CN105487665B - A kind of intelligent Mobile Service robot control method based on head pose identification - Google Patents
A kind of intelligent Mobile Service robot control method based on head pose identification Download PDFInfo
- Publication number
- CN105487665B CN105487665B CN201510872912.2A CN201510872912A CN105487665B CN 105487665 B CN105487665 B CN 105487665B CN 201510872912 A CN201510872912 A CN 201510872912A CN 105487665 B CN105487665 B CN 105487665B
- Authority
- CN
- China
- Prior art keywords
- head
- mobile service
- service robot
- head pose
- characteristic point
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Abstract
The present invention provides a kind of intelligent Mobile Service robot control method identified based on head pose, this method detects facial feature points using constraint partial model algorithm, the geometrical relationship of facial feature points is recycled to estimate current head posture, to enable the patient to control the movement of intelligent Mobile Service robot with head.Its specific implementation includes the following steps：Training head pose sample obtains the local template of facial feature points shape template and characteristic point；Using Ki nect acquisition head images and find face location；Facial feature points are detected by constraining partial model algorithm；Utilize the facial feature estimation current head posture detected；Different control commands is sent out to Information Mobile Service robot further according to different head pose parameters.Utilize the method for the present invention so that physical disabilities can also control intelligent Mobile Service robot using head steady, facilitate their trip.
Description
Technical field
The present invention relates to machine vision control fields, and in particular to a kind of intelligent Mobile Service based on head pose identification
Robot control method is applied in intelligent Mobile Service robot interactive system, is suitable for handing in intelligent Mobile Service robot
The head pose real-time estimation based on local restriction model (CLM) is realized in mutual system and intelligent mobile is taken to realize based on this
The control of business robot.
Background technology
It is faced jointly with the development of society, the problem of an aging population has become country in developed country and partial development
Challenge, predictive display, from 2015~2035 years 20 years, Aged in China population ratio will double, and reach
20%, wherein being no lack of since disease leads to the old man of inconvenient activity, in addition not due to various action caused by various disaster accidents
Just the quantity of physical disabilities is also gradually increasing, especially high paraplegia and arm handicapped person scholar, a part of in them
Travel activity receives great restriction.In order to trip facility with their quality of life of raising, it is in recent years, convenient to go on a journey
Walking-replacing tool becomes the hot spot of society and research institution's concern, and wherein intellect service robot more becomes hot spot therein and grinds
Study carefully object.Intelligent wheel chair as a kind of representative intelligent Mobile Service robot there is rocking bar to control, limb control,
Navigation, avoidance, the functions such as rehabilitation.Wherein rocking bar control possesses good control performance, but for both hands deformity or high paraplegia
Cause for the people that both hands can not move, rocking bar control and part of limb control such as the methods of gesture control receive very big
Restriction, therefore, impetus, which is controlled, becomes the hot spot of Recent study as a kind of novel man-machine interaction mode.
The method of current existing identification head appearance has：
1, wearable motion sensor method, such method on test object head by installing accelerometer and gyroscope biography
Sensor judges head appearance by receiving sensor exercise data, and such method precision is high, but needs wearable device, user experience compared with
Difference.
2, LED scaling methods capture headgear image, then root by the headgear with certain amount LED using camera
According to the current position of the position judgment of LED, equally, the method precision is high, but still needs to wearable device.
3, machine vision method, the method acquire head image by camera and judge head appearance by algorithm, are current
Main stream approach contacts, better user experience with test object zero, wherein existing means are mainly logical using algorithm for pattern recognition
Cross Image estimation head appearance (machine vision method), such as the random forest grader of traditional template matching algorithm, mainstream, face
Characteristic point geometrical relationship method etc., traditional CLM algorithms utilize merely two dimensional image, exist and different illumination conditions are interfered
Problem.
Invention content
Present invention aims at a kind of intelligent Mobile Service robot control based on head pose identification is provided, using base
Estimate head pose mode in the CLM methods of two dimensional image and depth image, is detected by local restriction model algorithm CLM
Human face characteristic point, further according to the geometrical relationship estimation head pose for the facial feature points that detection obtains, and then again to intelligent mobile
Service robot implements control.
The above-mentioned purpose of the present invention realizes that dependent claims are to select else or have by the technical characteristic of independent claims
The mode of profit develops the technical characteristic of independent claims.
To reach above-mentioned purpose, the present invention proposes a kind of intelligent Mobile Service robot control identified based on head pose
Method obtains head pose and controls Information Mobile Service robot based on constraining partial model algorithm, and specific implementation includes
Following steps：
S1 builds the shape template of facial feature points and the local feature template of characteristic point by head pose sample database
The purpose of the step be in sample database chromatic image and depth image PCA dimensionality reductions and be aligned, then into
Row feature extraction and study obtain the local feature template of the shape template and each characteristic point of characteristic point, and build corresponding
SVM is applied to subsequent matching step.
S2 by two dimensional image of the RGBD photographic devices acquisition comprising face and depth image and is aligned
Include the coloured image and depth image of face using Kinect acquisitions in this step, can and meanwhile acquisition resolution
All it is the coloured image and depth image of 640*480, acquisition rate is 30 frame per second.Due to the colour imagery shot and depth of Kinect
The image that degree camera has a certain distance, therefore collects two width needs correction function first to carry out registration process.
S3 detects face location in the picture using Viola-Jones detectors
The position of face is quickly positioned in collected RGB color image using the method, if with a rectangle frame packet
The face detected is enclosed, the characteristic point for following further detection face is ready.
S4 detects characteristic point using CLM algorithms on the face detected
The position for further searching for facial feature points in the step in the face location detected, carries out face first
The initial estimation of characteristic point, then the estimation of better place is carried out to the characteristic point of initial estimation, iteration is arrived until all characteristic points
The estimation of facial feature points is so far completed in optimum position.
S5 estimates head pose by the shape of the human face characteristic point detected
Current head pose is estimated in the step using the shape of the facial feature points detected in previous step, is estimated
Head appearance is divided by the pitch angle on head, roll angle, three angles of angle of rotation：Appearance without a head, left-hand rotation head, right-hand rotation head come back, bow
5 kinds of head poses altogether.
S6 realizes the control to Information Mobile Service robot by head pose recognition result
PC machine sends out different control commands according to current head pose and gives Information Mobile Service robot in this step, 5 kinds
The corresponding motion state of head appearance is：Stop, turning left, turn right, advance, retreats.Computer is by identifying obtained head appearance
Gesture result sends different orders to dsp controller by serial ports, and DSP retransmits control command and driven to motor to control intelligence
The movement of energy wheelchair.
It should be appreciated that as long as aforementioned concepts and all combinations additionally conceived that describe in greater detail below are at this
Sample design it is not conflicting in the case of can be viewed as the disclosure subject matter a part.In addition, required guarantor
All combinations of the theme of shield are considered as a part for the subject matter of the disclosure.
Can be more fully appreciated from the following description in conjunction with attached drawing present invention teach that foregoing and other aspect, reality
Apply example and feature.The feature and/or advantageous effect of other additional aspects such as illustrative embodiments of the present invention will be below
Description in it is obvious, or by according to present invention teach that specific implementation mode practice in learn.
Description of the drawings
Attached drawing is not intended to drawn to scale.In the accompanying drawings, identical or approximately uniform group each of is shown in each figure
It can be indicated by the same numeral at part.For clarity, in each figure, not each component part is labeled.
Now, by example and the embodiments of various aspects of the invention will be described in reference to the drawings, wherein：
Fig. 1 is the structure chart and Kinect camera schematic view of the mounting position of intelligent Mobile Service robot.
Fig. 2 is intelligent Mobile Service robot control system architecture figure.
Fig. 3 is head pose angle schematic diagram.
Fig. 4 is the flow chart of the intelligent Mobile Service robot control method identified based on head pose.
Fig. 5 is the schematic diagram for demarcating facial feature points.
Fig. 6 is detection face location schematic diagram.
Fig. 7 is to pass through facial feature estimation head pose schematic diagram.
Fig. 8 is facial feature points testing result and head appearance result.
Specific implementation mode
In order to know more about the technology contents of the present invention, spy lifts specific embodiment and institute's accompanying drawings is coordinated to be described as follows.
Various aspects with reference to the accompanying drawings to describe the present invention in the disclosure, shown in the drawings of the embodiment of many explanations.
It is not intended to cover all aspects of the invention for embodiment of the disclosure.It should be appreciated that a variety of designs and reality presented hereinbefore
Those of apply example, and describe in more detail below design and embodiment can in many ways in any one come it is real
It applies, this is because design disclosed in this invention and embodiment are not limited to any embodiment.In addition, disclosed by the invention one
A little aspects can be used alone, or otherwise any appropriately combined be used with disclosed by the invention.
Fig. 1 is the structure chart of intelligent Mobile Service robot, the tool of intelligent Mobile Service robot 100 there are one main body and
The each component being arranged in main body, these components include：Headrest 101, PC controllers 103, manipulates Kinect cameras 102
Bar 104, motor 105, battery 106, front-wheel 107, trailing wheel 108 and anti-hypsokinesis wheel 109.As shown, Kinect cameras 102
Front about 50cm mounted on user head or so place, it is ensured that can face head and by entire head be included in acquisition
In picture, distance 50cm or so can preferably acquire coloured image and depth image.
It should be appreciated that in Fig. 1, aforementioned headrest 101, PC controllers 103, control stick 104, motor 105, battery 106, front-wheel
107, trailing wheel 108 and anti-hypsokinesis wheel 109 are the common setting in intelligent Mobile Service robot, specific configuration, function
And/or details are not described herein for effect.These components installation site shown in FIG. 1 and/or a combination thereof are only exemplary, one
In a little needs or necessary example, make, combination and/or its installation site of these components can be any ways
's.
Fig. 2 illustratively illustrates that intelligent Mobile Service robot control system architecture figure, this control system include image
Acquisition module, image processing module, intelligent Mobile Service robot control module, intelligent Mobile Service robot.Wherein, image
Acquisition module, that is, Kinect cameras, image processing module use PC controllers, intelligent Mobile Service robot control module to make
With DSP, by PC controllers to the head pose information that is identified after image procossing come to intelligent Mobile Service robot into
Row activity control.
Fig. 3 show head pose angle schematic diagram, is divided into pitch angle, angle of rotation and roll angle.
Fig. 4 is the intelligent Mobile Service robot controlling party based on head pose identification according to certain embodiments of the invention
The flow chart of method, is as follows：
S1 builds the shape template of facial feature points and the local feature template of different characteristic point；
S2 acquires the two dimensional image on head and depth image and is aligned；
S3 detects face location in the image collected；
S4 detects characteristic point using CLM algorithms on face；
S5 estimates head pose by the geometrical relationship of the human face characteristic point detected；
S6 realizes the control to Information Mobile Service robot by head pose recognition result.
Shown in below in conjunction with the accompanying drawings, the aforementioned intelligent Mobile Service machine identified based on head pose is more specifically described
The exemplary realization of people's control method.
In step S1, the local feature template specific steps of the shape template and different characteristic point that build facial feature points are such as
Under：
S11 collects n head appearance training samples, it is desirable that each sample includes to compare clearly face image, each sample
Including gray scale and depth two images, and resolution ratio is all 640*480, it is desirable that gray scale and depth image have been aligned.
S12 demarcates the characteristic point on every width sample, as shown in figure 5, demarcating m characteristic point, each feature on every width sample
The corresponding fixed serial number of point, is connected into a vector as the shape vector of this sample using the coordinate of m characteristic point, obtains one
Shape vector：
Si={ xi1,yi1,xi2,yi2,...xim,yim}
(x in formulaij,yij) represent j-th of characteristic point of i-th of sample.
N face sample is demarcated, n shape vector is obtained.Representative point is chosen in the calibration of characteristic point as possible,
Such as the canthus in figure, the corners of the mouth, nose, at least one of chin.
S13, is normalized n shape vector of structure using Procrustes analysis methods and registration process, disappears
Except displacement, scaling, the influence of rotation；
S14 extracts the principal component of shape vector by PCA methods；
S15 trains the local classifiers of this feature point using corresponding image block to each characteristic point.
In some instances, in abovementioned steps S13, using Procrustes analysis methods to n shape vector of structure
The realization with registration process is normalized, includes the following steps：
S131 eliminates Influence of Displacement first, and each characteristic point position in the face is subtracted all features in the face
The mean place of point is to make sample set snap to origin：
S132, if the shape of all samples obtains sample set S=(S after all snapping to origin1,S2,S3...Sn)。
S133 selects S1As with reference to shape, suitable European similarity transformations are selected to other all shapes, with S1For mesh
Mark is aligned, and one group of new shape is obtained after alignment：
S'=(S1',S2',S3'...Sn')
S134 calculates the average shape of new shape collection：
S135, by new average shapeWith S1Alignment obtains
S136, to (S2',S3',...,Sn') the suitable transformation of selection, withAlignment.
S137 calculates new average shape again, if the variation of new average shape and last average shape is big
It is slight in a certain given threshold value, then it is believed that iteration terminate, otherwise go to S134.
It is in this way, by being aligned face shape vector, different shape vector is unified under the same coordinate system, in place
It sets, size, be comparable in rotation, so as to carry out the analysis of next step.
In some instances, abovementioned steps S14 extracts the principal component of shape vector by PCA methods, and specific steps are such as
Under：
S141 calculates the mean value of the shape vector collection S after Procrustes is analyzed：
S142 calculates the covariance matrix of shape vector collection：
Covariance matrix is carried out Eigenvalues Decomposition, obtains feature vector p by S143iCharacteristic value corresponding with its, and by shape
The difference value vector dS of shape vector and mean shape vectoriIt is expressed as the linear combination of principal component：
dSi=λi1p1+λi2p2+...+λi2mp2m
S144 enables P=(p1,p2,...,p2m), λi=(λi1,λi2,...,λi2m)T, can obtain：
dSi=P λi
S145 selects the corresponding feature vector P of preceding t characteristic value in feature value vectorSAs main shaft, obtain：
Wherein bsIt is weight vector, changes each component therein and can be obtained by new shape, limits component certain
Section variation can ensure new shape close to sample shape.B under normal circumstancessValue be：
In this way, the shape that arbitrary sample is concentrated can be by the linear weighted function approximation table of average shape and t feature vector
Show.
In some instances, abovementioned steps S15, the local classifiers of training single feature point, specific construction step is such as
Under：
S151, extracts p*p size areas around each characteristic point, is write the gray value in this region as column vector in order
Local feature of the form as this characteristic point the form of column vector is write as the depth value in this region for depth image
As local feature.
S152, a shared n sample, each sample have m characteristic point, then for each characteristic point, there are 2n parts
Feature, including n gray scale local feature and n depth local feature, respectively using this n local feature as proper characteristics, then
W local feature is randomly selected in the other positions of gray level image and depth image as error characteristic, then each feature respectively
Point possesses n+w gray scale local feature and n+w depth local feature, wherein comprising mistake and correct, illustrates ash first
Feature is spent, for each characteristic point, its training sample set is：
{x(1),x(2),...,x(n+w)}
Each training sample is a local feature vectors：
The output of S153, SVM are y(i)={ -1,1 }, 1 indicates correct sample, and -1 indicates the sample of mistake.Pass through input
The output of the interior product representation SVM of data set and support vector：
Wherein xjIndicate the subset of training sample, αjIndicate the weights of support vector, NsIndicate the number of support vector, b tables
Show deviation.
The output of S154, SVM can be expressed as the linear combination of input vector：
y(i)=wT·x(i)+θ
Wherein wT=(w1,w2,...,wn+w) indicating the weights of input element, θ indicates deviation, according to input data set and defeated
Go out function, w can be found outTAnd θ.
S155, the gray scale local feature collection that m characteristic point is respectively trained in repetition above step obtain m Linear SVM.
S156 trains the depth local feature collection of m characteristic point to obtain m Linear SVM again according to above step.
Step S2 acquires two dimensional image and depth image by Kinect cameras, and is sent into computer disposal, Kinect
Coloured image and depth image can be acquired simultaneously, due to having a certain distance between RGB cameras and depth camera, acquired
Two width pictures be not complete correspondence, need to carry out registration process, the function in library of being increased income using OPENNI is carried out at alignment
Reason, and RGB image is converted into gray level image.
Step S3, to the image collected handle and quickly obtains using Viola-Jones detectors in a computer
The position of face, it is assumed here that surround the face detected with a rectangle frame, as shown in Figure 6.
Step S4 detects characteristic point, in some instances, the detection using CLM algorithms on the face detected
Steps are as follows：
S41 does initial estimation to facial feature points, and average facial contours are put to the face position detected in step s3
It sets, the initial estimation as facial feature points.
S42 extracts the image block of (t+p/2) * (t+p/2) size around a characteristic point current location, for gray-scale map
As extraction gray level image block, depth image block extracted for depth image, t is indicated around search current location within the scope of t*t most
Good characteristic point position, p indicate the size of the topography of training.
S43 illustrates the operation for gray level image block first, and the gray level image block of extraction is corresponding by this feature point
Linear SVM obtains a response surface design R (x, y), indicates surrounding's image block of every bit probability similar with template.
S44 is fitted a quadratic function r (x, y) to R (x, y).
S45 finds out the global maximum point of quadratic function r (x, y), and it is the most similar to template to represent the regions p*p around the point,
Depth image optimal match point in the block is equally found out, by best in the optimal match point and depth image in gray level image
Next optimum position of the midpoint with point as this characteristic point.
S46 finds out next optimum position of all characteristic points, completes an iteration.
S47, repeat S42-S46 iterative process, until characteristic point twice between displacement distance be less than a threshold value or
Reach certain iterations, so far thinks that all characteristic points have found best position.
S5, by detecting the facial feature estimation head pose completed.As shown in figure 3, head pose is divided into roll angle, rotation
Angle, pitch angle.The method of estimation of three kinds of angles is introduced individually below：
S51, as shown in Fig. 7 (a), certain roll angle is presented in head, if two center point coordinates are respectively (x1,y1),(x2,
y2), then head roll angle can be by eye line and horizontal angle estimation：
S52, as shown in Fig. 7 (b), certain angle of rotation is presented in head, if two external eyes angular coordinates are respectively (x1,y1),
(x2,y2), nose coordinate is (x3,y3), then the angle difference of nose and two tail of the eyes is：
The angle difference β of nose and two inner eye corners is calculated simultaneously2With the angle difference β of nose and two corners of the mouths3, take three angles
Average value as head angle of rotation β estimation：
S53, as shown in Fig. 7 (c), certain pitch angle is presented in head, and the calculating of pitch angle relies solely on characteristic point seat
Mark is difficult estimation, needs to use depth data, if eyes and face are in same upright plane when appearance without a head, i.e., they are from camera
Distance it is identical, make when facing upward head action, eyes and camera distance are d1, face is d with camera distance2If head is average
Radius is r, and point coordinates is (x in two lines1,y1), lip center point coordinate is (x2,y2), then the estimation formulas of pitch angle
For：
The results are shown in Figure 8 for the facial feature points and head pose wherein detected.
Step S6 is controlled the movement of Information Mobile Service robot by the head pose angle of estimation, is as follows：
S61, if facing upward brilliance degree is more than given threshold χy, then advance command is sent out,
If angle of bowing is more than given threshold χd, then backward command is sent out.
S62, if head turns left, angle is more than given threshold βl, then flicker order is sent out,
If head turns right, angle is more than given threshold βr, then flicker order is sent out.
S63 is considered as appearance state without a head, sends out and cease and desist order if institute is angled to be respectively less than corresponding threshold value.
Control method in conjunction with described in above example, as shown in Figure 1, Figure 2, the interior moderate in illuminance,
User keeps right sitting position gesture, Kinect cameras to be placed on 40-50 centimeters in front of head on intelligent wheel chair, opens intelligence
The head appearance control function of wheelchair, user's new line to certain angle, wheelchair travel forward, and user bows to certain angle,
Wheelchair moves backward, and user head turns left certain angle, and wheelchair is turned left, and user head goes to the right certain angle
Degree, wheelchair are turned right, and user keeps appearance without a head, that is, face front, at this time wheelchair stop motion.
Although the present invention has been disclosed as a preferred embodiment, however, it is not to limit the invention.Skill belonging to the present invention
Has usually intellectual in art field, without departing from the spirit and scope of the present invention, when can be used for a variety of modifications and variations.Cause
This, the scope of protection of the present invention is defined by those of the claims.
Claims (3)
1. a kind of intelligent Mobile Service robot control method based on head pose identification, which is characterized in that this method is with about
Based on beam partial model algorithm, obtains head pose and control Information Mobile Service robot, specific implementation includes the following steps：
The local feature template of step 1, the shape template that facial feature points are built by head pose sample database and characteristic point；
Step 2 acquires the two dimensional image comprising face and depth image by RGBD photographic devices and is aligned；
Step 3 detects face location in the picture；
Step 4 detects characteristic point using CLM algorithms on face；
Step 5 estimates head pose by the shape of the human face characteristic point detected；
Step 6 realizes the control to Information Mobile Service robot by head pose recognition result；
In the step 1, the local feature template of the shape template and each characteristic point of facial feature points is built respectively, it is specific to wrap
Containing following steps：
(1) sample set of construction feature dot shape carries out Procrustes analyses and eliminates displacement, rotation, and the influence of scaling carries out
PCA dimension-reduction treatment；
(2) to each characteristic point, the local classifiers of this feature point are trained using corresponding image block；
In the step 4, by estimating characteristic point initial position and being detected using CLM algorithm iterations characteristic point to optimum position
The position of all characteristic points in face；
It is poor by the angle between nose and two by two lines and horizontal line angle estimation roll angle in the step 5
Estimate angle of rotation, estimates pitch angle with the distance between camera respectively by lip and eyes.
2. the intelligent Mobile Service robot control method according to claim 1 based on head pose identification, feature
It is, in abovementioned steps 3, face location is quickly detected in the picture using Viola-Jones detectors.
3. the intelligent Mobile Service robot control method according to claim 1 based on head pose identification, feature
It is, in abovementioned steps 6, controls advance by the pitch angle on head and retreat, control by the angle of rotation on head
And right-hand rotation, and control command is sent to the next machine controller by serial ports.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510872912.2A CN105487665B (en) | 2015-12-02 | 2015-12-02 | A kind of intelligent Mobile Service robot control method based on head pose identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510872912.2A CN105487665B (en) | 2015-12-02 | 2015-12-02 | A kind of intelligent Mobile Service robot control method based on head pose identification |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105487665A CN105487665A (en) | 2016-04-13 |
| CN105487665B true CN105487665B (en) | 2018-09-07 |
Family
ID=55674689
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510872912.2A Active CN105487665B (en) | 2015-12-02 | 2015-12-02 | A kind of intelligent Mobile Service robot control method based on head pose identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105487665B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105773633B (en) * | 2016-04-14 | 2018-04-20 | 中南大学 | Mobile robot man-machine control system based on face location and sensitivity parameter |
| CN105912120B (en) * | 2016-04-14 | 2018-12-21 | 中南大学 | Mobile robot man-machine interaction control method based on recognition of face |
| CN106974780B (en) * | 2017-03-13 | 2018-06-29 | 邝子佳 | Method for controlling intelligent wheelchair based on difference navigation attitude |
| CN107349570A (en) * | 2017-06-02 | 2017-11-17 | 南京邮电大学 | Rehabilitation training of upper limbs and appraisal procedure based on Kinect |
| CN107358154A (en) * | 2017-06-02 | 2017-11-17 | 广州视源电子科技股份有限公司 | A kind of head movement detection method and device and vivo identification method and system |
| CN107493531B (en) * | 2017-08-04 | 2019-11-08 | 歌尔科技有限公司 | A kind of head pose detection method, device and earphone |
| CN107621880A (en) * | 2017-09-29 | 2018-01-23 | 南京邮电大学 | A kind of robot wheel chair interaction control method based on improvement head orientation estimation method |
| CN107754307A (en) * | 2017-12-05 | 2018-03-06 | 野草莓影业（北京）有限公司 | Control method, control device and the swiveling seat of swiveling seat |
| US10528802B2 (en) * | 2018-01-09 | 2020-01-07 | Futurewei Technologies, Inc. | Head pose and distraction estimation |
| CN108427918B (en) * | 2018-02-12 | 2021-11-30 | 杭州电子科技大学 | Face privacy protection method based on image processing technology |
| CN108711175B (en) * | 2018-05-16 | 2021-10-01 | 浙江大学 | Head attitude estimation optimization method based on interframe information guidance |
| CN109086727B (en) * | 2018-08-10 | 2021-04-30 | 北京奇艺世纪科技有限公司 | Method and device for determining motion angle of human head and electronic equipment |
| CN109993073B (en) * | 2019-03-14 | 2021-07-02 | 北京工业大学 | Leap Motion-based complex dynamic gesture recognition method |
| CN110097024B (en) * | 2019-05-13 | 2020-12-25 | 河北工业大学 | Human body posture visual recognition method of transfer, transportation and nursing robot |
| CN110909596B (en) * | 2019-10-14 | 2022-07-05 | 广州视源电子科技股份有限公司 | Side face recognition method, device, equipment and storage medium |
| CN111346356B (en) * | 2020-03-09 | 2021-05-04 | 牡丹江医学院 | Sports teaching apparatus |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101968846A (en) * | 2010-07-27 | 2011-02-09 | 上海摩比源软件技术有限公司 | Face tracking method |
| CN103870843A (en) * | 2014-03-21 | 2014-06-18 | 杭州电子科技大学 | Head posture estimation method based on multi-feature-point set active shape model (ASM) |
| CN104463100A (en) * | 2014-11-07 | 2015-03-25 | 重庆邮电大学 | Intelligent wheelchair man-machine interaction system and method based on facial expression recognition mode |
-
2015
- 2015-12-02 CN CN201510872912.2A patent/CN105487665B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101968846A (en) * | 2010-07-27 | 2011-02-09 | 上海摩比源软件技术有限公司 | Face tracking method |
| CN103870843A (en) * | 2014-03-21 | 2014-06-18 | 杭州电子科技大学 | Head posture estimation method based on multi-feature-point set active shape model (ASM) |
| CN104463100A (en) * | 2014-11-07 | 2015-03-25 | 重庆邮电大学 | Intelligent wheelchair man-machine interaction system and method based on facial expression recognition mode |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105487665A (en) | 2016-04-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105487665B (en) | A kind of intelligent Mobile Service robot control method based on head pose identification | |
| US11699293B2 (en) | Neural network image processing apparatus | |
| TWI383325B (en) | Face expressions identification | |
| CN103177269B (en) | For estimating the apparatus and method of object gesture | |
| CN105005999B (en) | It is a kind of based on obstacle detection method of the computer stereo vision towards apparatus for guiding blind | |
| CN108345869A (en) | Driver's gesture recognition method based on depth image and virtual data | |
| CN110147738B (en) | Driver fatigue monitoring and early warning method and system | |
| CN106796449A (en) | Eye-controlling focus method and device | |
| EP3893090B1 (en) | Method for eye gaze tracking | |
| CN105159452B (en) | A kind of control method and system based on human face modeling | |
| CN111046734A (en) | Multi-modal fusion sight line estimation method based on expansion convolution | |
| CN112232128B (en) | Eye tracking based method for identifying care needs of old disabled people | |
| WO2022042203A1 (en) | Human body key point detection method and apparatus | |
| CN108256379A (en) | A kind of eyes posture identification method based on Pupil diameter | |
| CN113963416A (en) | Eye movement interaction method and system based on laser visual feedback | |
| CN111626226B (en) | Neck rotation recognition method and system | |
| Rahmaniar et al. | Touchless head-control (thc): Head gesture recognition for cursor and orientation control | |
| CN114639168B (en) | Method and system for recognizing running gesture | |
| CN110826495A (en) | Body left and right limb consistency tracking and distinguishing method and system based on face orientation | |
| WO2020065790A1 (en) | Estimation device, estimation method, and storage medium | |
| WO2023103145A1 (en) | Head pose truth value acquisition method, apparatus and device, and storage medium | |
| Gruendig et al. | 3d head pose estimation with symmetry based illumination model in low resolution video | |
| JP5688514B2 (en) | Gaze measurement system, method and program | |
| Pangestu et al. | Electric Wheelchair Control Mechanism Using Eye-mark Key Point Detection. | |
| Zhou et al. | Eye-tracking Control Wheelchair Method Based on Landmarks of Face |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| EE01 | Entry into force of recordation of patent licensing contract | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20160413 Assignee: Zhangjiagang Institute of Zhangjiagang Assignor: Nanjing Post & Telecommunication Univ. Contract record no.: X2019980001251 Denomination of invention: Method for controlling intelligent mobile service robot based on head posture recognition Granted publication date: 20180907 License type: Common License Record date: 20191224 |