred embodiment of the applicant's face gesture detecting method
  • FIG. 5 is a schematic diagram of feature points, vectors, and angles involved in step S30 of FIG. 4; FIG.
  • FIG. 6 is a schematic diagram of feature points, vectors, straight lines, and included angles in step S40 of FIG. 4;
  • FIG. 7 is a schematic diagram of a feature point and a distance between feature points involved in step S50 in FIG. 4;
  • FIG. 1 it is a schematic diagram of an operating environment of a preferred embodiment of the computing device 1 of the present application.
  • the computing device 1 may be a terminal device having a computing function such as a server, a smart phone, a tablet computer, a portable computer, or a desktop computer.
  • the computing device 1 includes an imaging device 11, a memory 12, a processor 13, a network interface 14, and a communication bus 15.
  • the camera device 11 is installed in a specific place, such as an office space, a monitoring area, and a smart car cab.
  • a real-time image is captured in real time for a target entering the specific place, and the captured real-time image is transmitted to the processor 13 through the network.
  • Network interface 14 may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface).
  • Communication bus 15 is used to implement connection communication between these components.
  • Memory 12 includes at least one type of readable storage medium.
  • the at least one type of readable storage medium may be a non-volatile storage medium such as a flash memory, a hard disk, a multimedia card, a card type memory, or the like.
  • the readable storage medium may be an internal storage unit of the computing device 1, such as a hard disk of the computing device 1.
  • the readable storage medium may also be an external memory of the computing device 1, such as a plug-in hard disk (Smart Media Card, SMC) equipped on the computing device 1, Secure Digital (SD) card, Flash Card, etc.
  • Smart Media Card Smart Media Card, SMC
  • SD Secure Digital
  • the readable storage medium of the memory 12 is generally used to store the face pose detection program 10 installed in the computing device 1, the real-time image obtained by the camera device 11, the face image, the initial data, and the like.
  • the memory 12 can also be used to temporarily store data that has been output or is about to be output.
  • the processor 13 may be a Central Processing Unit (CPU), microprocessor or other data processing chip for running program code or processing data stored in the memory 12, such as executing a human face.
  • the attitude detection program 10 and the like.
  • Figure 1 shows only computing device 1 having components 11-15 and face pose detection program 10, but it should be understood that not all illustrated components may be implemented and that more or fewer components may be implemented instead. .
  • the computing device 1 may further include a user interface
  • the user interface may include an input unit such as a keyboard, a voice input device such as a microphone, a device with a voice recognition function, a voice output device such as an audio, a headphone, and the like.
  • the user interface may also include a standard wired interface and a wireless interface.
  • the computing device 1 may also include a display, which may also be referred to as a display screen or display unit as appropriate.
  • a display may also be referred to as a display screen or display unit as appropriate.
  • it may be an LED display, a liquid crystal display, or a touch Liquid crystal display, OLED (Organic Light-Emitting Diode) touch sensor, and the like.
  • the display is used to display information processed in the computing device 1 and a user interface for displaying visualizations.
  • the computing device 1 further comprises a touch sensor.
  • the area provided by the touch sensor for the user to perform a touch operation is referred to as a touch area.
  • the touch sensor described herein may be a resistive touch sensor, a capacitive touch sensor, or the like.
  • the touch sensor includes not only a contact type touch sensor but also a proximity type touch sensor or the like.
  • the touch sensor may be a single sensor or a plurality of sensors arranged, for example, in an array.
  • the area of the display of the computing device 1 may be the same as or different from the area of the touch sensor.
  • a display is stacked with the touch sensor to form a touch display.
  • the computing device detects a user-triggered touch operation based on the touch display.
  • the computing device 1 may further include an RF (Radio Frequency) circuit, a sensor, an audio circuit, and the like, and details are not described herein.
  • RF Radio Frequency
  • the memory 12 as a computer readable storage medium may include data such as an operating system, a face gesture detecting program 10, a real-time image obtained by the camera 11, and a face image.
  • the processor 13 executes the face gesture detecting program 10 stored in the memory 12, the following steps are performed:
  • a key feature point extraction step extracting, from the N facial feature points, a first feature point P1 near the right ear, a second feature point P3 near the right ear, and a part near the left ear a third feature point P15, a fourth feature point P17 near the left ear, and extracting a nose point feature point as a fifth feature point P31;
  • a first calculating step forming the fifth feature point P31 and the first feature point P1 into a first vector, and the fifth feature point P31 and the fourth feature point P17 form a second vector (refer to FIG. 5), and calculating the first a first angle between the vector and the second vector, determining a rotation angle of the face in the face image about the X axis according to the change of the first angle and the initial angle;
  • a second calculating step forming the first feature point P1 and the fourth feature point P17 into a third vector, and extracting a line L1 parallel to the X axis from the fourth feature point P17 (refer to FIG. 6), and calculating the third a second angle between the vector and the straight line L1, and determining, according to the second angle, a rotation angle of the face in the face image about the Y axis;
  • a third calculating step the distance between the second feature point P3 and the fifth feature point P31 is recorded as a first distance, and the distance between the third feature point P15 and the fifth feature point P31 is recorded as a second distance (refer to FIG. 7
  • the ratio of the first distance to the second distance is calculated, and the rotation direction of the face around the Z axis in the face image is determined according to the ratio of the first distance to the second distance.
  • the first calculating step includes:
  • the first angle is greater than 120°, it indicates that the face in the face image rotates around the X axis, and the rotation angle is the difference between the first angle and the initial angle;
  • the first angle is less than 120°, it indicates that the face in the face image is rotated downward about the X axis, and the rotation angle is the difference between the initial angle and the first angle;
  • the first angle is equal to 120°, it indicates that the face in the face image is in an initial position with respect to the X axis, that is, not raised or bowed.
  • the second calculating step includes:
  • the second angle When the second angle is 0°, it indicates that the face in the face image is in an initial position with respect to the Y axis, that is, the head is not left or right;
  • the third calculating step includes:
  • the ratio of the first distance to the second distance is less than 1, it indicates that the face in the face image is rotated right around the Z axis.
  • the face detection algorithm is a dlib face feature point detection algorithm.
  • 68 feature points are extracted from the face image by using the dlib face feature point detection algorithm, including 17 facial contour feature points, 10 eyebrow feature points, and 4 nose wing feature points, including the tip of the nose.
  • the facial contour feature points the first feature point near the right ear is symmetric with the fourth feature point near the left ear, and the second feature point near the right ear is symmetric with the third feature point near the left ear.
  • the X, Y, and Z coordinate axes are defined as: the face is considered to be a cylinder, and the central axis of the cylinder is used as the vertical coordinate axis Z axis of the face data, and the nose point feature point is taken and the cylindrical center axis is taken.
  • a straight line intersecting perpendicularly is taken as the Y axis, and a cross direction of the Z axis and the Y axis is taken as the X axis.
  • the face gesture detection program 10 can also be partitioned into one or more modules that are stored in the memory 12 and executed by the processor 13 to complete the application.
  • a module as referred to in this application refers to a series of computer program instructions that are capable of performing a particular function.
  • FIG. 2 it is a functional block diagram of the face pose detecting program 10 of Fig. 1.
  • the face gesture detection program 10 can be divided into an acquisition module 110, an extraction module 120, and a calculation module 130.
  • the obtaining module 110 is configured to acquire a real-time image captured by the camera 11 and extract a real-time facial image from the real-time image by using a face recognition algorithm.
  • the camera 11 captures a real-time image
  • the camera 11 transmits the real-time image to the processor 13.
  • the acquisition module 110 extracts the real-time using the face recognition algorithm. Face image.
  • the face recognition algorithm for extracting the real-time facial image from the real-time image may be a geometric feature-based method, a local feature analysis method, a feature face method, an elastic model-based method, a neural network method, or the like.
  • the extracting module 120 is configured to extract the facial feature points and the key feature points of the real-time facial image.
  • the extraction module 120 extracts 68 human face feature points P1-P68 in the real-time facial image by the dlib face feature point detection algorithm, as shown in FIG. Further, five key feature points are extracted from the 68 facial feature points: a first feature point P1 near the right ear, a second feature point P3 near the right ear, and a third feature near the left ear. Point P15, near the fourth feature point P17 above the left ear, and the fifth feature point P31 at the tip of the nose.
  • the calculation module 130 is configured to calculate a face rotation direction and an angle in the real-time face image according to coordinate values of five key feature points in the real-time face image.
  • the calculation module 130 is configured to:
  • the computing device 1 proposed in this embodiment extracts a real-time facial image from a real-time image, extracts a facial feature point in the real-time facial image by using a dlib facial feature point detection algorithm, and further extracts a key feature point, and then extracts a key feature point, and then The face rotation direction and angle in the face image are calculated according to the coordinate values of the key feature points, realizing the real-time detection of the face pose.
  • the present application also provides a face gesture detection method.
  • FIG. 4 it is a flowchart of a preferred embodiment of the applicant's face gesture detection method.
  • the method can be performed by a computing device, which can be implemented by software and/or hardware.
  • the face gesture detection method includes:
  • the extraction module 120 extracts N facial feature points from the facial image by using a face detection algorithm.
  • the camera 11 captures a real-time image and transmits the real-time image to the processor 13.
  • the processor 13 extracts a real-time facial image by using a face recognition algorithm, and then acquires N through a face detection algorithm.
  • personal face feature points extracting the real image from the real image
  • the face recognition algorithm of the face image may be a geometric feature based method, a local feature analysis method, a feature face method, an elastic model based method, a neural network method, or the like.
  • the face detection algorithm for extracting facial feature points from the face image is a dlib face feature point detection algorithm.
  • step S20 the extraction module 120 extracts key feature points in the feature points of the face image.
  • the extraction module 120 further extracts a first feature point P1 near the right ear, a second feature point P3 near the right ear, a third feature point P15 near the left ear, and a fourth feature point P17 near the left ear.
  • the nose tip feature point is extracted as the fifth feature point P31.
  • step S30 the calculation module 130 calculates a rotation angle of the face in the face image about the X axis.
  • the fifth feature point P31 and the first feature point P1 form a first vector
  • the fifth feature point P31 and the fourth feature point P17 form a second vector
  • the first vector and the second vector are calculated.
  • a first angle determining a rotation angle of the face in the face image about the X axis according to the change of the first angle and the initial angle;
  • step S40 the calculation module 130 calculates a rotation angle of the face in the face image around the Y axis.
  • the first feature point P1 and the fourth feature point P17 form a third vector
  • a line L1 parallel to the X axis is extracted from the fourth feature point P17, and the third vector and the line L1 are calculated.
  • the second angle determines the rotation angle of the face in the face image about the Y axis according to the second angle.
  • step S50 the calculation module 130 calculates a rotation direction of the face in the face image about the Z axis.
  • the distance between the second feature point P3 and the fifth feature point P31 is recorded as a first distance
  • the distance between the third feature point P15 and the fifth feature point P31 is recorded as a second distance
  • the first calculation is performed.
  • the ratio of the distance to the second distance determines the direction of rotation of the face about the Z axis based on the ratio of the first distance to the second distance.
  • the face pose detection method proposed in this embodiment uses the dlib face feature point detection algorithm to acquire the face feature points in the real-time face image, and further extracts the key feature points, and calculates the real-time according to the coordinate values of the key feature points.
  • the face rotation direction and angle in the face image realize real-time detection of the face pose.
  • the embodiment of the present application further provides a computer readable storage medium, where the computer readable storage medium includes a face gesture detection program, and when the face gesture detection program is executed by the processor, the following operations are implemented:
  • a key feature point extraction step extracting, from the N facial feature points, a first feature point P1 near the right ear, a second feature point P3 near the right ear, and a part near the left ear a third feature point P15, a fourth feature point P17 near the left ear, and extracting a nose point feature point as a fifth feature point P31;
  • a first calculating step forming the fifth feature point P31 and the first feature point P1 into a first vector, and the fifth feature point P31 and the fourth feature point P17 form a second vector (refer to FIG. 5), and calculating the first The first angle between the vector and the second vector, and the face map is determined according to the change of the first angle and the initial angle The angle of rotation of the human face around the X axis;
  • a second calculating step forming the first feature point P1 and the fourth feature point P17 into a third vector, and extracting a line L1 parallel to the X axis from the fourth feature point P17 (refer to FIG. 6), and calculating the third a second angle between the vector and the straight line L1, and determining, according to the second angle, a rotation angle of the face in the face image about the Y axis;
  • a third calculating step the distance between the second feature point P3 and the fifth feature point P31 is recorded as a first distance, and the distance between the third feature point P15 and the fifth feature point P31 is recorded as a second distance (refer to FIG. 7
  • the ratio of the first distance to the second distance is calculated, and the rotation direction of the face around the Z axis in the face image is determined according to the ratio of the first distance to the second distance.
  • the first calculating step includes:
  • the first angle is greater than 120°, it indicates that the face in the face image rotates around the X axis, and the rotation angle is the difference between the first angle and the initial angle;
  • the first angle is less than 120°, it indicates that the face in the face image is rotated downward about the X axis, and the rotation angle is the difference between the initial angle and the first angle;
  • the first angle is equal to 120°, it indicates that the face in the face image is in an initial position with respect to the X axis, that is, not raised or bowed.
  • the second calculating step includes:
  • the second angle When the second angle is 0°, it indicates that the face in the face image is in an initial position with respect to the Y axis, that is, the head is not left or right;
  • the third calculating step includes:
  • the ratio of the first distance to the second distance is less than 1, it indicates that the face in the face image is rotated right around the Z axis.
  • a disk including a number of instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the methods described in the various embodiments of the present application.
  • a terminal device which may be a mobile phone, a computer, a server, or a network device, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Computer Interaction (AREA)
  • Multimedia (AREA)
  • Geometry (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

一种人脸姿态检测方法，该方法包括：通过人脸检测算法从人脸图像中提取N个人脸特征点；从所述N个人脸特征点中提取关键特征点；根据所述关键特征点的坐标值计算人脸绕坐标轴的旋转方向和角度等人脸姿态信息。根据人脸特征点的坐标值计算人脸图像中人脸的姿态信息，实现了对人脸姿态的实时检测。相应的，还提出一种计算装置以及一种计算机可读存储介质。

Description

人脸姿态检测方法、装置及存储介质
本专利申请以2017年8月17日提交的申请号为201710707941.2，名称为“人脸姿态检测方法、装置及存储介质”的中国发明专利申请为基础，并要求其优先权。
技术领域
本申请涉及计算机视觉处理技术领域，尤其涉及一种人脸姿态检测方法、装置及存储介质。
背景技术
计算机视觉处理技术始于20世纪60年代初，包含检测、跟踪、识别等多个研究方向。随着计算机和网络技术的飞速发展，计算机视觉处理技术已广泛地应用于工业检测、安全监控、临床诊治、交通、娱乐、军事等诸多重要领域。其中，计算机人脸姿态检测技术可用于辅助驾驶、人机交互等智能系统，具有广阔的应用前景。
现有的人脸姿态检测系统多是采用深度学习直接对不同姿态的人脸样本数据进行训练，然后使用训练后的模型对人脸姿态进行判断，但这种检测方法需要大量样本、精确度较差，且只能对已训练的姿态进行检测，无法检测所有的人脸姿态。
发明内容
本申请提供一种人脸姿态检测方法、装置及存储介质，其主要目的在于根据人体脸部关键特征点的坐标值计算人脸姿态信息，提高人脸姿态检测的速度和精度，实现任意角度的人脸姿态检测。
为实现上述目的，本申请提供一种人脸姿态检测方法，该方法包括：
人脸特征点提取步骤：通过人脸检测算法从人脸图像中提取N个人脸特征点；
关键特征点提取步骤：从所述N个人脸特征点中提取脸部轮廓特征点中靠近右耳上方的第一特征点、靠近右耳下方的第二特征点、靠近左耳下方的第三特征点、靠近左耳上方的第四特征点，并提取鼻尖特征点为第五特征点；
第一计算步骤：将所述第五特征点与第一特征点组成第一向量，第五特征点与第四特征点组成第二向量，计算第一向量与第二向量的第一夹角，根据第一夹角与初始夹角的变化判断所述人脸图像中人脸绕X轴的旋转角度；
第二计算步骤：将所述第一特征点与第四特征点组成第三向量，从第四特征点引出一条与X轴平行的直线L1，计算第三向量与直线L1的第二夹角，根据该第二夹角判断所述人脸图像中人脸绕Y轴的旋转角度；
第三计算步骤：将所述第二特征点与第五特征点的距离记为第一距离，第三特征点与第五特征点的距离记为第二距离，计算第一距离与第二距离的 比值，根据第一距离与第二距离的比值判断所述人脸图像中人脸绕Z轴的旋转方向。
为实现上述目的，本申请还提供一种计算装置，该装置包括摄像装置、存储器和处理器，所述存储器中包括人脸姿态检测程序，所述人脸姿态检测程序被所述处理器执行时遵循如下步骤：
人脸特征点提取步骤：通过人脸检测算法从人脸图像中提取N个人脸特征点；
关键特征点提取步骤：从所述N个人脸特征点中提取脸部轮廓特征点中靠近右耳上方的第一特征点、靠近右耳下方的第二特征点、靠近左耳下方的第三特征点、靠近左耳上方的第四特征点，并提取鼻尖特征点为第五特征点；
第一计算步骤：将所述第五特征点与第一特征点组成第一向量，第五特征点与第四特征点组成第二向量，计算第一向量与第二向量的第一夹角，根据第一夹角与初始夹角的变化判断所述人脸图像中人脸绕X轴的旋转角度；
第二计算步骤：将所述第一特征点与第四特征点组成第三向量，从第四特征点引出一条与X轴平行的直线L1，计算第三向量与直线L1的第二夹角，根据该第二夹角判断所述人脸图像中人脸绕Y轴的旋转角度；
第三计算步骤：将所述第二特征点与第五特征点的距离记为第一距离，第三特征点与第五特征点的距离记为第二距离，计算第一距离与第二距离的比值，根据第一距离与第二距离的比值判断所述人脸图像中人脸绕Z轴的旋转方向。
此外，为实现上述目的，本申请还提供一种计算机可读存储介质，所述计算机可读存储介质中包括人脸姿态检测程序。所述人脸姿态检测程序被所述处理器执行时，实现如上所述的人脸姿态检测方法中的任意步骤。
本申请提出的人脸姿态检测方法、装置及计算机可读存储介质，通过提取人体面部关键特征点，根据所述关键特征点的坐标值计算人脸旋转角度等姿态信息，提高人脸姿态检测的速度和精度，实现任意角度的人脸姿态检测。
附图说明
图1为本申请计算装置较佳实施例的运行环境示意图；
图2为图1中人脸姿态检测程序的功能模块图；
图3为图2中提取模块提取的人脸特征点的示意图；
图4为本申请人脸姿态检测方法较佳实施例的流程图；
图5为图4中步骤S30涉及的特征点、向量和夹角的示意图；
图6为图4中步骤S40涉及的特征点、向量、直线和夹角的示意图；
图7为图4中步骤S50涉及的特征点及特征点之间距离的示意图；
本申请目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。
具体实施方式
应当理解，此处所描述的具体实施例仅仅用以解释本申请，并不用于限定本申请。
参照图1所示，为本申请计算装置1较佳实施例的运行环境示意图。
该计算装置1可以是服务器、智能手机、平板电脑、便携计算机、桌上型计算机等具有运算功能的终端设备。
参照图1所示，该计算装置1包括摄像装置11、存储器12、处理器13、网络接口14及通信总线15。其中，摄像装置11安装于特定场所，如办公场所、监控区域、智能汽车驾驶室，对进入该特定场所的目标实时拍摄得到实时图像，通过网络将拍摄得到的实时图像传输至处理器13。网络接口14可选地可以包括标准的有线接口、无线接口(如WI-FI接口)。通信总线15用于实现这些组件之间的连接通信。
存储器12包括至少一种类型的可读存储介质。所述至少一种类型的可读存储介质可为如闪存、硬盘、多媒体卡、卡型存储器等的非易失性存储介质。在一些实施例中，所述可读存储介质可以是所述计算装置1的内部存储单元，例如该计算装置1的硬盘。在另一些实施例中，所述可读存储介质也可以是所述计算装置1的外部存储器，例如所述计算装置1上配备的插接式硬盘，智能存储卡(Smart Media Card,SMC)，安全数字(Secure Digital,SD)卡，闪存卡(Flash Card)等。
在本实施例中，所述存储器12的可读存储介质通常用于存储安装于所述计算装置1的人脸姿态检测程序10、摄像装置11获得的实时图像、人脸图像及初始数据等。所述存储器12还可以用于暂时地存储已经输出或者将要输出的数据。
处理器13，在一些实施例中可以是一中央处理器(Central Processing Unit,CPU)，微处理器或其他数据处理芯片，用于运行存储器12中存储的程序代码或处理数据，例如执行人脸姿态检测程序10等。
图1仅示出了具有组件11-15以及人脸姿态检测程序10的计算装置1，但是应理解的是，并不要求实施所有示出的组件，可以替代的实施更多或者更少的组件。
可选地，该计算装置1还可以包括用户接口，用户接口可以包括输入单元比如键盘(Keyboard)、语音输入装置比如麦克风(microphone)等具有语音识别功能的设备、语音输出装置比如音响、耳机等，可选地用户接口还可以包括标准的有线接口、无线接口。
可选地，所述计算装置1还可以包括显示器，显示器也可以适当的称为显示屏或显示单元。在一些实施例中可以是LED显示器、液晶显示器、触控 式液晶显示器以及OLED(Organic Light-Emitting Diode，有机发光二极管)触摸器等。显示器用于显示在计算装置1中处理的信息以及用于显示可视化的用户界面。
可选地，该计算装置1还包括触摸传感器。所述触摸传感器所提供的供用户进行触摸操作的区域称为触控区域。此外，这里所述的触摸传感器可以为电阻式触摸传感器、电容式触摸传感器等。而且，所述触摸传感器不仅包括接触式的触摸传感器，也可包括接近式的触摸传感器等。此外，所述触摸传感器可以为单个传感器，也可以为例如阵列布置的多个传感器。
此外，该计算装置1的显示器的面积可以与所述触摸传感器的面积相同，也可以不同。可选地，将显示器与所述触摸传感器层叠设置，以形成触摸显示屏。该计算装置基于触摸显示屏侦测用户触发的触控操作。
可选地，该计算装置1还可以包括RF(Radio Frequency，射频)电路，传感器、音频电路等等，在此不再赘述。
作为一种计算机可读存储介质的存储器12，可以包括操作系统、人脸姿态检测程序10、摄像装置11获得的实时图像和人脸图像等数据。处理器13执行存储器12中存储的人脸姿态检测程序10时执行如下步骤：
人脸特征点提取步骤：通过人脸检测算法从人脸图像中提取N个人脸特征点。如图3所示，当应用dlib人脸特征点检测算法时，可以从人脸图像中提取人脸特征点P1-P68(N＝68)；
关键特征点提取步骤：从所述N个人脸特征点中提取脸部轮廓特征点中靠近右耳上方的第一特征点P1、靠近右耳下方的第二特征点P3、靠近左耳下方的第三特征点P15、靠近左耳上方的第四特征点P17，并提取鼻尖特征点为第五特征点P31；
第一计算步骤：将所述第五特征点P31与第一特征点P1组成第一向量，第五特征点P31与第四特征点P17组成第二向量(参照图5所示)，计算第一向量与第二向量的第一夹角，根据第一夹角与初始夹角的变化判断所述人脸图像中人脸绕X轴的旋转角度；
第二计算步骤：将所述第一特征点P1与第四特征点P17组成第三向量，从第四特征点P17引出一条与X轴平行的直线L1(参照图6所示)，计算第三向量与直线L1的第二夹角，根据该第二夹角判断所述人脸图像中人脸绕Y轴的旋转角度；
第三计算步骤：将所述第二特征点P3与第五特征点P31的距离记为第一距离，第三特征点P15与第五特征点P31的距离记为第二距离(参照图7所示)，计算第一距离与第二距离的比值，根据第一距离与第二距离的比值判断所述人脸图像中人脸绕Z轴的旋转方向。
其中，所述第一计算步骤包括：
计算人脸在初始位置时所述第一向量与第二向量的初始夹角，可选地， 若该初始夹角为120°，则：
当所述第一夹角大于120°时，表明所述人脸图像中人脸绕X轴向上旋转，旋转角度为第一夹角与初始夹角之差；
当所述第一夹角小于120°时，表明所述人脸图像中人脸绕X轴向下旋转，旋转角度为初始夹角与第一夹角之差；
当所述第一夹角等于120°时，表明所述人脸图像中人脸相对于X轴处于初始位置，即未抬头或低头。
所述第二计算步骤包括：
当所述第二夹角为0°时，表明所述人脸图像中人脸相对于Y轴处于初始位置，即未向左或向右歪头；
当所述第二夹角不为0°且所述第一特征点在所述直线L1上方时，表明所述人脸图像中人脸绕Y轴向左旋转，旋转角度为第二夹角的角度值；
当所述第二夹角不为0°且所述第一特征点在所述直线L1下方时，表明所述人脸图像中人脸绕Y轴向右旋转，旋转角度为第二夹角的角度值。
所述第三计算步骤包括：
当所述第一距离与第二距离的比值等于1时，表明所述人脸图像中人脸相对于Z轴处于初始位置，即未向左或向右转头；
当所述第一距离与第二距离的比值大于1时，表明所述人脸图像中人脸绕Z轴向左旋转；
当所述第一距离与第二距离的比值小于1时，表明所述人脸图像中人脸绕Z轴向右旋转。
在本实施例中，所述人脸检测算法为dlib人脸特征点检测算法。参照图3所示，利用dlib人脸特征点检测算法从人脸图像中提取68个特征点，包括17个脸部轮廓特征点，10个眉部特征点，4个鼻翼特征点，其中包括鼻尖特征点，以及其他37个鼻孔、眼部、唇部特征点。所述脸部轮廓特征点中，靠近右耳上方的第一特征点与靠近左耳上方的第四特征点对称，靠近右耳下方的第二特征点与靠近左耳下方的第三特征点对称。
所述X、Y、Z坐标轴的定义为：把人脸认为是一个柱面，则柱面的中心轴作为人脸数据的垂直坐标轴Z轴，取过鼻尖特征点且与柱面中心轴垂直相交的直线作为Y轴，以Z轴和Y轴的叉积方向作为X轴。
在其他实施例中，人脸姿态检测程序10还可以被分割为一个或者多个模块，该一个或者多个模块被存储于存储器12中，并由处理器13执行，以完成本申请。本申请所称的模块是指能够完成特定功能的一系列计算机程序指令段。
参照图2所示，为图1中人脸姿态检测程序10的功能模块图。
所述人脸姿态检测程序10可以被分割为获取模块110、提取模块120及计算模块130。
获取模块110，用于获取摄像装置11拍摄的实时图像，利用人脸识别算法从该实时图像中提取一张实时脸部图像。当摄像装置11拍摄到一张实时图像，摄像装置11将这张实时图像发送到处理器13，当处理器13接受到该实时图像后，所述获取模块110利用人脸识别算法提取出实时的脸部图像。
具体地，从该实时图像中提取实时脸部图像的人脸识别算法可以为基于几何特征的方法、局部特征分析方法、特征脸方法、基于弹性模型的方法、神经网络方法，等等。
提取模块120，用于将该实时脸部图像的人脸特征点和关键特征点提取出来。
例如，所述提取模块120通过dlib人脸特征点检测算法将实时脸部图像中的68个人脸特征点P1-P68提取出来，如图3所示。进一步的，从所述68个人脸特征点中提取出5个关键特征点：靠近右耳上方的第一特征点P1，靠近右耳下方的第二特征点P3，靠近左耳下方的第三特征点P15，靠近左耳上方的第四特征点P17，以及鼻尖处的第五特征点P31。
计算模块130，用于根据该实时人脸图像中5个关键特征点的坐标值计算得到该实时人脸图像中人脸旋转方向及角度。
具体地，所述计算模块130用于：
计算所述第五特征点和第一特征点组成的第一向量与第五特征点和第四特征点组成的第二向量的第一夹角，根据第一夹角与初始夹角的变化判断所述人脸图像中人脸绕X轴的旋转角度；
计算所述第一特征点和第四特征点组成的第三向量与从第四特征点引出的与X轴平行的直线L1的第二夹角，以此判断所述人脸图像中人脸绕Y轴的旋转角度；及
计算所述第二特征点和第五特征点的第一距离与第三特征点和第五特征点的第二距离的比值，以此判断所述人脸图像中人脸绕Z轴的旋转方向。
本实施例提出的计算装置1，从实时图像中提取实时脸部图像，利用dlib人脸特征点检测算法提取出该实时脸部图像中的人脸特征点，并进一步提取出关键特征点，然后根据关键特征点的坐标值计算人脸图像中人脸旋转方向和角度，实现对人脸姿态的实时检测。
此外，本申请还提供一种人脸姿态检测方法。参照图4所示，为本申请人脸姿态检测方法较佳实施例的流程图。该方法可以由一个计算装置执行，该计算装置可以由软件和/或硬件实现。
在本实施例中，人脸姿态检测方法包括：
步骤S10，提取模块120通过人脸检测算法从人脸图像中提取N个人脸特征点。摄像装置11拍摄一张实时图像并将该实时图像发送到处理器13，处理器13接受到该实时图像后，利用人脸识别算法提取出实时的脸部图像，然后通过人脸检测算法获取N个人脸特征点。可选地，从该实时图像中提取实 时的脸部图像的人脸识别算法可以为基于几何特征的方法、局部特征分析方法、特征脸方法、基于弹性模型的方法、神经网络方法，等等。优选地，从该人脸图像中提取人脸特征点的人脸检测算法为dlib人脸特征点检测算法。
步骤S20，提取模块120提取人脸图像特征点中的关键特征点。提取模块120进一步提取靠近右耳上方的第一特征点P1、靠近右耳下方的第二特征点P3、靠近左耳下方的第三特征点P15、靠近左耳上方的第四特征点P17，并提取鼻尖特征点为第五特征点P31。
步骤S30，计算模块130计算所述人脸图像中人脸绕X轴的旋转角度。参照图5所示，将所述第五特征点P31与第一特征点P1组成第一向量，第五特征点P31与第四特征点P17组成第二向量，计算第一向量和第二向量的第一夹角，根据第一夹角与初始夹角的变化判断所述人脸图像中人脸绕X轴的旋转角度；
步骤S40，计算模块130计算所述人脸图像中人脸绕Y轴的旋转角度。参照图6所示，将所述第一特征点P1与第四特征点P17组成第三向量，从第四特征点P17引出一条与X轴平行的直线L1，计算第三向量与直线L1的第二夹角，根据该第二夹角判断所述人脸图像中人脸绕Y轴的旋转角度。
步骤S50，计算模块130计算所述人脸图像中人脸绕Z轴的旋转方向。参照图7所示，将所述第二特征点P3与第五特征点P31的距离记为第一距离，第三特征点P15与第五特征点P31的距离记为第二距离，计算第一距离与第二距离的比值，根据第一距离与第二距离的比值判断人脸绕Z轴的旋转方向。
本实施例提出的人脸姿态检测方法，利用dlib人脸特征点检测算法获取该实时脸部图像中的人脸特征点，并进一步提取出关键特征点，根据关键特征点的坐标值计算该实时脸部图像中人脸旋转方向和角度，实现对人脸姿态的实时检测。
此外，本申请实施例还提出一种计算机可读存储介质，所述计算机可读存储介质中包括人脸姿态检测程序，所述人脸姿态检测程序被所述处理器执行时实现如下操作：
人脸特征点提取步骤：通过人脸检测算法从人脸图像中提取N个人脸特征点。如图3所示，当应用dlib人脸特征点检测算法时，可以从人脸图像中提取人脸特征点P1-P68(N＝68)；
关键特征点提取步骤：从所述N个人脸特征点中提取脸部轮廓特征点中靠近右耳上方的第一特征点P1、靠近右耳下方的第二特征点P3、靠近左耳下方的第三特征点P15、靠近左耳上方的第四特征点P17，并提取鼻尖特征点为第五特征点P31；
第一计算步骤：将所述第五特征点P31与第一特征点P1组成第一向量，第五特征点P31与第四特征点P17组成第二向量(参照图5所示)，计算第一向量与第二向量的第一夹角，根据第一夹角与初始夹角的变化判断所述人脸图 像中人脸绕X轴的旋转角度；
第二计算步骤：将所述第一特征点P1与第四特征点P17组成第三向量，从第四特征点P17引出一条与X轴平行的直线L1(参照图6所示)，计算第三向量与直线L1的第二夹角，根据该第二夹角判断所述人脸图像中人脸绕Y轴的旋转角度；
第三计算步骤：将所述第二特征点P3与第五特征点P31的距离记为第一距离，第三特征点P15与第五特征点P31的距离记为第二距离(参照图7所示)，计算第一距离与第二距离的比值，根据第一距离与第二距离的比值判断所述人脸图像中人脸绕Z轴的旋转方向。
其中，所述第一计算步骤包括：
计算人脸在初始位置时所述第一向量与第二向量的初始夹角，若该初始夹角为120°，则：
当所述第一夹角大于120°时，表明所述人脸图像中人脸绕X轴向上旋转，旋转角度为第一夹角与初始夹角之差；
当所述第一夹角小于120°时，表明所述人脸图像中人脸绕X轴向下旋转，旋转角度为初始夹角与第一夹角之差；
当所述第一夹角等于120°时，表明所述人脸图像中人脸相对于X轴处于初始位置，即未抬头或低头。
所述第二计算步骤包括：
当所述第二夹角为0°时，表明所述人脸图像中人脸相对于Y轴处于初始位置，即未向左或向右歪头；
当所述第二夹角不为0°且所述第一特征点在所述直线L1上方时，表明所述人脸图像中人脸绕Y轴向左旋转，旋转角度为第二夹角的角度值；
当所述第二夹角不为0°且所述第一特征点在所述直线L1下方时，表明所述人脸图像中人脸绕Y轴向右旋转，旋转角度为第二夹角的角度值。
所述第三计算步骤包括：
当所述第一距离与第二距离的比值等于1时，表明所述人脸图像中人脸相对于Z轴处于初始位置，即未向左或向右转头；
当所述第一距离与第二距离的比值大于1时，表明所述人脸图像中人脸绕Z轴向左旋转；
当所述第一距离与第二距离的比值小于1时，表明所述人脸图像中人脸绕Z轴向右旋转。
本申请之人脸姿态检测计算机可读存储介质的具体实施方式与上述人脸姿态检测方法的具体实施方式大致相同，在此不再赘述。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、装置、物品或者方法不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是 还包括为这种过程、装置或方法所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、装置或方法中还存在另外的相同要素。
上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在如上所述的一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是手机，计算机，服务器，或者网络设备等)执行本申请各个实施例所述的方法。
以上仅为本申请的优选实施例，并非因此限制本申请的专利范围，凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。

Claims (20)

  1. 一种人脸姿态检测方法，其特征在于，该方法包括：
    人脸特征点提取步骤：通过人脸检测算法从人脸图像中提取N个人脸特征点；
    关键特征点提取步骤：从所述N个人脸特征点中提取脸部轮廓特征点中靠近右耳上方的第一特征点、靠近右耳下方的第二特征点、靠近左耳下方的第三特征点、靠近左耳上方的第四特征点，并提取鼻尖特征点为第五特征点；
    第一计算步骤：将所述第五特征点与第一特征点组成第一向量，第五特征点与第四特征点组成第二向量，计算第一向量与第二向量的第一夹角，根据第一夹角与初始夹角的变化判断所述人脸图像中人脸绕X轴的旋转角度；
    第二计算步骤：将所述第一特征点与第四特征点组成第三向量，从第四特征点引出一条与X轴平行的直线L1，计算第三向量与直线L1的第二夹角，根据该第二夹角判断所述人脸图像中人脸绕Y轴的旋转角度；
    第三计算步骤：将所述第二特征点与第五特征点的距离记为第一距离，第三特征点与第五特征点的距离记为第二距离，计算第一距离与第二距离的比值，根据第一距离与第二距离的比值判断所述人脸图像中人脸绕Z轴的旋转方向。
  2. 根据权利要求1所述的人脸姿态检测方法，其特征在于，所述人脸检测算法为dlib人脸特征点检测算法。
  3. 根据权利要求1所述的人脸姿态检测方法，其特征在于，所述N个人脸特征点包括脸部轮廓特征点、眉部特征点、鼻翼特征点、鼻孔特征点、眼部特征点和唇部特征点。
  4. 根据权利要求1所述的人脸姿态检测方法，其特征在于，所述脸部轮廓特征点中，靠近右耳上方的第一特征点与靠近左耳上方的第四特征点对称，靠近右耳下方的第二特征点与靠近左耳下方的第三特征点对称。
  5. 根据权利要求1所述的人脸姿态检测方法，其特征在于，所述第一计算步骤包括：
    计算人脸在初始位置时所述第一向量与第二向量的初始夹角；
    当所述第一夹角大于初始夹角时，表明所述人脸图像中人脸绕X轴向上旋转，旋转角度为第一夹角与初始夹角之差；
    当所述第一夹角小于初始夹角时，表明所述人脸图像中人脸绕X轴向下旋转，旋转角度为初始夹角与第一夹角之差；
    当所述第一夹角等于初始夹角时，表明所述人脸图像中人脸相对于X轴处于初始位置。
  6. 根据权利要求1所述的人脸姿态检测方法，其特征在于，所述第二计算步骤包括：
    当所述第二夹角为0°时，表明所述人脸图像中人脸相对于Y轴处于初始位置；
    当所述第二夹角不为0°且所述第一特征点在所述直线L1上方时，表明所述人脸图像中人脸绕Y轴向左旋转，旋转角度为第二夹角的角度值；
    当所述第二夹角不为0°且所述第一特征点在所述直线L1下方时，表明所述人脸图像中人脸绕Y轴向右旋转，旋转角度为第二夹角的角度值。
  7. 根据权利要求1所述的人脸姿态检测方法，其特征在于，所述第三计算步骤包括：
    当所述第一距离与第二距离的比值等于1时，表明所述人脸图像中人脸相对于Z轴处于初始位置；
    当所述第一距离与第二距离的比值大于1时，表明所述人脸图像中人脸绕Z轴向左旋转；
    当所述第一距离与第二距离的比值小于1时，表明所述人脸图像中人脸绕Z轴向右旋转。
  8. 一种计算装置，包括摄像装置、存储器和处理器，其特征在于，所述存储器中包括人脸姿态检测程序，所述人脸姿态检测程序被所述处理器执行时实现如下步骤：
    人脸特征点提取步骤：通过人脸检测算法从人脸图像中提取N个人脸特征点；
    关键特征点提取步骤：从所述N个人脸特征点中提取脸部轮廓特征点中靠近右耳上方的第一特征点、靠近右耳下方的第二特征点、靠近左耳下方的第三特征点、靠近左耳上方的第四特征点，并提取鼻尖特征点为第五特征点；
    第一计算步骤：将所述第五特征点与第一特征点组成第一向量，第五特征点与第四特征点组成第二向量，计算第一向量与第二向量的第一夹角，根据第一夹角与初始夹角的变化判断所述人脸图像中人脸绕X轴的旋转角度；
    第二计算步骤：将所述第一特征点与第四特征点组成第三向量，从第四特征点引出一条与X轴平行的直线L1，计算第三向量与直线L1的第二夹角，根据该第二夹角判断所述人脸图像中人脸绕Y轴的旋转角度；
    第三计算步骤：将所述第二特征点与第五特征点的距离记为第一距离，第三特征点与第五特征点的距离记为第二距离，计算第一距离与第二距离的比值，根据第一距离与第二距离的比值判断所述人脸图像中人脸绕Z轴的旋转方向。
  9. 根据权利要求8所述的计算装置，其特征在于，所述人脸检测算法为 dlib人脸特征点检测算法。
  10. 根据权利要求8所述的计算装置，其特征在于，所述N个人脸特征点包括脸部轮廓特征点、眉部特征点、鼻翼特征点、鼻孔特征点、眼部特征点和唇部特征点。
  11. 根据权利要求8所述的计算装置，其特征在于，所述脸部轮廓特征点中，靠近右耳上方的第一特征点与靠近左耳上方的第四特征点对称，靠近右耳下方的第二特征点与靠近左耳下方的第三特征点对称。
  12. 根据权利要求8所述的计算装置，其特征在于，所述第一计算步骤包括：
    计算人脸在初始位置时所述第一向量与第二向量的初始夹角；
    当所述第一夹角大于初始夹角时，表明所述人脸图像中人脸绕X轴向上旋转，旋转角度为第一夹角与初始夹角之差；
    当所述第一夹角小于初始夹角时，表明所述人脸图像中人脸绕X轴向下旋转，旋转角度为初始夹角与第一夹角之差；
    当所述第一夹角等于初始夹角时，表明所述人脸图像中人脸相对于X轴处于初始位置。
  13. 根据权利要求8所述的计算装置，其特征在于，所述第二计算步骤包括：
    当所述第二夹角为0°时，表明所述人脸图像中人脸相对于Y轴处于初始位置；
    当所述第二夹角不为0°且所述第一特征点在所述直线L1上方时，表明所述人脸图像中人脸绕Y轴向左旋转，旋转角度为第二夹角的角度值；
    当所述第二夹角不为0°且所述第一特征点在所述直线L1下方时，表明所述人脸图像中人脸绕Y轴向右旋转，旋转角度为第二夹角的角度值。
  14. 根据权利要求8所述的计算装置，其特征在于，所述第三计算步骤包括：
    当所述第一距离与第二距离的比值等于1时，表明所述人脸图像中人脸相对于Z轴处于初始位置；
    当所述第一距离与第二距离的比值大于1时，表明所述人脸图像中人脸绕Z轴向左旋转；
    当所述第一距离与第二距离的比值小于1时，表明所述人脸图像中人脸绕Z轴向右旋转。
  15. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质 中包括人脸姿态检测程序，所述人脸姿态检测程序被所述处理器执行时，实现如下步骤：
    人脸特征点提取步骤：通过人脸检测算法从人脸图像中提取N个人脸特征点；
    关键特征点提取步骤：从所述N个人脸特征点中提取脸部轮廓特征点中靠近右耳上方的第一特征点、靠近右耳下方的第二特征点、靠近左耳下方的第三特征点、靠近左耳上方的第四特征点，并提取鼻尖特征点为第五特征点；
    第一计算步骤：将所述第五特征点与第一特征点组成第一向量，第五特征点与第四特征点组成第二向量，计算第一向量与第二向量的第一夹角，根据第一夹角与初始夹角的变化判断所述人脸图像中人脸绕X轴的旋转角度；
    第二计算步骤：将所述第一特征点与第四特征点组成第三向量，从第四特征点引出一条与X轴平行的直线L1，计算第三向量与直线L1的第二夹角，根据该第二夹角判断所述人脸图像中人脸绕Y轴的旋转角度；
    第三计算步骤：将所述第二特征点与第五特征点的距离记为第一距离，第三特征点与第五特征点的距离记为第二距离，计算第一距离与第二距离的比值，根据第一距离与第二距离的比值判断所述人脸图像中人脸绕Z轴的旋转方向。
  16. 根据权利要求15所述的计算机可读存储介质，其特征在于，所述人脸检测算法为dlib人脸特征点检测算法；
    所述N个人脸特征点包括脸部轮廓特征点、眉部特征点、鼻翼特征点、鼻孔特征点、眼部特征点和唇部特征点。
  17. 根据权利要求15所述的计算机可读存储介质，其特征在于，所述脸部轮廓特征点中，靠近右耳上方的第一特征点与靠近左耳上方的第四特征点对称，靠近右耳下方的第二特征点与靠近左耳下方的第三特征点对称。
  18. 根据权利要求15所述的计算机可读存储介质，其特征在于，所述第一计算步骤包括：
    计算人脸在初始位置时所述第一向量与第二向量的初始夹角；
    当所述第一夹角大于初始夹角时，表明所述人脸图像中人脸绕X轴向上旋转，旋转角度为第一夹角与初始夹角之差；
    当所述第一夹角小于初始夹角时，表明所述人脸图像中人脸绕X轴向下旋转，旋转角度为初始夹角与第一夹角之差；
    当所述第一夹角等于初始夹角时，表明所述人脸图像中人脸相对于X轴处于初始位置。
  19. 根据权利要求15所述的计算机可读存储介质，其特征在于，所述第二计算步骤包括：
    当所述第二夹角为0°时，表明所述人脸图像中人脸相对于Y轴处于初始位置；
    当所述第二夹角不为0°且所述第一特征点在所述直线L1上方时，表明所述人脸图像中人脸绕Y轴向左旋转，旋转角度为第二夹角的角度值；
    当所述第二夹角不为0°且所述第一特征点在所述直线L1下方时，表明所述人脸图像中人脸绕Y轴向右旋转，旋转角度为第二夹角的角度值。
  20. 根据权利要求15所述的计算机可读存储介质，其特征在于，所述第三计算步骤包括：
    当所述第一距离与第二距离的比值等于1时，表明所述人脸图像中人脸相对于Z轴处于初始位置；
    当所述第一距离与第二距离的比值大于1时，表明所述人脸图像中人脸绕Z轴向左旋转；
    当所述第一距离与第二距离的比值小于1时，表明所述人脸图像中人脸绕Z轴向右旋转。
PCT/CN2017/108803 2017-08-17 2017-10-31 人脸姿态检测方法、装置及存储介质 Ceased WO2019033576A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
JP2019556821A JP6815707B2 (ja) 2017-08-17 2017-10-31 顔姿勢検出方法、装置及び記憶媒体
SG11202001403XA SG11202001403XA (en) 2017-08-17 2017-10-31 Face pose detection method, device and storage medium
US16/632,371 US11062124B2 (en) 2017-08-17 2017-10-31 Face pose detection method, device and storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710707941.2A CN107679446B (zh) 2017-08-17 2017-08-17 人脸姿态检测方法、装置及存储介质
CN2017107079412 2017-08-17

Publications (1)

Publication Number Publication Date
WO2019033576A1 true WO2019033576A1 (zh) 2019-02-21

Family

ID=61134676

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/108803 Ceased WO2019033576A1 (zh) 2017-08-17 2017-10-31 人脸姿态检测方法、装置及存储介质

Country Status (5)

Country Link
US (1) US11062124B2 (zh)
JP (1) JP6815707B2 (zh)
CN (1) CN107679446B (zh)
SG (1) SG11202001403XA (zh)
WO (1) WO2019033576A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113642368A (zh) * 2020-05-11 2021-11-12 杭州海康威视数字技术股份有限公司 人脸姿态的确定方法、装置、设备及存储介质

Families Citing this family (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10867161B2 (en) * 2017-09-06 2020-12-15 Pixart Imaging Inc. Auxiliary filtering device for face recognition and starting method for electronic device
CN108549487A (zh) * 2018-04-23 2018-09-18 网易（杭州）网络有限公司 虚拟现实交互方法与装置
AU2018420579A1 (en) * 2018-04-25 2020-07-02 Beijing Didi Infinity Technology And Development Co., Ltd. Systems and methods for shaking action recognition based on facial feature points
CN109086727B (zh) * 2018-08-10 2021-04-30 北京奇艺世纪科技有限公司 一种确定人体头部的运动角度的方法、装置及电子设备
CN109522863B (zh) * 2018-11-28 2020-11-27 北京达佳互联信息技术有限公司 耳部关键点检测方法、装置及存储介质
CN110222651A (zh) * 2019-06-10 2019-09-10 Oppo广东移动通信有限公司 一种人脸姿态检测方法、装置、终端设备及可读存储介质
CN112307817B (zh) * 2019-07-29 2024-03-19 中国移动通信集团浙江有限公司 人脸活体检测方法、装置、计算设备及计算机存储介质
CN111104846B (zh) * 2019-10-16 2022-08-30 平安科技（深圳）有限公司 数据检测方法、装置、计算机设备和存储介质
JP7643546B2 (ja) 2021-06-15 2025-03-11 日本電信電話株式会社 学習装置、推定装置、それらの方法、およびプログラム
KR102866156B1 (ko) * 2022-04-27 2025-10-02 주식회사 세오로보틱스 무인 순찰 로봇 및 이의 사람 자세 분석 방법
JP7481398B2 (ja) * 2022-07-04 2024-05-10 ソフトバンク株式会社 判定装置、プログラム、及び判定方法
KR102538783B1 (ko) * 2022-09-27 2023-06-01 주식회사 미러로이드 가상 헤어 스타일링 체험을 위한 민머리 인물 이미지 생성 장치 및 방법
KR102894819B1 (ko) * 2022-10-17 2025-12-03 연세대학교 산학협력단 동적 스마일 분석을 위한 얼굴 정면 여부 판단 방법 및 장치

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040062424A1 (en) * 1999-11-03 2004-04-01 Kent Ridge Digital Labs Face direction estimation using a single gray-level image
CN102054291A (zh) * 2009-11-04 2011-05-11 厦门市美亚柏科信息股份有限公司 一种基于单幅人脸图像实现三维人脸重建的方法及其装置
CN103605965A (zh) * 2013-11-25 2014-02-26 苏州大学 一种多姿态人脸识别方法和装置
KR20150127381A (ko) * 2014-05-07 2015-11-17 포항공과대학교 산학협력단 얼굴 특징점 추출 방법 및 이를 수행하는 장치
CN105389570A (zh) * 2015-11-19 2016-03-09 吴建忠 一种人脸角度判定方法及其系统
CN106991367A (zh) * 2016-01-21 2017-07-28 腾讯科技（深圳）有限公司 确定人脸转动角度的方法和装置

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3590321B2 (ja) * 2000-03-28 2004-11-17 株式会社メディックエンジニアリング 人物照合システム
US6671391B1 (en) * 2000-05-26 2003-12-30 Microsoft Corp. Pose-adaptive face detection system and process
JP4204336B2 (ja) * 2003-01-30 2009-01-07 富士通株式会社 顔の向き検出装置、顔の向き検出方法及びコンピュータプログラム
JP4946730B2 (ja) * 2007-08-27 2012-06-06 ソニー株式会社 顔画像処理装置及び顔画像処理方法、並びにコンピュータ・プログラム
JP5432532B2 (ja) * 2008-01-22 2014-03-05 株式会社 資生堂 化粧方法、化粧シミュレーション装置、及び化粧シミュレーションプログラム
EP2942980A1 (en) * 2014-05-08 2015-11-11 GN Store Nord A/S Real-time control of an acoustic environment
CN106462937B (zh) * 2014-06-06 2019-12-31 夏普株式会社 图像处理装置以及图像显示装置
JP5950486B1 (ja) * 2015-04-01 2016-07-13 みずほ情報総研株式会社 加齢化予測システム、加齢化予測方法及び加齢化予測プログラム
WO2016164859A1 (en) * 2015-04-10 2016-10-13 Bespoke, Inc. Systems and methods for creating eyewear with multi-focal lenses
WO2019110115A1 (en) * 2017-12-07 2019-06-13 Koninklijke Philips N.V. A frame unit
JP6981325B2 (ja) * 2018-03-22 2021-12-15 株式会社Ｊｖｃケンウッド 処理装置、処理方法、及びプログラム
KR102537784B1 (ko) * 2018-08-17 2023-05-30 삼성전자주식회사 전자 장치 및 그 제어 방법
CN109345253A (zh) * 2018-09-04 2019-02-15 阿里巴巴集团控股有限公司 资源转移方法、装置及系统
EP4022386A1 (en) * 2019-09-24 2022-07-06 Bespoke, Inc. d/b/a Topology Eyewear Systems and methods for adjusting stock eyewear frames using a 3d scan of facial features

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040062424A1 (en) * 1999-11-03 2004-04-01 Kent Ridge Digital Labs Face direction estimation using a single gray-level image
CN102054291A (zh) * 2009-11-04 2011-05-11 厦门市美亚柏科信息股份有限公司 一种基于单幅人脸图像实现三维人脸重建的方法及其装置
CN103605965A (zh) * 2013-11-25 2014-02-26 苏州大学 一种多姿态人脸识别方法和装置
KR20150127381A (ko) * 2014-05-07 2015-11-17 포항공과대학교 산학협력단 얼굴 특징점 추출 방법 및 이를 수행하는 장치
CN105389570A (zh) * 2015-11-19 2016-03-09 吴建忠 一种人脸角度判定方法及其系统
CN106991367A (zh) * 2016-01-21 2017-07-28 腾讯科技（深圳）有限公司 确定人脸转动角度的方法和装置

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113642368A (zh) * 2020-05-11 2021-11-12 杭州海康威视数字技术股份有限公司 人脸姿态的确定方法、装置、设备及存储介质
CN113642368B (zh) * 2020-05-11 2023-08-18 杭州海康威视数字技术股份有限公司 人脸姿态的确定方法、装置、设备及存储介质

Also Published As

Publication number Publication date
CN107679446B (zh) 2019-03-15
CN107679446A (zh) 2018-02-09
SG11202001403XA (en) 2020-03-30
US20200226354A1 (en) 2020-07-16
JP2020518051A (ja) 2020-06-18
US11062124B2 (en) 2021-07-13
JP6815707B2 (ja) 2021-01-20

Similar Documents

Publication Publication Date Title
US11062124B2 (en) Face pose detection method, device and storage medium
CN109657533B (zh) 行人重识别方法及相关产品
JP7106742B2 (ja) 顔認識方法、装置、電子機器及びコンピュータ不揮発性読み取り可能な記憶媒体
WO2022166243A1 (zh) 捏合手势检测识别方法、装置及系统
CN106030610B (zh) 移动设备的实时3d姿势识别和跟踪系统
CN110647811A (zh) 人脸姿态检测方法、装置及计算机可读存储介质
WO2020215565A1 (zh) 手部图像的分割方法、装置及计算机设备
WO2019033571A1 (zh) 面部特征点检测方法、装置及存储介质
TW201120681A (en) Method and system for operating electric apparatus
WO2019041519A1 (zh) 目标跟踪装置、方法及计算机可读存储介质
JP2021524951A (ja) 空中手書きを識別するための方法、装置、デバイス及びコンピュータ読み取り可能な記憶媒体
WO2018028546A1 (zh) 一种关键点的定位方法及终端、计算机存储介质
CN105095829A (zh) 一种人脸识别方法及系统
WO2019033567A1 (zh) 眼球动作捕捉方法、装置及存储介质
WO2019033570A1 (zh) 嘴唇动作分析方法、装置及存储介质
CN111444764A (zh) 一种基于深度残差网络的手势识别方法
WO2017113407A1 (zh) 一种手势识别方法、装置及电子设备
WO2019000817A1 (zh) 手势识别控制方法和电子设备
CN109993021A (zh) 人脸正脸检测方法、装置及电子设备
WO2022022365A1 (zh) 一种应用于指纹识别的图像匹配方法及相关装置
CN111754571A (zh) 一种姿态识别方法、装置及其存储介质
WO2020124993A1 (zh) 活体检测方法、装置、电子设备及存储介质
WO2018161893A1 (zh) 用户身份识别方法及装置
WO2018064800A1 (en) Apparatus, method and computer program product for distance estimation between samples
CN113158912B (zh) 手势识别方法及装置、存储介质及电子设备

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: 17921934

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2019556821

Country of ref document: JP

Kind code of ref document: A

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 (EPO FORM 1205A DATED 24.09.2020)

122 Ep: pct application non-entry in european phase

Ref document number: 17921934

Country of ref document: EP

Kind code of ref document: A1