CN108021902A - Head pose recognition methods, head pose identification device and storage medium - Google Patents

Head pose recognition methods, head pose identification device and storage medium Download PDF

Info

Publication number
CN108021902A
CN108021902A CN201711373258.6A CN201711373258A CN108021902A CN 108021902 A CN108021902 A CN 108021902A CN 201711373258 A CN201711373258 A CN 201711373258A CN 108021902 A CN108021902 A CN 108021902A
Authority
CN
China
Prior art keywords
triangle
eye position
position
default
head
Prior art date
Application number
CN201711373258.6A
Other languages
Chinese (zh)
Inventor
杨志勇
林连生
苗瑛
Original Assignee
珠海瞳印科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 珠海瞳印科技有限公司 filed Critical 珠海瞳印科技有限公司
Priority to CN201711373258.6A priority Critical patent/CN108021902A/en
Publication of CN108021902A publication Critical patent/CN108021902A/en

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K9/00Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K9/00221Acquiring or recognising human faces, facial parts, facial sketches, facial expressions
    • G06K9/00228Detection; Localisation; Normalisation
    • G06K9/00248Detection; Localisation; Normalisation using facial parts and geometric relationships

Abstract

The present invention provides a kind of head pose recognition methods, head pose identification device and storage medium, and head pose recognition methods includes:Gather user's head image;Identification head region simultaneously obtains the two-dimensional position information at multiple positions;The two-dimensional position information at multiple positions is compared with preset standard positional information, and draws comparison result;Judged according to the comparison result and draw posture state, real time image collection can be carried out to user's head by camera, and identify multiple positions of user's head and the two-dimensional position information at multiple positions, then utilize to the two-dimensional position information ratio at position for preset standard positional information, it then can determine whether to draw the head of active user is in state how, then the identification to head pose can be realized, the image gathered due to the camera utilized, its applicability is high, it can be suitable for a variety of application places, and one camera of energy gathers the head of multiple users and is identified at the same time.

Description

Head pose recognition methods, head pose identification device and storage medium

Technical field

The present invention relates to human bioequivalence field, more particularly to a kind of head pose recognition methods, head pose identification device And computer-readable recording medium.

Background technology

The situation that child myopia becomes younger is increasingly severe, and existing myopia mean of defense is usually to utilize Nearsightedness-preventing pen, Nearsightedness-preventing pen has generally comprised the cap for brush, penholder, pen cap, nib and nib lid, and infrared receiving/transmission spy is dynamically provided with penholder Circuit board, actuating mechanism and the battery of head, its major function is that infrared receiving/transmission mechanism is provided with penholder, and passes through its detection To the signal being reflected back from writer face, whether control pen core upper end is headed on by thrust bar, if face is too near to penholder, electricity Road driving thrust bar, which releases, contacts pen core, and pen core can bounce back once writing and being pressurized, and can not use, reach protect eyes Purpose.But not only cost is big for this Nearsightedness-preventing pen, and heavier Nearsightedness-preventing pen also influences whether normal usage experience, can be big It is big to hinder writing, and be under children's service condition.

Therefore, ultrasonic wave and remote infrared is usually used in the prior art at the same time, it can only be by identifying that distance becomes Change and judge posture, therefore be there are technical bottleneck, cannot promote the use of and be generalized in other industry on a large scale and use, respectively Each industry of row needs a set of general solution there are occupational disease or pose problem.

The content of the invention

The first object of the present invention is to provide a kind of head pose recognition methods that head pose identification is carried out using image.

The second object of the present invention is to provide a kind of head pose identification device that head pose identification is carried out using image.

The third object of the present invention is to provide a kind of computer-readable recording medium for realizing picture headers gesture recognition.

In order to realize the first object of the present invention, the present invention provides a kind of head pose recognition methods, including:

Gather user's head image;

Identification head region simultaneously obtains the two-dimensional position information at multiple positions;

The two-dimensional position information at multiple positions is compared with preset standard positional information, and draws comparison result;

Judged according to the comparison result and draw posture state.

By such scheme as it can be seen that real time image collection, and identification account can be carried out to user's head by camera Multiple positions in portion and the two-dimensional position information at multiple positions, then utilize to the two-dimensional position information ratio at position for pre- bidding Quasi- positional information, then can determine whether to draw the head of active user is in state how, can then realize to head pose Identification, due to the image that the camera utilized gathers, its applicability is high, can be suitable for a variety of application places, and energy one is taken the photograph As head gathers the head of multiple users at the same time and is identified.

Further scheme is that the position includes left eye position, right eye position and nose areas, the preset standard Positional information includes default triangle;

The positional information at the left eye position, the right eye position and the nose areas is used for the default triangle into line position Put comparison.

Further scheme is when being compared, to judge the left eye position, the right eye position and the nose portion Whether position is respectively distributed on three angles of the default triangle;

Three angles of triangle are preset as described in being respectively distributed to the left eye position, the right eye position and the nose areas On, then the posture state is standard posture.

Further scheme is, the left eye position, the right eye position and the nose areas surround real-time three Angle;

When being compared, when judging that the area of the triangle in real time is more than the area of the default triangle;

Area such as the triangle in real time is more than the area of the default triangle, then the posture state is posture too far.

Further scheme is, the left eye position, the right eye position and the nose areas surround real-time three Angle;

When being compared, when judging that the area of the triangle in real time is less than the area of the default triangle;

Area such as the triangle in real time is less than the area of the default triangle, then the posture state was nearly posture.

Further scheme is that the default triangle is set in inverted triangle;

When being compared, such as left eye position is located at the upside of the default triangle, the right eye position positioned at described default The left side of triangle and the nose areas are located at the right side of the default triangle, then the posture state is posture bending to the right;

As the left eye position is located at the right side of the default triangle, the right eye position be located at the default triangle upside and The nose areas is located at the left side of the default triangle, then the posture state is posture bending to the left.

Further scheme is that the default triangle is set in inverted triangle;

When being compared, such as left eye position is located at the right side of the default triangle, the right eye position positioned at described default The left side of triangle and the nose areas are located at the inner side of the default triangle, then the posture state for cocked-up attitude or is bowed Posture.

Therefore by the position to left eye position, right eye position and nose areas relative to default triangle, can be to mistake Far, excessively closely, to the right, to the left, face upward head or the bad posture such as bow is identified.

Further scheme is that the position includes left eye position, right eye position, nose areas, left ear position, auris dextra Position and face mask;

According to left ear position, the auris dextra position and the face mask, by the left eye position, the right eye position and The positional information of the nose areas is arranged to the preset standard position.

Therefore since the face arrangement of different people has difference, when setting default triangle or in real-time judge, Left ear position, auris dextra position and face mask can be introduced and carry out framing, therefore can further improve image recognition Precision.

In order to realize the second object of the present invention, the present invention provides a kind of head pose identification device, it is characterised in that institute Stating head pose identification device includes processor, and the processor is used to realize when performing the computer program stored in memory As described in such scheme any one the step of recognition methods.

In order to realize the third object of the present invention, the present invention provides a kind of computer-readable recording medium, is stored thereon with Computer program, it is characterised in that:Realized when the computer program is executed by processor as described in any one of such scheme The step of recognition methods.

By such scheme as it can be seen that real time image collection, and identification account can be carried out to user's head by camera Multiple positions in portion and the two-dimensional position information at multiple positions, then utilize to the two-dimensional position information ratio at position for pre- bidding Quasi- positional information, then can determine whether to draw the head of active user is in state how, can then realize to head pose Identification, due to the image that the camera utilized gathers, its applicability is high, can be suitable for a variety of application places, and energy one is taken the photograph As head gathers the head of multiple users at the same time and is identified.

Brief description of the drawings

Fig. 1 is the application principle schematic diagram of head pose recognition methods embodiment of the present invention.

Fig. 2 is head pose recognition methods embodiment Plays gesture recognition schematic diagram of the present invention.

Fig. 3 is gesture recognition schematic diagram too far in head pose recognition methods embodiment of the present invention.

Fig. 4 is excessively near gesture recognition schematic diagram in head pose recognition methods embodiment of the present invention.

Fig. 5 is gesture recognition schematic diagram to the right in head pose recognition methods embodiment of the present invention.

Fig. 6 is gesture recognition schematic diagram to the left in head pose recognition methods embodiment of the present invention.

Fig. 7 is that head or nose-down attitude identification schematic diagram are faced upward in head pose recognition methods embodiment of the present invention.

Fig. 8 is the flow chart of head pose recognition methods embodiment of the present invention.

The invention will be further described with reference to the accompanying drawings and embodiments.

Embodiment

Head pose recognition methods embodiment:

With reference to Fig. 1, Fig. 2 and Fig. 3, head pose recognition methods then realizes that identification head is in correct by gathering head region Sitting posture or wryneck, too remote, too near, the abnormal sitting posture looked up etc., specific recognition methods include:

Step S11 is performed, user's head is shot using camera device and gathers user's head image, when shooting can be the same as screen collection The head of more people, the image of multiple user's heads can be drawn by analysis.

Step S12 is performed, identifies the two-dimensional position information at multiple crucial positions of user in image, plurality of position Including left eye position, right eye position, nose areas, left ear position, auris dextra position and face mask, then and above-mentioned position is obtained Two-dimensional position information.

And in advance, structure connection information intrinsic between human face characteristic point is stored, and can be in data lab setting Default corresponding preset standard positional information, this preset standard positional information can also be in initial setting up according to different use Family carries out default adjustment, and adjustment mode can be according to left ear position, auris dextra position and face mask and by left eye position, right eye position Preset standard position is arranged to the positional information of nose areas, that is, performs step S13.

Step S14 is performed, the two-dimensional position information at multiple positions is compared with preset standard positional information and can draw Comparison result, the specific principle that compares can refer to Fig. 2, and preset standard position is in that default inverted triangle is set, itself and left eye position, the right side Eye position is associated with nose areas and position compares, and R points represent right eye position, and L points represent left eye position, and N points represent nose portion Position, O points represent the center of preset standard position, it is pre- to judge whether left eye position, right eye position and nose areas are respectively distributed to If on three angles of triangle, if so, then comparison result is standard posture.

With reference to Fig. 3, the real-time triangle that left eye position, right eye position and nose areas surround, when being compared, judges real-time Whether the area of triangle is more than the area of default triangle, and the area of 3 points of outsides for being respectively positioned on default triangle, such as real-time triangle More than the area of default triangle, then comparison result is posture too far.

With reference to Fig. 4, the real-time triangle that left eye position, right eye position and nose areas surround, when being compared, judges real-time Whether the area of triangle is less than the area of default triangle, and 3 points of inside for being respectively positioned on default triangle, if so, then comparison result is Cross nearly posture.

It is located at the upside for presetting triangle, right eye position positioned at the left side of default triangle and nose with reference to Fig. 5, such as left eye position Position is located at the right side of default triangle, it is necessary to which explanation, upside, the orientation of left or right side can be understood as on two dimensional surface Two-dimensional position relation, then show positioned at upside it is in larger distance in Y-axis, then show positioned at left side it is in small distance in X-axis, and With reference to that can need center O points, by above-mentioned judgement, then it is posture bending to the right that comparison result, which can be calculated,.

It is located at the right side for presetting triangle, right eye position positioned at the upside of default triangle and nose with reference to Fig. 6, such as left eye position Position is located at the left side of default triangle, then posture state is posture bending to the left.

It is located at the right side for presetting triangle, right eye position positioned at the left side of default triangle and nose with reference to Fig. 7, such as left eye position Position is located at the inner side of default triangle, then posture state is cocked-up attitude or nose-down attitude.

By above-mentioned judgement and current head pose is can be calculated, followed by step S16, it is more to continue collection A image is continuously analyzed, and after the head pose of continuous certain time is identified, can perform step S17, according to And result of analysing and comparing judges and draws posture state, judge whether current sitting posture posture is normal.

In addition, joint left and right ear group is can also be in addition to including default triangle on preset standard positional information Into default pentagon, and can achieve the object of the present invention.

Above-mentioned head pose method can be widely used in classroom, fatigue drive of car monitoring, office, production line operation Deng place, multiple personnel are monitored at the same time using camera, can be the state of attending class that teacher obtains student in real time, or supervise Whether the sitting posture for controlling driver, or the working status of clerk, or product line worker can then occupational disease etc..

Head pose identification device head pose identification device includes processor, and processor is used to perform in memory to store Computer program when realize as in such scheme the step of recognition methods.

Computer-readable recording medium, is stored thereon with computer program, is realized when computer program is executed by processor As in such scheme the step of recognition methods.

Therefore real time image collection can be carried out to user's head by camera, and identify the more of user's head A position and the two-dimensional position information at multiple positions, then utilize to the two-dimensional position information ratio at position for preset standard position Information, then can determine whether to draw the head of active user is in state how, can then realize the identification to head pose, by In the image of the camera collection utilized, its applicability is high, can be suitable for a variety of application places, and one camera of energy is at the same time Gather the head of multiple users and be identified.

Claims (10)

1. head pose recognition methods, it is characterised in that including:
Gather user's head image;
Identification head region simultaneously obtains the two-dimensional position information at multiple positions;
The two-dimensional position information at multiple positions is compared with preset standard positional information, and draws comparison result;
Judged according to the comparison result and draw posture state.
2. head pose recognition methods according to claim 1, it is characterised in that:
The position includes left eye position, right eye position and nose areas, and the preset standard positional information includes default triangle;
The positional information at the left eye position, the right eye position and the nose areas is used for the default triangle into line position Put comparison.
3. head pose recognition methods according to claim 2, it is characterised in that:
When being compared, it is described to judge whether the left eye position, the right eye position and the nose areas are respectively distributed to On three angles of default triangle;
Three angles of triangle are preset as described in being respectively distributed to the left eye position, the right eye position and the nose areas On, then the posture state is standard posture.
4. head pose recognition methods according to claim 2, it is characterised in that:
The real-time triangle that the left eye position, the right eye position and the nose areas surround;
When being compared, judge whether the area of the triangle in real time is more than the area of the default triangle;
Area such as the triangle in real time is more than the area of the default triangle, then the posture state is posture too far.
5. head pose recognition methods according to claim 2, it is characterised in that:
The real-time triangle that the left eye position, the right eye position and the nose areas surround;
When being compared, judge whether the area of the triangle in real time is less than the area of the default triangle;
Area such as the triangle in real time is less than the area of the default triangle, then the posture state was nearly posture.
6. head pose recognition methods according to claim 2, it is characterised in that:
The default triangle is set in inverted triangle;
When being compared, such as left eye position is located at the upside of the default triangle, the right eye position positioned at described default The left side of triangle and the nose areas are located at the right side of the default triangle, then the posture state is posture bending to the right;
As the left eye position is located at the right side of the default triangle, the right eye position be located at the default triangle upside and The nose areas is located at the left side of the default triangle, then the posture state is posture bending to the left.
7. head pose recognition methods according to claim 2, it is characterised in that:
The default triangle is set in inverted triangle;
When being compared, such as left eye position is located at the right side of the default triangle, the right eye position positioned at described default The left side of triangle and the nose areas are located at the inner side of the default triangle, then the posture state for cocked-up attitude or is bowed Posture.
8. head pose recognition methods according to any one of claims 1 to 7, it is characterised in that:
The position includes left eye position, right eye position, nose areas, left ear position, auris dextra position and face mask;
According to left ear position, the auris dextra position and the face mask, by the left eye position, the right eye position and The positional information of the nose areas is arranged to the preset standard position.
9. head pose identification device, it is characterised in that the head pose identification device includes processor, and the processor is used The step of the recognition methods as described in any one in claim 1 to 8 is realized when the computer program stored in memory is performed Suddenly.
10. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that:The computer program Realized when being executed by processor as described in any one of claim 1 to 8 the step of recognition methods.
CN201711373258.6A 2017-12-19 2017-12-19 Head pose recognition methods, head pose identification device and storage medium CN108021902A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711373258.6A CN108021902A (en) 2017-12-19 2017-12-19 Head pose recognition methods, head pose identification device and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711373258.6A CN108021902A (en) 2017-12-19 2017-12-19 Head pose recognition methods, head pose identification device and storage medium

Publications (1)

Publication Number Publication Date
CN108021902A true CN108021902A (en) 2018-05-11

Family

ID=62074241

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711373258.6A CN108021902A (en) 2017-12-19 2017-12-19 Head pose recognition methods, head pose identification device and storage medium

Country Status (1)

Country Link
CN (1) CN108021902A (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110158546A1 (en) * 2009-12-25 2011-06-30 Primax Electronics Ltd. System and method for generating control instruction by using image pickup device to recognize users posture
CN102254164A (en) * 2011-08-04 2011-11-23 蔡翔 Face recognition method with prompt function
CN103369214A (en) * 2012-03-30 2013-10-23 华晶科技股份有限公司 An image acquiring method and an image acquiring apparatus
CN105046246A (en) * 2015-08-31 2015-11-11 广州市幸福网络技术有限公司 Identification photo camera capable of performing human image posture photography prompting and human image posture detection method
CN107103309A (en) * 2017-05-31 2017-08-29 天津中科智能识别产业技术研究院有限公司 A kind of sitting posture of student detection and correcting system based on image recognition

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110158546A1 (en) * 2009-12-25 2011-06-30 Primax Electronics Ltd. System and method for generating control instruction by using image pickup device to recognize users posture
CN102254164A (en) * 2011-08-04 2011-11-23 蔡翔 Face recognition method with prompt function
CN103369214A (en) * 2012-03-30 2013-10-23 华晶科技股份有限公司 An image acquiring method and an image acquiring apparatus
CN105046246A (en) * 2015-08-31 2015-11-11 广州市幸福网络技术有限公司 Identification photo camera capable of performing human image posture photography prompting and human image posture detection method
CN107103309A (en) * 2017-05-31 2017-08-29 天津中科智能识别产业技术研究院有限公司 A kind of sitting posture of student detection and correcting system based on image recognition

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
赵俊宝: "《驾驶员头部姿态估计系统的设计与实现》", 《中国优秀硕士学位论文全文数据库信息科技辑》 *

Similar Documents

Publication Publication Date Title
US10269222B2 (en) System with wearable device and haptic output device
Plouffe et al. Static and dynamic hand gesture recognition in depth data using dynamic time warping
Monkaresi et al. Automated detection of engagement using video-based estimation of facial expressions and heart rate
Al-Rahayfeh et al. Eye tracking and head movement detection: A state-of-art survey
US20200050281A1 (en) Machine responsiveness to dynamic user movements and gestures
US9547798B2 (en) Gaze tracking for a vehicle operator
US10062303B2 (en) Object tracking for artificial vision
US8824779B1 (en) Apparatus and method for determining eye gaze from stereo-optic views
US8814357B2 (en) System and method for identifying the existence and position of text in visual media content and for determining a subject's interactions with the text
CN102749991B (en) A kind of contactless free space sight tracing being applicable to man-machine interaction
US10165176B2 (en) Methods, systems, and computer readable media for leveraging user gaze in user monitoring subregion selection systems
US9538219B2 (en) Degree of interest estimating device and degree of interest estimating method
US10007336B2 (en) Apparatus, system, and method for mobile, low-cost headset for 3D point of gaze estimation
US8199126B1 (en) Use of potential-touch detection to improve responsiveness of devices
CN102081918B (en) Video image display control method and video image display device
Ehinger et al. Modelling search for people in 900 scenes: A combined source model of eye guidance
Braunagel et al. Driver-activity recognition in the context of conditionally autonomous driving
US6154559A (en) System for classifying an individual's gaze direction
CN102693047B (en) The method of numerical information and image is shown in cooperative surroundings
KR101765961B1 (en) Vision testing and/or training using adaptable visual indicia
US20130271584A1 (en) User wearable visual assistance device
US7043056B2 (en) Facial image processing system
EP2634727A2 (en) Method and portable terminal for correcting gaze direction of user in image
CN104463152B (en) A kind of gesture identification method, system, terminal device and Wearable
CN101466305B (en) Method for determining and analyzing a location of visual interest

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination