CN107784280A - A kind of dynamic pupil tracking method - Google Patents

A kind of dynamic pupil tracking method Download PDF

Info

Publication number
CN107784280A
CN107784280A CN201710973082.1A CN201710973082A CN107784280A CN 107784280 A CN107784280 A CN 107784280A CN 201710973082 A CN201710973082 A CN 201710973082A CN 107784280 A CN107784280 A CN 107784280A
Authority
CN
China
Prior art keywords
mtd
pupil
mtr
mrow
tracking
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.)
Withdrawn
Application number
CN201710973082.1A
Other languages
Chinese (zh)
Inventor
黄靖宇
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.)
Zhangjiagang Quan Zhi Electronic Technology Co Ltd
Original Assignee
Zhangjiagang Quan Zhi Electronic Technology 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 Zhangjiagang Quan Zhi Electronic Technology Co Ltd filed Critical Zhangjiagang Quan Zhi Electronic Technology Co Ltd
Priority to CN201710973082.1A priority Critical patent/CN107784280A/en
Publication of CN107784280A publication Critical patent/CN107784280A/en
Withdrawn legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/18—Eye characteristics, e.g. of the iris
    • G06V40/193—Preprocessing; Feature extraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/207—Analysis of motion for motion estimation over a hierarchy of resolutions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/24—Aligning, centring, orientation detection or correction of the image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/62—Extraction of image or video features relating to a temporal dimension, e.g. time-based feature extraction; Pattern tracking

Abstract

The present invention discloses a kind of dynamic pupil tracking method, comprises the following steps：1) in image sequence, it is assumed that tracking target t coordinate beMotion prediction is carried out using Kalman filtering, then the coordinate at t+1 moment isCovariance is ∑ (xk, yk)；2) during pupil tracking, the motion of pupil in front and rear two frame is regarded as at the uniform velocity, the feature available position of pupil movement and speed describe, if (ct, rt) be t pupil position, (ut, vt) it is speed of the t in c directions and r directions, so the state vector of t pupil is xt=(ct, tt, ut, vt)t, the state model of system is expressed as：xt+1=Φ xi+wi；Wherein wtFor system noise；The dynamic pupil tracking method accuracy of detection is high and real-time.

Description

A kind of dynamic pupil tracking method
Technical field
Present invention relates particularly to a kind of dynamic pupil tracking method.
Background technology
The process that computer vision system gets up the feature in image from piece image to another width images match is referred to as The tracking of characteristics of image.Signature tracking technology includes based drive method and the method based on template, and the former is using motion point Cut the tracking target motion such as technology, Kalman filtering；The latter obtains the priori of target first, constructs object module, then Template matches are carried out by sliding window to each two field picture of input.
The position of pupil is assessed in the tracking of pupil real-time continuously in one group of image sequence.At present, for face and The detection of pupil is with tracking oneself through there is substantial amounts of achievement in research.But some of track algorithm precision are higher and real-time compared with Difference, some algorithm real-times can meet and precision is not high.
The content of the invention
It is an object of the invention to provide a kind of high and real-time dynamic pupil tracking method of accuracy of detection.
In order to solve the above-mentioned technical problem, the technical scheme is that：
A kind of dynamic pupil tracking method, comprises the following steps：
1) in image sequence, it is assumed that tracking target t coordinate beTransported using Kalman filtering Dynamic prediction, then the coordinate at t+1 moment isCovariance is ∑ (xk, yk)；
2) during pupil tracking, the motion of pupil in front and rear two frame is regarded as at the uniform velocity, the feature of pupil movement Available position and speed describe, if (ct, rt) be t pupil position, (ut, vt) it is t in c directions and r directions Speed, so the state vector of t pupil is xt=(ct, tt, ut, vt)t, the state model of system is expressed as：xt+1=Φ xt +wt；Wherein wtFor system noise；
3) when pupil at the uniform velocity moves between two field pictures, it is set as in state-transition matrix：
Observed quantityFor the position of t pupil, the measurement model of system is：zt=Hxt+vt, wherein vt For the white noise of zero-mean.
The technology of the present invention effect major embodiment is in the following areas：First have to detect and position in initial frame when being tracked Go out the position of pupil, and construct pupil template, tracked target is then estimated in next frame according to the movable information of image Position, and scanned in the region that is likely to occur of estimation target.Obtain and what To Template distribution of color was most like is For tracked target.Due to having carried out motion estimation, so as to substantially reduce hunting zone, so compared with using exhaustive search Algorithm is more fast and effective.
Embodiment
In the present embodiment, it is necessary to which explanation, such as first and second or the like relational terms are used merely to one Individual entity or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or operate it Between any this actual relation or order be present.Moreover, term " comprising ", "comprising" or its any other variant are intended to Cover including for nonexcludability, so that process, method, article or equipment including a series of elements not only include those Key element, but also the other element including being not expressly set out, or also include for this process, method, article or set Standby intrinsic key element.
In addition, the connection between part or fixed form if not otherwise specified in this embodiment, it is connected or solid It can be to be fixed by bolt commonly used in the prior art or pin is fixed to determine mode, or the mode such as bearing pin connection, therefore, at this No longer it is described in detail in embodiment.
Embodiment
A kind of dynamic pupil tracking method, in image sequence, it is assumed that tracking target t coordinate beAdopt Motion prediction is carried out with Kalman filtering, then the coordinate at t+1 moment isCovariance is ∑ (xk, yk)；Pupil with During track, the motion of pupil in front and rear two frame is regarded as at the uniform velocity, the feature available position of pupil movement and speed are retouched State, if (ct, rt) be t pupil position, (ut, vt) it is speed of the t in c directions and r directions, so t pupil State vector be xt=(ct, tt, ut, vt)t, the state model of system is expressed as：xt+1=Φ xt+wt；Wherein wtFor system noise Sound；It is assumed that the displacement very little that pupil moves between two field pictures, and be at the uniform velocity to move, it is set as in state-transition matrix：Observed quantityFor the position of t pupil, the measurement model of system is：zt =Hxt+vt, wherein vtFor the white noise of zero-mean.
Due to ztIt is only relevant with position, can set H asIf provide target initial position and Speed, in aforementioned manners can estimate the state vector x of pupil in next two field picturet, and covariance matrix ∑t+1 (xt+1, yt+1)。
The technology of the present invention effect major embodiment is in the following areas：First have to detect and position in initial frame when being tracked Go out the position of pupil, and construct pupil template, tracked target is then estimated in next frame according to the movable information of image Position, and scanned in the region that is likely to occur of estimation target.Obtain and what To Template distribution of color was most like is For tracked target.Due to having carried out motion estimation, so as to substantially reduce hunting zone, so compared with using exhaustive search Algorithm is more fast and effective.
The foregoing is only a specific embodiment of the invention, but protection scope of the present invention is not limited thereto, any The change or replacement expected without creative work, it should all be included within the scope of the present invention.

Claims (1)

  1. A kind of 1. dynamic pupil tracking method, it is characterised in that comprise the following steps：
    1) in image sequence, it is assumed that tracking target t coordinate beUsing Kalman filtering move pre- Survey, then the coordinate at t+1 moment isCovariance is ∑ (xk, yk)；
    2) during pupil tracking, the motion of pupil in front and rear two frame is regarded at the uniform velocity as the feature of pupil movement can use Position and speed describe, if (ct, rt) be t pupil position, (ut, vt) it is speed of the t in c directions and r directions Degree, so the state vector of t pupil is xt=(ct, tt, ut, vt)t, the state model of system is expressed as：xt+1=Φ xt+ wt；Wherein wtFor system noise；
    3) when pupil at the uniform velocity moves between two field pictures, it is set as in state-transition matrix：
    <mrow> <mi>&amp;Phi;</mi> <mo>=</mo> <mfenced open = "[" close = "]"> <mtable> <mtr> <mtd> <mn>1</mn> </mtd> <mtd> <mn>0</mn> </mtd> <mtd> <mrow> <mi>&amp;Delta;</mi> <mi>t</mi> </mrow> </mtd> <mtd> <mn>0</mn> </mtd> </mtr> <mtr> <mtd> <mn>0</mn> </mtd> <mtd> <mn>1</mn> </mtd> <mtd> <mn>0</mn> </mtd> <mtd> <mrow> <mi>&amp;Delta;</mi> <mi>t</mi> </mrow> </mtd> </mtr> <mtr> <mtd> <mn>0</mn> </mtd> <mtd> <mn>0</mn> </mtd> <mtd> <mn>1</mn> </mtd> <mtd> <mn>0</mn> </mtd> </mtr> <mtr> <mtd> <mn>0</mn> </mtd> <mtd> <mn>0</mn> </mtd> <mtd> <mn>0</mn> </mtd> <mtd> <mn>1</mn> </mtd> </mtr> </mtable> </mfenced> <mo>;</mo> </mrow>
    Observed quantityFor the position of t pupil, the measurement model of system is：zt=Hxt+vt, wherein vtIt is zero equal The white noise of value.
CN201710973082.1A 2017-10-18 2017-10-18 A kind of dynamic pupil tracking method Withdrawn CN107784280A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710973082.1A CN107784280A (en) 2017-10-18 2017-10-18 A kind of dynamic pupil tracking method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710973082.1A CN107784280A (en) 2017-10-18 2017-10-18 A kind of dynamic pupil tracking method

Publications (1)

Publication Number Publication Date
CN107784280A true CN107784280A (en) 2018-03-09

Family

ID=61434738

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710973082.1A Withdrawn CN107784280A (en) 2017-10-18 2017-10-18 A kind of dynamic pupil tracking method

Country Status (1)

Country Link
CN (1) CN107784280A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108595008A (en) * 2018-04-27 2018-09-28 北京计算机技术及应用研究所 Man-machine interaction method based on eye movement control
CN112749604A (en) * 2019-10-31 2021-05-04 Oppo广东移动通信有限公司 Pupil positioning method and related device and product
CN113838086A (en) * 2021-08-23 2021-12-24 广东电网有限责任公司 Attention assessment test method, attention assessment test device, electronic equipment and storage medium
CN115147462A (en) * 2022-07-08 2022-10-04 浙江大学 Gaze characteristic tracking method based on three-dimensional eyeball model and Kalman filtering

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101699510A (en) * 2009-09-02 2010-04-28 北京科技大学 Particle filtering-based pupil tracking method in sight tracking system
CN105373766A (en) * 2014-08-14 2016-03-02 由田新技股份有限公司 Pupil positioning method and device

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101699510A (en) * 2009-09-02 2010-04-28 北京科技大学 Particle filtering-based pupil tracking method in sight tracking system
CN105373766A (en) * 2014-08-14 2016-03-02 由田新技股份有限公司 Pupil positioning method and device

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
党治: "实时瞳孔检测与跟踪方法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108595008A (en) * 2018-04-27 2018-09-28 北京计算机技术及应用研究所 Man-machine interaction method based on eye movement control
CN112749604A (en) * 2019-10-31 2021-05-04 Oppo广东移动通信有限公司 Pupil positioning method and related device and product
CN113838086A (en) * 2021-08-23 2021-12-24 广东电网有限责任公司 Attention assessment test method, attention assessment test device, electronic equipment and storage medium
CN113838086B (en) * 2021-08-23 2024-03-22 广东电网有限责任公司 Attention assessment test method, device, electronic equipment and storage medium
CN115147462A (en) * 2022-07-08 2022-10-04 浙江大学 Gaze characteristic tracking method based on three-dimensional eyeball model and Kalman filtering

Similar Documents

Publication Publication Date Title
Henein et al. Dynamic SLAM: The need for speed
CN107784280A (en) A kind of dynamic pupil tracking method
US9607401B2 (en) Constrained key frame localization and mapping for vision-aided inertial navigation
Nerurkar et al. C-KLAM: Constrained keyframe-based localization and mapping
CN103383776B (en) A kind of laddering Stereo Matching Algorithm based on two stage cultivation and Bayesian Estimation
US20210150755A1 (en) Device and method with simultaneous implementation of localization and mapping
Dani et al. Globally exponentially stable observer for vision-based range estimation
CN102405483A (en) Object tracking device, object tracking method, and object tracking program
Yoon et al. Object tracking from image sequences using adaptive models in fuzzy particle filter
JP6061770B2 (en) Camera posture estimation apparatus and program thereof
Shin et al. Hybrid approach for facial feature detection and tracking under occlusion
KR20120094102A (en) Similarity degree calculation device, similarity degree calculation method, and program
Koledić et al. MOFT: Monocular odometry based on deep depth and careful feature selection and tracking
Nitsche et al. Constrained-covisibility marginalization for efficient on-board stereo SLAM
Shang et al. Lane detection using steerable filters and FPGA-based implementation
CN108731683B (en) Unmanned aerial vehicle autonomous recovery target prediction method based on navigation information
Huang et al. MC-VEO: A Visual-Event Odometry With Accurate 6-DoF Motion Compensation
Lin et al. A New Approach for Vision-based Rear Vehicle Tracking
CN101976446A (en) Tracking method of multiple feature points of microscopic sequence image
CN111784680B (en) Detection method based on consistency of key points of left and right eye views of binocular camera
JP2012215549A (en) Tracking device
Leykin et al. Real-time estimation of human attention field in LWIR and color surveillance videos
Pan et al. Correlation tracking algorithm based on adaptive template update
Teshima et al. Estimation of FOE Without Optical Flow for ehicle Lateral Position Detection.
Askar et al. Optimized uav object tracking framework based on integrated particle filter with ego-motion transformation matrix

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
WW01 Invention patent application withdrawn after publication

Application publication date: 20180309

WW01 Invention patent application withdrawn after publication