CN110362264A - A kind of dynamic gesture identification method based on touch tablet, device and system - Google Patents

A kind of dynamic gesture identification method based on touch tablet, device and system Download PDF

Info

Publication number
CN110362264A
CN110362264A CN201910579656.6A CN201910579656A CN110362264A CN 110362264 A CN110362264 A CN 110362264A CN 201910579656 A CN201910579656 A CN 201910579656A CN 110362264 A CN110362264 A CN 110362264A
Authority
CN
China
Prior art keywords
moment
vector
motion
sequence
gesture
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.)
Granted
Application number
CN201910579656.6A
Other languages
Chinese (zh)
Other versions
CN110362264B (en
Inventor
李林峰
林碧云
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.)
Wuhan Haiwei Technology Co ltd
Original Assignee
Wuhan Haiwei 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 Wuhan Haiwei Technology Co Ltd filed Critical Wuhan Haiwei Technology Co Ltd
Priority to CN201910579656.6A priority Critical patent/CN110362264B/en
Publication of CN110362264A publication Critical patent/CN110362264A/en
Application granted granted Critical
Publication of CN110362264B publication Critical patent/CN110362264B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • G06F18/241Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00Input 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
    • G06F3/01Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/048Interaction techniques based on graphical user interfaces [GUI]
    • G06F3/0487Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser
    • G06F3/0488Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser using a touch-screen or digitiser, e.g. input of commands through traced gestures
    • G06F3/04883Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser using a touch-screen or digitiser, e.g. input of commands through traced gestures for inputting data by handwriting, e.g. gesture or text
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44Arrangements for executing specific programs
    • G06F9/451Execution arrangements for user interfaces

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Human Computer Interaction (AREA)
  • Probability & Statistics with Applications (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Image Analysis (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

The present invention relates to a kind of dynamic gesture identification methods based on touch tablet, device and system, including obtain the corresponding coordinate point sequence of gesture on touch tablet;According to the coordinate points at each adjacent two moment in coordinate point sequence to obtaining direction of motion sequence vector;Whether the type that gesture is judged according to direction of motion sequence vector is picture circle operation, if, then according to the direction of motion vector between the two neighboring moment all in direction of motion sequence vector, obtain the center-of-mass coordinate and the direction of motion of picture circle operation, and type and the output of picture circle operation are identified according to center-of-mass coordinate and the direction of motion, if not, touch tablet is left to finger, multiple characteristic vector sequences are extracted from direction of motion sequence vector, and type and the output of gesture are identified according to all characteristic vector sequences.Calculating is enormously simplified, recognition efficiency is enhanced, gesture identification can be exported in real time as a result, user experience is strong, also can be integrated into small embedded systems, the touch tablet especially suitable for vehicle steering wheel.

Description

A kind of dynamic gesture identification method based on touch tablet, device and system
Technical field
The present invention relates to human-computer interaction technique field more particularly to a kind of dynamic gesture identification method based on touch tablet, Device and system.
Background technique
In field of human-computer interaction, gesture refers to that the mankind have by what the combination of finger or the different gestures of palm was formed The set of the information of specific meanings.Currently, touch screen terminal using more more and more universal, also started setting up in vehicle steering wheel Touch tablet is operated with some functions to automobile, such as multimedia function (such as: music and volume of music are adjusted), Communication function (such as: receiving calls) and navigation feature.People can use gesture and operate, for example, user is in touch tablet The different touch gestures of upper operation (such as picture circle, horizontal stroke, perpendicular stroke and diagonally opposing corner are drawn), can represent different meanings, realization is not Same function (such as page turning, amplification, diminution and opening file), so that the experience sense of user obtains larger promotion.
In the prior art, the gesture reference template that the Gesture Recognition Algorithm based on touch tablet is mostly used and pre-set is straight The method of row template matching is tapped into identify, this method calculating is complex, and computationally intensive, recognition efficiency is not high, user experience Feel poor, for some small embedded systems, it has not been convenient to transplant, poor universality.
Summary of the invention
The technical problem to be solved by the present invention is to solve the above shortcomings of the prior art and to provide a kind of based on touch tablet Dynamic gesture identification method, device and system simplify calculating, enhance recognition efficiency, can export gesture identification knot in real time Fruit, user experience is strong, can be integrated into the small embedded systems such as single-chip microcontroller, the touch especially suitable for vehicle steering wheel Plate.
The technical scheme to solve the above technical problems is that
A kind of dynamic gesture identification method based on touch tablet, comprising the following steps:
Step 1: obtaining and acting on the corresponding coordinate point sequence of the gesture on touch tablet with finger;
Step 2: according to the coordinate points pair at each adjacent two moment in coordinate point sequence, obtaining direction of motion vector sequence Column;
Step 3: whether the type that the gesture is judged according to direction of motion sequence vector is picture circle operation, if so, carrying out Step 4, if it is not, carrying out step 5;
Step 4: according to the direction of motion vector between all two neighboring moment in direction of motion sequence vector, obtaining Center-of-mass coordinate and the direction of motion of the circle operation on the touch tablet are drawn, and picture circle is identified according to center-of-mass coordinate and the direction of motion The type of operation and output;
Step 5: leaving the touch tablet to the finger, multiple feature vector sequences are extracted from direction of motion sequence vector It arranges, and identifies type and the output of the gesture according to all characteristic vector sequences.
The beneficial effects of the present invention are: obtaining the coordinate point sequence of gesture on a touchpad, the coordinate point sequence packet first The corresponding coordinate points in touch point containing each moment, therefore pass through the coordinate points pair at each adjacent two moment, available movement Direction vector sequence;It (including operates, grasp downwards upwards since the gesture on touch tablet can be divided into the operation of picture circle and linear operating Make, several classes such as operate and operate to the right to the left), and draw that circle operation is different from the feature of linear operating, and recognition methods is also different, Therefore the direction of motion sequence vector can be first passed through, judges whether the type of gesture currently entered is picture circle operation, then Different recognition methods is taken according to the result judged to identify gesture and export, and is operated when judging that the type of gesture is justified for picture When, it can determine mass center and the movement side of picture circle operation respectively by the direction of motion vector between all two neighboring moment To can determine the type (draw clockwise round or draw counterclockwise circle) of picture circle operation and export;When the type for judging gesture It is operated for non-picture circle, then gesture is upward operation, downward operation, several classes, such gesture such as operates and operate to the right to the left and only need Determine that the direction of gesture can determine the type of gesture, therefore can be by extracting multiple spies in direction of motion sequence vector It levies sequence vector, then identifies that the direction of gesture may recognize that the type of gesture and defeated by all characteristic vector sequences Out;
The present invention is based on the dynamic gesture identification methods of touch tablet, by first carrying out just classification to gesture, by gesture classification Draw circle operation and the circle operation of non-pictures, then different recognition methods be respectively adopted to identify the concrete type of gesture, be not necessarily to it is preparatory The reference template of setting is matched, and is enormously simplified calculating, is enhanced recognition efficiency, can export gesture identification knot in real time Fruit, user experience is strong, also can be integrated into the small embedded systems such as single-chip microcontroller, the touch especially suitable for vehicle steering wheel Plate, to realize specific vehicle-mounted function.
Based on the above technical solution, the present invention can also be improved as follows:
Further, the specific steps of the step 2 include:
Step 201: according to the coordinate points of the coordinate points of the t moment in coordinate point sequence and t-1 moment, being calculated and sit The corresponding adjacent coordinates point distance of t moment in punctuation sequence;
Adjacent coordinates point distance between the coordinate points and the coordinate points at t-1 moment of t moment are as follows:
Wherein, d is adjacent coordinates point distance corresponding with t moment, xtIt is right on the touch tablet for finger described in t moment The x coordinate answered, ytFor finger described in t moment on the touch tablet corresponding y-coordinate, xt-1It is finger described in the t-1 moment in institute State corresponding x coordinate, y on touch tablett-1For finger described in the t-1 moment on the touch tablet corresponding y-coordinate;
Step 202: according to the method for the step 201, being calculated and corresponded with the moment each in coordinate point sequence Multiple adjacent coordinates point distances, and Screening Treatments are carried out to all adjacent coordinates points distance according to preset screening criterion, obtained To multiple effective coordinate points;
The screening criterion are as follows:
Wherein, dthreFor preset distance threshold, label (d) is the output function of the screening criterion, as label (d) Output valve when being 0 apart from the coordinate points at a corresponding moment be invalid sit with adjacent coordinates point then in coordinate point sequence Punctuate, when the output valve of label (d) is 1, then in coordinate point sequence, with adjacent coordinates point apart from a corresponding moment Coordinate points be effective coordinate points;
Step 203: all effective coordinate points in chronological sequence sequentially being sorted, effective coordinate point sequence is obtained;
Step 204: according to effective coordinate at -1 moment of effective coordinate points at the t ' moment in effective coordinate point sequence and t ' Point obtains direction of motion vector corresponding with the t ' moment;
Step 205: according to the method for the step 204, obtaining corresponding with the moment each in effective coordinate point sequence Multiple direction of motion vectors, obtain direction of motion sequence vector according to all direction of motion vectors.
The beneficial effect of above-mentioned further scheme is: during obtaining direction of motion sequence vector, passing through meter first The coordinate points of t moment and the distance between the coordinate points at t-1 moment in calculation coordinate point sequence, i.e. adjacent coordinates point distance d, thus Adjacent coordinates point distance corresponding with the moment each in coordinate point sequence can be obtained after the same method;Again by preset Distance threshold and preset screening criterion carry out Screening Treatment to all adjacent coordinates points distance in coordinate point sequence, can be with The coordinate points that distance is too small between the coordinate points at some two neighboring moment, i.e. invalid coordinates point are abandoned, are known so as to improve Other efficiency and accuracy rate;It is sorted, is obtained according to chronological order by all effective coordinate points that Screening Treatment obtains again Effective coordinate point sequence obtains between each adjacent two moment according to effective coordinate point sequence (such as effectively sitting convenient for subsequent Effective coordinate points of effective coordinate points at t ' moment and -1 moment of t ' in punctuation sequence) direction of motion vector, convenient for according to institute Some direction of motion vectors obtain direction of motion sequence vector, to facilitate the subsequent type according to direction of motion sequence to gesture It whether is that picture circle operation is judged;
Wherein, adjacent coordinates point distance corresponding with t moment in coordinate point sequence, when referring to the coordinate points and t-1 of t moment The distance between the coordinate points at quarter, therefore adjacent coordinates point distance corresponding with the moment each in coordinate point sequence, each mean seat The distance between the coordinate points of the coordinate points at a moment and a moment thereon in punctuation sequence;Similarly, with effective coordinate points T ' moment corresponding direction of motion vector in sequence, refer to the t ' moment effective coordinate points and -1 moment of t ' effective coordinate points it Between direction of motion vector, therefore direction of motion vector corresponding with the moment each in effective coordinate point sequence each means effectively Effective coordinate points at moment and the direction of motion vector between effective coordinate points at a moment thereon in coordinate point sequence; Wherein, when the output valve of label (d) is 0, then when with an adjacent coordinates point in coordinate point sequence apart from corresponding one The coordinate points at quarter are invalid coordinates point, for example, carrying out when to adjacent coordinates point corresponding with the t moment distance in coordinate point sequence Screening Treatment, when the value for obtaining its label (d) is 0, then the coordinate points of t moment are invalid coordinates point;When the output of label (d) When value is 1, then coordinate points at the time of corresponding with an adjacent coordinates point distance in coordinate point sequence are effective coordinate points, example Such as, when carrying out Screening Treatment to adjacent coordinates point corresponding with the t-1 moment distance in coordinate point sequence, its label (d) is obtained Value be 1 when, then the coordinate points at t-1 moment be effective coordinate points.
Further, the specific steps of the step 3 include:
Step 301: according to the movement at -1 moment of direction of motion vector sum t ' at the t ' moment in direction of motion sequence vector It is poor that movement direction angle corresponding with the t ' moment in direction of motion sequence vector is calculated in direction vector;
Movement direction angle between the direction of motion vector at -1 moment of direction of motion vector and t ' at t ' moment is poor are as follows:
θ=angel (vt′)-angel(vt′-1); (3)
Wherein, θ is that movement direction angle corresponding with the t ' moment is poor, vt′For the direction of motion vector at t ' moment, vt′-1For The direction of motion vector at -1 moment of t ', angel (vt′) be the t ' moment direction of motion vector angle, angel (vt′-1) it is t '- The angle of the direction of motion vector at 1 moment;
Step 302: according to the method for the step 301, respectively obtain t ' moment in direction of motion sequence vector it Preceding continuous multiple moment, corresponding multiple movement direction angles were poor;
Step 303: being judged according to the positive-negative polarity of all movement direction angle differences, when movement direction angle difference is continuous When the number for being positive or being continuously negative is greater than or equal to preset times, then judges that the type of the gesture operates for picture circle, go forward side by side Otherwise row step 4 carries out step 5.
The beneficial effect of above-mentioned further scheme is: due to the operation of picture circle each moment and its last moment it Between movement direction angle difference can continuously be positive or continuously be negative, therefore, as t ' moment and t ' -1 in direction of motion sequence vector Movement direction angle difference between moment is positive, and (movement direction angle i.e. corresponding with the t ' moment in direction of motion sequence vector is poor It is positive), and the number that is positive of t ' moment corresponding movement direction angle difference and continuous multiple moment pair before the t ' moment When the total degree that the polarity for the continuous multiple movement direction angle differences answered continuously is positive is greater than or equal to preset times, for example, it is default Number is 5 times, and -1 moment of t ' corresponding movement direction angle is poor, movement direction angle corresponding with -2 moment of t ' is poor and t ' - The polarity of 3 moment corresponding movement direction angle difference and movement direction angle difference corresponding with -4 moment of t ' is also positive, then may be used Judge that the type of the gesture operates for picture circle;Alternatively, when the fortune between -1 moment of t ' moment and t ' in direction of motion sequence vector Dynamic orientation angle difference is negative (movement direction angle difference i.e. corresponding with the t ' moment in direction of motion sequence vector is negative), and when t ' It is corresponding with continuous multiple moment before the t ' moment continuous multiple to carve the number that corresponding movement direction angle difference is negative When the number that the polarity of movement direction angle difference is continuously negative is greater than or equal to preset times, such as preset times are 5 times, with - 1 moment of t ' corresponding movement direction angle is poor and -2 moment of t ' corresponding movement direction angle is poor, corresponding with -3 moment of t ' The polarity of movement direction angle difference and movement direction angle difference corresponding with -4 moment of t ' is also negative, then can determine whether the gesture Type is picture circle operation;Therefore the direction of motion vector at -1 moment of direction of motion vector sum t ' corresponding with the t ' moment is obtained first Between movement direction angle it is poor, i.e., movement direction angle corresponding with the t ' moment is poor, calculates the t ' moment further according to formula (3) Corresponding multiple movement direction angles of continuous multiple moment before are poor, finally judge that all movement direction angle differences are continuously Whether the number just or being continuously negative is greater than or equal to preset times, that is, can determine whether the type of gesture is picture circle operation;
Judge whether gesture-type is picture circle operation, and method is simple, and calculation amount is small, and facilitates subsequent through the above steps It is judged that result identifies gesture-type to take different recognition methods, recognition efficiency height is improved, recognition result is more acurrate.
Further, in the step 4, the specific steps for obtaining center-of-mass coordinate and the direction of motion include:
Step 401: one by one according to the direction of motion vector between the two neighboring moment all in direction of motion sequence vector Corresponding effective coordinate points calculate the center-of-mass coordinate of picture circle operation;
Calculate the formula of center-of-mass coordinate are as follows:
Wherein, (xmc,ymc) it is center-of-mass coordinate, (xt′,yt′) it is that direction of motion vector corresponding with the t ' moment is corresponding effectively The coordinate of coordinate points, N are the sum of effective coordinate points;
Step 402: according to the seat of center-of-mass coordinate and the corresponding effective coordinate points of direction of motion vector corresponding with the t ' moment Mark, obtains the picture circular direction vector at t ' moment, corresponding according to center-of-mass coordinate and direction of motion vector corresponding with -1 moment of t ' The coordinate of effective coordinate points, obtains the picture circular direction vector at -1 moment of t ';
Step 403: according to preset direction of motion criterion, to -1 moment of angle and t ' of the picture circular direction vector at t ' moment The angle of picture circular direction vector judged, obtain the direction of motion of picture circle operation;
The direction of motion criterion are as follows:
Wherein, ut′For the picture circular direction vector at t ' moment, angel (ut′) be the t ' moment picture circular direction vector angle, ut′-1For the picture circular direction vector at -1 moment of t ', angel (ut′-1) be -1 moment of t ' picture circular direction vector angle, Clockwise is the output function of the direction of motion criterion, and when the output valve of Clockwise is 1, then the direction of motion is suitable Hour hands, when the output valve of Clockwise is 0, then the direction of motion is as you were, when the output valve of Clockwise is -1, Then the direction of motion is counterclockwise.
The beneficial effect of above-mentioned further scheme is: the direction of motion for drawing circle operation due to one only exists clockwise Or counterclockwise, the change direction between two moment of arbitrary neighborhood is necessarily consistent, therefore when step 3 has determined gesture class When type is picture circle operation, the direction of motion of picture circle operation can be determined according to the change direction between two moment of arbitrary neighborhood; By the direction of motion vector between the gesture corresponding all two neighboring moment, effective coordinate points are (i.e. correspondingly first Effective coordinate point sequence that effective coordinate points at all moment are constituted) center-of-mass coordinate is calculated, and when according to center-of-mass coordinate and with t ' The coordinate of effective coordinate points corresponding to the corresponding direction of motion vector carved obtains the picture circular direction vector at the t ' moment, and adopts The picture circular direction vector that -1 moment of t ' is obtained with same method, when the angle of the picture circular direction vector at the t ' moment is less than t ' -1 When the angle of the picture circular direction vector at moment, illustrate that the direction of motion of picture circle operation is clockwise, direction of motion criterion The output valve of Clockwise is 1;When the angle of the picture circular direction vector at the t ' moment is greater than the picture circular direction vector at -1 moment of t ' Angle when, illustrate picture circle operation the direction of motion be it is counterclockwise, the output valve of direction of motion criterion Clockwise is -1;And When the angle of the picture circular direction vector at the t ' moment is equal to the angle of -1 moment of t ' corresponding picture circular direction vector, illustrate picture circle The direction of motion of operation is (i.e. current time does not move temporarily) as you were, and the output valve of direction of motion criterion Clockwise is 0;The direction of motion obtained by the above-mentioned center-of-mass coordinate being calculated and direction of motion criterion can determine picture circle operation Concrete type, be it is counterclockwise draw round or draw circle clockwise, and recognition result is exported, recognition efficiency is high, accuracy rate is high.
Further, when judging that the type of the gesture operates for picture circle, further including following step before the step 4 It is rapid:
Step 411: according to predetermined angle, the touch tablet being divided into K touch area;And according to the step 2 Method obtains the direction of motion sequence vector in each touch area respectively;Wherein, K >=2;
Step 412: choosing preset hysteresis region in the critical zone of each touch area, and respectively in each hysteresis area Add hysteresis loop comparator in domain;
The then step 4 specifically:
Based on the hysteresis loop comparator of each touch area addition, according in direction of motion sequence vector in each touch area All two neighboring moment between direction of motion vector, respectively obtain the circle operation of each touch area inside-paint in the touch Center-of-mass coordinate and the direction of motion on plate, and respectively according in each touch area center-of-mass coordinate and the direction of motion identify pair The round type operated of the picture answered and output.
The beneficial effect of above-mentioned further scheme is:, can be according to pre- when step 3 judges that gesture-type operates for picture circle If touch tablet is divided into K touch area by angle, such as is divided according to 90 °, 4 touch areas, including 315 ° are obtained ~0 ° and 0 °~45 ° of 0 degree of region, 45 °~135 ° of 90 degree of regions, 135 °~225 ° of 180 degree region, 225 °~315 ° 270 degree of regions can obtain the corresponding direction of motion sequence vector of gesture in each touch area;By in each touch Preset hysteresis region is chosen in the critical zone in region, and adds hysteresis loop comparator, such as chooses hysteresis 45 ° of critical zone 45 ° ± 5 ° of region, the then threshold value of the hysteresis loop comparator added are 5 °;The hysteresis of touch area and addition based on above-mentioned division Comparator, then in each touch area, respectively according between all two neighboring moment in each touch area Direction of motion vector respectively obtains the center-of-mass coordinate and the direction of motion of the circle operation of each touch area inside-paint, to identify The corresponding type for drawing circle operation in each touch area and output;For example, when the finger for drawing circle operation is located at 0 degree of region, When being moved to 45 ° of+5 ° of hysteresis regions counterclockwise, final recognition result can export a gesture counterclockwise;When picture circle operation When finger is located at 90 degree of regions, when being moved to 45 ° of -5 ° of hysteresis regions clockwise, final recognition result can export a up time Needle gesture;It the step of by above-mentioned division touch area and addition hysteresis loop comparator, can be carried out on a touchpad to avoid finger Data dithering caused by circle operation is drawn, guarantees stability, to further increase recognition accuracy.
Further, characteristic vector sequence is the fortune in direction of motion sequence vector with the same direction in the step 5 The set of dynamic direction vector;
The specific steps for identifying the type of the gesture according to all characteristic vector sequences include:
Step 501: all direction of motion vector field homoemorphisms in each characteristic vector sequence adding up respectively, obtain every The mould of a characteristic vector sequence is long;
Step 502: the mould according to all characteristic vector sequences is long, and the total mould for obtaining the gesture is long;
Step 503: the ratio between total mould of the mould length and the gesture that calculate separately each characteristic vector sequence is long obtains The probability occurred in the gesture to each characteristic vector sequence, and all characteristic vector sequences are obtained according to all probability Probability sequence;
The probability that ith feature sequence vector occurs in the gesture are as follows:
Wherein, giMould for ith feature sequence vector is long, p (gi) it is ith feature sequence vector in the gesture The probability of appearance, M are characterized the sum of sequence vector,It is long for total mould of the gesture;
Step 504: traversal probability sequence, and according to the corresponding characteristic vector sequence of maximum value in probability sequence Direction determines the type of the gesture.
The beneficial effect of above-mentioned further scheme is: (the then gesture when judging that gesture-type operates for non-picture circle in step 3 Type is linear operating), it is only necessary to determine the direction of gesture is the concrete type that can determine gesture;And user grasps on a touchpad When making, due to individual difference, it can not be operated in strict accordance with predefined gesture, will appear with the direction of predefined gesture Difference, therefore for the gestures direction of user's input is recognized accurately, can extract has phase Tongfang in direction of motion sequence vector To direction of motion vector, all direction of motion vectors in the same direction on each direction constitute a characteristic vector sequences, lead to The total mould for crossing the mould length and gesture that calculate each characteristic vector sequence is long, then the mould by calculating each characteristic vector sequence The long ratio long with total mould of gesture can be obtained the probability of each characteristic vector sequence appearance, and when a feature vector The maximum probability that sequence occurs then illustrates that the corresponding direction of this feature sequence vector is the direction of the gesture of user's input, because , by traversal probability sequence, finding out the corresponding characteristic vector sequence of maximum probability value can determine gesture-type and exports for this;
The non-gesture for drawing circle action type of above-mentioned judgement, method is simple, and calculation amount is small, without referring to mould with preset gesture Plate is matched, and recognition efficiency is high.
Further, before the step 501 further include:
Step 500: each characteristic vector sequence being filtered according to the long threshold value of preset dynamic analog.
The beneficial effect of above-mentioned further scheme is: by the long threshold value of preset dynamic analog to each characteristic vector sequence It is filtered, the long shorter direction of motion vector of some moulds in characteristic vector sequence, the long shorter fortune of such mould can be removed Dynamic direction vector is usually distracter, therefore discrimination can be improved by filtering processing;Wherein, the long threshold value of preset dynamic analog It can be arranged according to the actual situation.
Another aspect according to the present invention provides a kind of dynamic hand gesture recognition device based on touch tablet, including data Obtain module, data processing module, judgment module, the first identification module, extraction module and the second identification module;
The data acquisition module, for obtaining the corresponding coordinate points sequence of the gesture that is being acted on touch tablet with finger Column;
The data processing module is obtained for the coordinate points pair according to each adjacent two moment in coordinate point sequence Direction of motion sequence vector;
The judgment module, for judging whether the type of the gesture is picture circle behaviour according to direction of motion sequence vector Make:
First identification module, when for judging that the type of the gesture operates for picture circle when the judgment module, root According to the direction of motion vector between all two neighboring moment in direction of motion sequence vector, picture circle operation is obtained in the touching Center-of-mass coordinate and the direction of motion in template, and identify according to center-of-mass coordinate and the direction of motion type and defeated of picture circle operation Out;
The extraction module, when for judging that the type of the gesture operates for non-picture circle when the judgment module, to institute It states finger and leaves the touch tablet, multiple characteristic vector sequences are extracted from direction of motion sequence vector;
Second identification module, for identifying type and the output of the gesture according to all characteristic vector sequences.
The beneficial effects of the present invention are: obtaining the coordinate points sequence of gesture on a touchpad by data acquisition module first Column, the coordinate point sequence include the corresponding coordinate points in touch point at each moment, therefore by data processing module according to every phase The coordinate points pair at adjacent two moment, available direction of motion sequence vector, further according to judgment module according to direction of motion vector Sequence judges whether the type of gesture currently entered is that picture circle operation can when judging that the type of gesture operates for picture circle To determine picture circle operation respectively according to the direction of motion vector between all two neighboring moment by the first identification module Mass center and the direction of motion can determine the type (draw circle clockwise or draw circle counterclockwise) of picture circle operation and exports;When sentencing The type for gesture of cutting off the hands is the circle operation of non-picture, therefore multiple features in direction of motion sequence vector can be extracted by extraction module Sequence vector, then identify that the direction of gesture may recognize that hand according to all characteristic vector sequences by the second identification module The type of gesture and output;
Dynamic hand gesture recognition device based on touch tablet of the invention is not necessarily to and the progress of pre-set reference template Match, enormously simplify calculating, enhance recognition efficiency, gesture identification can be exported in real time as a result, user experience is strong, can also collected At arriving in the small embedded systems such as single-chip microcontroller, especially suitable for the touch tablet of vehicle steering wheel, to realize specific vehicle-mounted function Energy.
Another aspect according to the present invention provides a kind of dynamic hand gesture recognition device based on touch tablet, including processing Device, memory and storage in the memory and may operate at computer program on the processor, the computer journey The step in a kind of dynamic gesture identification method based on touch tablet of the invention is realized when sort run.
The beneficial effects of the present invention are: the computer program by storage on a memory, and run on a processor, it is real The existing dynamic hand gesture recognition of the invention based on touch tablet is greatly simplified without being matched with pre-set reference template It calculates, enhances recognition efficiency, can gesture identification be exported in real time as a result, user experience is strong, also can be integrated into single-chip microcontroller In equal small embedded systems, especially suitable for the touch tablet of vehicle steering wheel, to realize specific vehicle-mounted function.
Another aspect according to the present invention provides a kind of dynamic hand gesture recognition system based on touch tablet, including touches Plate, and the dynamic hand gesture recognition device of the invention based on touch tablet being arranged in the touch tablet, the touch tablet with The dynamic hand gesture recognition device electrical connection based on touch tablet.
The beneficial effects of the present invention are: being constituted by touch tablet and the Dynamic Recognition device based on touch tablet of the invention Dynamic recognition system based on touch tablet enormously simplifies calculating, enhances without being matched with pre-set reference template Recognition efficiency can export gesture identification in real time as a result, user experience is strong, also can be integrated into the low profile edges such as single-chip microcontroller In system, especially suitable for the touch tablet of vehicle steering wheel, to realize specific vehicle-mounted function.
Detailed description of the invention
Fig. 1 is the flow diagram of the dynamic gesture identification method based on touch tablet in the embodiment of the present invention one;
Fig. 2 is to obtain the flow diagram of direction of motion sequence vector in the embodiment of the present invention one;
Fig. 3 is the model schematic divided in the embodiment of the present invention one to the angle of direction of motion vector;
Fig. 4 is to judge whether the type of gesture is the round flow diagram operated of picture in the embodiment of the present invention one;
Fig. 5 illustrates for the process of the center-of-mass coordinate and the direction of motion that obtain picture circle operation in the S4 of the embodiment of the present invention one Figure;
Fig. 6 justifies the type of operation and the flow diagram of output to obtain picture in the S4 of the embodiment of the present invention one;
Fig. 7 is the model schematic that touch area is divided in the S411 of the embodiment of the present invention one;
Fig. 8 is the model schematic that hysteresis region is chosen in the S412 of the embodiment of the present invention one;
Fig. 9 is the flow diagram one that the type of gesture is identified in the S5 of the embodiment of the present invention one;
Figure 10 is the flow diagram two that the type of gesture is identified in the S5 of the embodiment of the present invention one;
Figure 11 is the structural schematic diagram of the dynamic hand gesture recognition device based on touch tablet in the embodiment of the present invention two.
Specific embodiment
The principle and features of the present invention will be described below with reference to the accompanying drawings, and the given examples are served only to explain the present invention, and It is non-to be used to limit the scope of the invention.
With reference to the accompanying drawing, the present invention will be described.
Embodiment one, as shown in Figure 1, a kind of dynamic gesture identification method based on touch tablet, comprising the following steps:
S1: it obtains and is acting on the corresponding coordinate point sequence of the gesture on touch tablet with finger;
S2: according to the coordinate points pair at each adjacent two moment in coordinate point sequence, direction of motion sequence vector is obtained;
S3: whether the type that the gesture is judged according to direction of motion sequence vector is picture circle operation, if so, S4 is carried out, If it is not, carrying out S5;
S4: according to the direction of motion vector between all two neighboring moment in direction of motion sequence vector, picture is obtained Center-of-mass coordinate and the direction of motion of the circle operation on the touch tablet, and picture circle behaviour is identified according to center-of-mass coordinate and the direction of motion The type of work and output;
S5: leaving the touch tablet to the finger, multiple characteristic vector sequences extracted from direction of motion sequence vector, And type and the output of the gesture are identified according to all characteristic vector sequences.
Gesture classification is drawn circle behaviour by first carrying out just classification to gesture by the dynamic gesture identification method based on touch tablet Make and the circle operation of non-pictures, then different recognition methods be respectively adopted to identify the concrete type of gesture, be not necessarily to it is pre-set Reference template is matched, and is enormously simplified calculating, is enhanced recognition efficiency, can export gesture identification in real time as a result, user Experience sense is strong, also can be integrated into the small embedded systems such as single-chip microcontroller, especially suitable for the touch tablet of vehicle steering wheel, with reality Now specific vehicle-mounted function.
It should be noted that carrying out the prior art that output is comparative maturity to the gesture identified, specifically repeat no more.
Preferably, as shown in Fig. 2, the specific steps of S2 include:
S201: it according to the coordinate points of the coordinate points of the t moment in coordinate point sequence and t-1 moment, is calculated and coordinate The corresponding adjacent coordinates point distance of t moment in point sequence;
Adjacent coordinates point distance between the coordinate points and the coordinate points at t-1 moment of t moment are as follows:
Wherein, d is adjacent coordinates point distance corresponding with t moment, xtIt is right on the touch tablet for finger described in t moment The x coordinate answered, ytFor finger described in t moment on the touch tablet corresponding y-coordinate, xt-1It is finger described in the t-1 moment in institute State corresponding x coordinate, y on touch tablett-1For finger described in the t-1 moment on the touch tablet corresponding y-coordinate;
S202: it according to the method for S201, is calculated one-to-one multiple adjacent with the moment each in coordinate point sequence Coordinate points distance, and Screening Treatment is carried out to all adjacent coordinates point distances according to preset screening criterion, it obtains multiple effective Coordinate points;
The screening criterion are as follows:
Wherein, dthreFor preset distance threshold, label (d) is the output function of the screening criterion, as label (d) Output valve when being 0 apart from the coordinate points at a corresponding moment be invalid sit with adjacent coordinates point then in coordinate point sequence Punctuate, when the output valve of label (d) is 1, then in coordinate point sequence, with adjacent coordinates point apart from a corresponding moment Coordinate points be effective coordinate points;
S203: all effective coordinate points are in chronological sequence sequentially sorted, effective coordinate point sequence is obtained;
S204: according to effective coordinate points at -1 moment of effective coordinate points at the t ' moment in effective coordinate point sequence and t ', Obtain direction of motion vector corresponding with t ' current time;
S205: it according to the method for S204, obtains and moment each in effective coordinate point sequence multiple movements correspondingly Direction vector, and direction of motion sequence vector is obtained according to all direction of motion vectors.
During obtaining direction of motion sequence vector, pass through the coordinate points of t moment in coordinates computed point sequence first The distance between the coordinate points at t-1 moment, i.e. adjacent coordinates point distance d, so as to obtain and sit after the same method Each moment corresponding adjacent coordinates point distance in punctuation sequence;Again by preset distance threshold and preset screening criterion, Screening Treatment is carried out to all adjacent coordinates points distance in coordinate point sequence, the coordinate at some two neighboring moment can be abandoned The too small coordinate points of distance, i.e. invalid coordinates point between point, so as to improve recognition efficiency and accuracy rate;Again by screening It manages obtained all effective coordinate points to sort according to chronological order, obtains effective coordinate point sequence, be convenient for subsequent basis Effective coordinate point sequence obtain between all two neighboring moment (such as in effective coordinate point sequence the t ' moment effective coordinate Point and -1 moment of t ' effective coordinate points) direction of motion vector, convenient for obtaining movement side according to all direction of motion vectors To sequence vector, so that whether facilitate subsequent according to direction of motion sequence is that picture circle operation judges to the type of gesture;Its In, preset distance threshold can be arranged according to the actual situation.
Specifically, it is being obtained and moment each in effective coordinate point sequence one-to-one multiple movement sides in the present embodiment When to vector, each direction of motion vector field homoemorphism and angle are recorded, wherein the direction of motion vector at t ' moment is vt′, mould and Angle is respectively angel (vt′) and mod (vt′);It to simplify the calculation, will be obtained all using an angle threshold theta The angle of direction of motion vector is divided into 0 ° of direction, 45 ° of directions, 90 ° of directions, 135 ° of directions, 180 ° of directions, 225 ° of directions, 270 ° Totally eight directions indicate for direction and 315 ° of directions;Such as when theta=30 °, the angle of direction of motion vector is calculated For 30 °≤angel (vt′) 60 ° of < when, then by the angle recordings of the direction of motion vector be 45 ° of directions;What angle specifically divided Schematic diagram is as shown in Figure 3.
Preferably, as shown in figure 4, the specific steps of S3 include:
S301: according to the direction of motion at -1 moment of direction of motion vector sum t ' at the t ' moment in direction of motion sequence vector It is poor that movement direction angle corresponding with the t ' moment in direction of motion sequence vector is calculated in vector;
Movement direction angle between the direction of motion vector at -1 moment of direction of motion vector and t ' at t ' moment is poor are as follows:
θ=angel (vt′)-angel(vt′-1); (3)
Wherein, θ is that movement direction angle corresponding with the t ' moment is poor, vt' be the t ' moment direction of motion vector, vt′-1For The direction of motion vector at -1 moment of t ', angel (vt′) be the t ' moment direction of motion vector angle, angel (vt′-1) it is t '- The angle of the direction of motion vector at 1 moment;
S302: it according to the method for S301, respectively obtains continuous more before the t ' moment in direction of motion sequence vector A moment, corresponding multiple movement direction angles were poor;
S303: being judged according to the positive-negative polarity of all movement direction angle differences, when movement direction angle difference is continuously When the number just or being continuously negative is greater than or equal to preset times, then judge that the type of the gesture operates for picture circle, and carry out Otherwise S4 carries out S5.
Since movement direction angle difference of the picture circle operation between each moment and its last moment can continuously be It just or is continuously negative, therefore, when the movement direction angle difference between -1 moment of t ' moment in direction of motion sequence vector and t ' is Positive (movement direction angle difference i.e. corresponding with the t ' moment in direction of motion sequence vector is positive), and t ' moment corresponding movement side The number being positive to differential seat angle continuous multiple movement direction angles corresponding with continuous multiple moment before the t ' moment are poor The total degree that is continuously positive of polarity when being greater than or equal to preset times;Alternatively, in the direction of motion sequence vector t ' moment with Movement direction angle difference between -1 moment of t ' is negative (direction of motion angle i.e. corresponding with the t ' moment in direction of motion sequence vector Degree difference is negative), and when the number that is negative of t ' moment corresponding movement direction angle difference is with continuous multiple before the t ' moment When carving number that the polarity of corresponding continuous multiple movement direction angle differences is continuously negative more than or equal to preset times, then The type that may determine that the gesture is picture circle operation;Therefore direction of motion vector sum corresponding with the t ' moment and t ' -1 are obtained first Movement direction angle between moment corresponding direction of motion vector is poor, i.e., movement direction angle corresponding with the t ' moment is poor, then Corresponding multiple movement direction angles of continuous multiple moment before calculating the t ' moment according to formula (3) are poor, and finally judgement should Whether the number that all movement direction angle differences are continuously positive or are continuously negative is greater than or equal to preset times, that is, can determine whether hand Whether the type of gesture is picture circle operation;
Judge whether gesture-type is picture circle operation, and method is simple, and calculation amount is small, and facilitates subsequent through the above steps It is judged that result identifies gesture-type to take different recognition methods, recognition efficiency height is improved, recognition result is more acurrate.
Specifically, the preset times in the present embodiment are set as 5 times, and movement direction angle difference is continuously positive or is continuously negative Number when being greater than or equal to 5 times, i.e., and t ' moment corresponding movement direction angle is poor, corresponding with -1 moment of the t ' direction of motion Differential seat angle and -2 moment of t ' corresponding movement direction angle is poor, corresponding with -3 moment of t ' movement direction angle difference and with t ' -4 The moment polarity of corresponding movement direction angle difference is positive, that is, judges that the type of gesture operates for picture circle;Alternatively, with the t ' moment Corresponding movement direction angle is poor and -1 moment of t ' corresponding movement direction angle is poor, the direction of motion corresponding with -2 moment of t ' The polarity of differential seat angle and -3 moment of t ' corresponding movement direction angle difference and movement direction angle difference corresponding with -4 moment of t ' is equal It is negative, that is, judges that the type of gesture operates for picture circle.
Preferably, as shown in figure 5, in S4, the specific steps for obtaining center-of-mass coordinate and the direction of motion include:
S401: it is corresponded according to the direction of motion vector between the two neighboring moment all in direction of motion sequence vector Available point coordinate points, calculate picture circle operation center-of-mass coordinate;
Calculate the formula of center-of-mass coordinate are as follows:
Wherein, (xmc,ymc) it is center-of-mass coordinate, (xt′,yt′) it is that direction of motion vector corresponding with the t ' moment is corresponding effectively The coordinate of coordinate points, N are the sum of effective coordinate points;
S402: according to the coordinate of center-of-mass coordinate and the corresponding effective coordinate points of direction of motion vector corresponding with the t ' moment, The picture circular direction vector at t ' moment is obtained, it is corresponding effectively according to center-of-mass coordinate and direction of motion vector corresponding with -1 moment of t ' The coordinate of coordinate points obtains the picture circular direction vector at -1 moment of t ';
S403: according to preset direction of motion criterion, to -1 moment of angle and t ' of the picture circular direction vector at t ' moment The angle for drawing circular direction vector is judged, the direction of motion of picture circle operation is obtained;
The direction of motion criterion are as follows:
Wherein, ut′For the picture circular direction vector at t ' moment, angel (ut′) be the t ' moment picture circular direction vector angle, ut′-1For the picture circular direction vector at -1 moment of t ', angel (ut′-1) be -1 moment of t ' picture circular direction vector angle, Clockwise is the output function of direction of motion criterion, and when the output valve of Clockwise is 1, then the direction of motion is up time Needle, when the output valve of Clockwise is 0, then the direction of motion is as you were, when the output valve of Clockwise is -1, then The direction of motion is counterclockwise.
When it is picture circle operation that gesture-type has been determined in S3, the gesture corresponding all two neighboring moment can be passed through Between direction of motion vector effective coordinate points (effective coordinate that effective coordinate points at i.e. all moment are constituted correspondingly Point sequence) calculate center-of-mass coordinate, and effective seat according to corresponding to center-of-mass coordinate and direction of motion vector corresponding with the t ' moment The coordinate of punctuate obtains the picture circular direction vector at the t ' moment, and using same method obtain the picture circular direction at -1 moment of t ' to Amount determines the concrete type of picture circle operation further according to preset direction of motion criterion, is to draw round or draw counterclockwise clockwise Circle, and recognition result is exported, recognition efficiency is high, accuracy rate is high.
Specifically, the fortune of circle operation is then drawn when the output valve of direction of motion criterion Clockwise is 1 in the present embodiment Dynamic direction is clockwise, picture circle as clockwise;When the output valve of direction of motion criterion Clockwise is -1, then round behaviour is drawn The direction of motion of work is counterclockwise, picture circle as counterclockwise;And when the output valve of direction of motion criterion Clockwise is 0, then The direction of motion for drawing circle operation is (i.e. current time does not move temporarily) as you were.
It should be noted that identifying that picture circle operation is the prior art according to center-of-mass coordinate and the direction of motion, specifically no longer It repeats.
It preferably, further include following when judging that the type of the gesture operates for picture circle as shown in fig. 6, before S4 Step:
S411: according to predetermined angle, the touch tablet is divided into K touch area;And according to the method for S2, obtain respectively Take the direction of motion sequence vector in each touch area;Wherein, K >=2;
S412: preset hysteresis region is chosen in the critical zone of each touch area, and respectively in each hysteresis region Add hysteresis loop comparator;
Then S4 specifically:
Based on the hysteresis loop comparator of each touch area addition, according in direction of motion sequence vector in each touch area All two neighboring moment between direction of motion vector, respectively obtain the circle operation of each touch area inside-paint in the touch Center-of-mass coordinate and the direction of motion on plate, and respectively according in each touch area center-of-mass coordinate and the direction of motion identify pair The round type operated of the picture answered and output.
It the step of by above-mentioned division touch area and addition hysteresis loop comparator, can be carried out on a touchpad to avoid finger Data dithering caused by circle operation is drawn, guarantees stability, to further increase recognition accuracy.
Specifically, it when judging that gesture-type operates for picture circle in S3, will be touched in the present embodiment according to 90 ° of predetermined angle Template is divided into 4 touch areas, respectively includes 315 °~0 ° and 0 °~45 ° of 0 degree of region, 45 °~135 ° of 90 degree of areas Domain, 135 °~225 ° of 180 degree region, 225 °~315 ° of 270 degree of regions, as shown in Figure 7;Therefore, in each touch area It is interior to obtain the corresponding direction of motion sequence vector of gesture.
Specifically, in the present embodiment, preset hysteresis region is chosen by the critical zone in above-mentioned each touch area, And hysteresis loop comparator is added, as shown in figure 8,45 ° ± 5 ° of hysteresis region, that is, the hysteresis ratio added are chosen in the critical zone at 45 ° Threshold value compared with device is 5 °, chooses 135 ° ± 5 ° of hysteresis region 135 ° of critical zones, chooses hysteresis 225 ° of critical zones 315 ° ± 5 ° of hysteresis region is chosen 315 ° of critical zones in 225 ° ± 5 ° of region;Touch area based on above-mentioned division and The hysteresis loop comparator of addition, then in each touch area, according to method described in S401 to S403, respectively according to each described The direction of motion vector of all adjacent moments in touch area respectively obtains the mass center of each touch area inside-paint circle operation Coordinate and the direction of motion, to identify the corresponding type for drawing circle operation in each touch area and output;For example, when picture circle When the finger of operation is located at 0 degree of region, when being moved to 45 ° of+5 ° of hysteresis regions counterclockwise, final recognition result can export one Gesture counterclockwise;When the finger for drawing circle operation is located at 90 degree of regions, when being moved to 45 ° of -5 ° of hysteresis regions clockwise, finally Recognition result can export a gesture clockwise.
Preferably, as shown in figure 9, in S5, characteristic vector sequence is to have the same direction in direction of motion sequence vector Direction of motion vector set;
The specific steps for identifying the type of the gesture according to all characteristic vector sequences include:
S501: all direction of motion vector field homoemorphisms in each characteristic vector sequence are added up respectively, are obtained each The mould of characteristic vector sequence is long;
S502: the mould according to all characteristic vector sequences is long, and the total mould for obtaining the gesture is long;
S503: the ratio between total mould of the mould length and the gesture that calculate separately each characteristic vector sequence is long obtains The probability that each characteristic vector sequence occurs in the gesture, and the general of all characteristic vector sequences is obtained according to all probability Rate sequence;
The probability that ith feature sequence vector occurs in the gesture are as follows:
Wherein, giMould for ith feature sequence vector is long, p (gi) it is ith feature sequence vector in the gesture The probability of appearance, M are characterized the sum of sequence vector,It is long for total mould of the gesture;
S504: traversal probability sequence, and according to the side of the corresponding characteristic vector sequence of maximum value in probability sequence To determining the type of the gesture.
When judging that gesture-type operates for non-picture circle in S3 (then gesture-type is linear operating), it is only necessary to determine gesture Direction be that can determine the concrete type of gesture;And user is when operating on a touchpad, it, can not be in strict accordance with due to individual difference Predefined gesture is operated, and will appear difference with the direction of predefined gesture, therefore for user's input is recognized accurately Gestures direction, the direction of motion vector in direction of motion sequence vector with the same direction can be extracted, on each direction All direction of motion vectors in the same direction constitute a characteristic vector sequences, it is long by the mould for calculating each characteristic vector sequence It is long with total mould of gesture, then the long ratio long with total mould of gesture of the mould by calculating each characteristic vector sequence, it can obtain The probability that each characteristic vector sequence occurs is obtained, and when the maximum probability that a characteristic vector sequence occurs, then illustrate the spy The corresponding direction of sign sequence vector is the direction of the gesture of user's input, therefore by traversal probability sequence, finds out probability most Being worth corresponding characteristic vector sequence greatly i.e. can determine gesture-type and exports;
The non-gesture for drawing circle action type of above-mentioned judgement, method is simple, and calculation amount is small, without referring to mould with preset gesture Plate is matched, and recognition efficiency is high.
Preferably, as shown in Figure 10, before S501 further include:
S500: each characteristic vector sequence is filtered according to the long threshold value of preset dynamic analog.
Each characteristic vector sequence is filtered by the long threshold value of preset dynamic analog, can remove feature to The long shorter direction of motion vector of some moulds in sequence is measured, the long shorter direction of motion vector of such mould is usually distracter, therefore Discrimination can be improved by filtering processing.
Specifically, in the present embodiment, the long threshold value of dynamic analog is used as by the 5% of total mould length of gesture.
Embodiment two, a kind of as shown in figure 11, dynamic hand gesture recognition device based on touch tablet, including data acquisition mould Block, data processing module, judgment module, the first identification module, extraction module and the second identification module;
The data acquisition module, for obtaining the corresponding coordinate points sequence of the gesture that is being acted on touch tablet with finger Column;
The data processing module is obtained for the coordinate points pair according to each adjacent two moment in coordinate point sequence Direction of motion sequence vector;
The judgment module, for judging whether the type of the gesture is picture circle behaviour according to direction of motion sequence vector Make:
First identification module, when for judging that the type of the gesture operates for picture circle when the judgment module, root According to the direction of motion vector between all two neighboring moment in direction of motion sequence vector, picture circle operation is obtained in the touching Center-of-mass coordinate and the direction of motion in template, and identify according to center-of-mass coordinate and the direction of motion type and defeated of picture circle operation Out;
The extraction module, when for judging that the type of the gesture operates for non-picture circle when the judgment module, to institute It states finger and leaves the touch tablet, multiple characteristic vector sequences are extracted from direction of motion sequence vector;
Second identification module, for identifying type and the output of the gesture according to all characteristic vector sequences.
The coordinate point sequence of gesture on a touchpad is obtained by data acquisition module first, which includes every The corresponding coordinate points in the touch point at a moment, therefore by data processing module according to the coordinate points pair at each adjacent two moment, Available direction of motion sequence vector is judged currently entered further according to judgment module according to direction of motion sequence vector Whether the type of gesture is that picture circle operation can pass through the first identification module root when judging that the type of gesture operates for picture circle The mass center and the direction of motion for determining picture circle operation respectively according to the direction of motion vector between all two neighboring moment, can be true The type (i.e. picture is round clockwise or draws circle counterclockwise) of fixed picture circle operation simultaneously exports;It is grasped when judging that the type of gesture is justified for non-picture Make, therefore multiple characteristic vector sequences in direction of motion sequence vector can be extracted by extraction module, then know by second Other module identifies that the direction of gesture may recognize that type and the output of gesture according to all characteristic vector sequences;
The dynamic hand gesture recognition device based on touch tablet of the present embodiment is not necessarily to and the progress of pre-set reference template Match, enormously simplify calculating, enhance recognition efficiency, gesture identification can be exported in real time as a result, user experience is strong, can also collected At arriving in the small embedded systems such as single-chip microcontroller, especially suitable for the touch tablet of vehicle steering wheel, to realize specific vehicle-mounted function Energy.
Embodiment three is based on embodiment one and embodiment two, and the present embodiment also provides another dynamic based on touch tablet Gesture identifying device including processor, memory and stores in the memory and may operate at the meter on the processor Calculation machine program, the computer program realize the specific steps of S1 to S5 as shown in Figure 1 when running.
It by storing computer program on a memory, and runs on a processor, realizes of the invention based on touch The dynamic hand gesture recognition of plate enormously simplifies calculating, enhances identification effect without being matched with pre-set reference template Rate can export gesture identification in real time as a result, user experience is strong, also can be integrated into the small embedded systems such as single-chip microcontroller, Especially suitable for the touch tablet of vehicle steering wheel, to realize specific vehicle-mounted function.
In the present embodiment S1 to S5 do not use up details, the specific descriptions of detailed in Example one and Fig. 1 to Figure 10, herein no longer It repeats.
Example IV is based on embodiment one and embodiment two, and the present embodiment also provides a kind of dynamic hand based on touch tablet Gesture identifying system, including touch tablet, and be arranged in the touch tablet as embodiment three the dynamic hand based on touch tablet Gesture identification device, the touch tablet are electrically connected with the dynamic hand gesture recognition device based on touch tablet.
The dynamic of the invention based on touch tablet is constituted by touch tablet and the Dynamic Recognition device based on touch tablet to know Other system enormously simplifies calculating, enhances recognition efficiency, Ke Yishi without being matched with pre-set reference template When output gesture identification as a result, user experience is strong, also can be integrated into the small embedded systems such as single-chip microcontroller, especially suitable for The touch tablet of vehicle steering wheel, to realize specific vehicle-mounted function.
Specifically, in the present embodiment, touch tablet is provided in vehicle steering wheel, and be embedded in MCU in touch tablet, it should Comprising the dynamic hand gesture recognition device based on touch tablet described in embodiment three in MCU, seat when touching is acquired by touch tablet Punctuation sequence, and be real-time transmitted in MCU, blind operation multi-functional to steering wheel is completed by gesture.Wherein it is possible to by not With gesture may be implemented multimedia function, such as the big minor adjustment of volume of music is realized by picture circle operation, by operate upwards/ Downward operation (or operation to the left/operate to the right) realizes musically one first/next switching etc.;It can also realize call function Can, when there is call to access, by operate upwards/downward operation (or operate/to the right operate) realize receiving talking/to the left and hang up, The big minor adjustment of In Call is realized by drawing circle operation.
Dynamic hand gesture recognition device in the present embodiment based on touch tablet does not use up details, and detailed in Example three is specifically retouched It states, details are not described herein again.
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.

Claims (10)

1. a kind of dynamic gesture identification method based on touch tablet, which comprises the following steps:
Step 1: obtaining and acting on the corresponding coordinate point sequence of the gesture on touch tablet with finger;
Step 2: according to the coordinate points pair at each adjacent two moment in coordinate point sequence, obtaining direction of motion sequence vector;
Step 3: whether the type that the gesture is judged according to direction of motion sequence vector is picture circle operation, if so, carrying out step 4, if it is not, carrying out step 5;
Step 4: according to the direction of motion vector between all two neighboring moment in direction of motion sequence vector, obtaining picture circle Center-of-mass coordinate and the direction of motion on the touch tablet are operated, and picture circle operation is identified according to center-of-mass coordinate and the direction of motion Type and output;
Step 5: the touch tablet is left to the finger, multiple characteristic vector sequences are extracted from direction of motion sequence vector, And type and the output of the gesture are identified according to all characteristic vector sequences.
2. the dynamic gesture identification method according to claim 1 based on touch tablet, which is characterized in that the step 2 Specific steps include:
Step 201: according to the coordinate points of the coordinate points of the t moment in coordinate point sequence and t-1 moment, being calculated and coordinate points The corresponding adjacent coordinates point distance of t moment in sequence;
Adjacent coordinates point distance between the coordinate points and the coordinate points at t-1 moment of t moment are as follows:
Wherein, d is adjacent coordinates point distance corresponding with t moment, xtFor finger described in t moment on the touch tablet corresponding x Coordinate, ytFor finger described in t moment on the touch tablet corresponding y-coordinate, xt-1It is finger described in the t-1 moment in the touching Corresponding x coordinate, y in templatet-1For finger described in the t-1 moment on the touch tablet corresponding y-coordinate;
Step 202: according to the method for the step 201, being calculated more correspondingly with the moment each in coordinate point sequence A adjacent coordinates point distance, and Screening Treatment is carried out to all adjacent coordinates point distances according to preset screening criterion, it obtains more A effective coordinate points;
The screening criterion are as follows:
Wherein, dthreFor preset distance threshold, label (d) is the output function of the screening criterion, defeated as label (d) It then apart from the coordinate points at a corresponding moment is invalid coordinates with adjacent coordinates point in coordinate point sequence when value is 0 out Point, when the output valve of label (d) is 1, then in coordinate point sequence, with adjacent coordinates point apart from a corresponding moment Coordinate points are effective coordinate points;
Step 203: all effective coordinate points in chronological sequence sequentially being sorted, effective coordinate point sequence is obtained;
Step 204: according to effective coordinate points at -1 moment of effective coordinate points at the t ' moment in effective coordinate point sequence and t ', obtaining To direction of motion vector corresponding with the t ' moment in effective coordinate point sequence;
Step 205: according to the method for the step 204, obtaining more correspondingly with the moment each in effective coordinate point sequence A direction of motion vector, and direction of motion sequence vector is obtained according to all direction of motion vectors.
3. the dynamic gesture identification method according to claim 2 based on touch tablet, which is characterized in that the step 3 Specific steps include:
Step 301: according to the direction of motion at -1 moment of direction of motion vector sum t ' at the t ' moment in direction of motion sequence vector It is poor that movement direction angle corresponding with the t ' moment in direction of motion sequence vector is calculated in vector;
Movement direction angle between the direction of motion vector at -1 moment of direction of motion vector and t ' at t ' moment is poor are as follows:
θ=angel (vt′)-angel(vt′-1); (3)
Wherein, θ is that movement direction angle corresponding with the t ' moment is poor, vt′For the direction of motion vector at t ' moment, vt′-1For t ' -1 The direction of motion vector at moment, angel (vt′) be the t ' moment direction of motion vector angle, angel (vt′-1) be -1 t ' when The angle of the direction of motion vector at quarter;
Step 302: according to the method for the step 301, respectively obtaining before the t ' moment in direction of motion sequence vector Continuously corresponding multiple movement direction angles of multiple moment are poor;
Step 303: being judged according to the positive-negative polarity of all movement direction angle differences, when movement direction angle difference is continuously positive Or the number being continuously negative then judges that the type of the gesture is picture circle operation, and is walked when being greater than or equal to preset times Rapid 4, otherwise, carry out step 5.
4. the dynamic gesture identification method according to claim 3 based on touch tablet, which is characterized in that in the step 4 In, the specific steps for obtaining center-of-mass coordinate and the direction of motion include:
Step 401: being corresponded according to the direction of motion vector between the two neighboring moment all in direction of motion sequence vector Effective coordinate points, calculate picture circle operation center-of-mass coordinate;
Calculate the formula of center-of-mass coordinate are as follows:
Wherein, (xmc,ymc) it is center-of-mass coordinate, (xt′,yt′) it is the corresponding effective coordinate of direction of motion vector corresponding with the t ' moment The coordinate of point, N are the sum of effective coordinate points;
Step 402: according to the coordinate of center-of-mass coordinate and the corresponding effective coordinate points of direction of motion vector corresponding with the t ' moment, obtaining To the picture circular direction vector at t ' moment, according to center-of-mass coordinate and the corresponding effective seat of direction of motion vector corresponding with -1 moment of t ' The coordinate of punctuate obtains the picture circular direction vector at -1 moment of t ';
Step 403: according to preset direction of motion criterion, to the picture at -1 moment of angle and t ' of the picture circular direction vector at t ' moment The angle of circular direction vector is judged, the direction of motion of picture circle operation is obtained;
The direction of motion criterion are as follows:
Wherein, ut′For the picture circular direction vector at t ' moment, angel (ut′) be the t ' moment picture circular direction vector angle, ut′-1 For the picture circular direction vector at -1 moment of t ', angel (ut′-1) be -1 moment of t ' picture circular direction vector angle, Clockwise For the output function of the direction of motion criterion, when the output valve of Clockwise is 1, then the direction of motion is clockwise, when When the output valve of Clockwise is 0, then the direction of motion is as you were, when the output valve of Clockwise is -1, is then moved Direction is counterclockwise.
5. the dynamic gesture identification method according to any one of claims 1 to 4 based on touch tablet, which is characterized in that It is further comprising the steps of when judging that the type of the gesture operates for picture circle before the step 4:
Step 411: according to predetermined angle, the touch tablet being divided into K touch area;And according to the method for the step 2, The direction of motion sequence vector in each touch area is obtained respectively;Wherein, K >=2;
Step 412: choosing preset hysteresis region in the critical zone of each touch area, and add respectively in each hysteresis region Add hysteresis loop comparator;
The then step 4 specifically:
Based on the hysteresis loop comparator of each touch area addition, according to the institute in direction of motion sequence vector in each touch area There is the direction of motion vector between the two neighboring moment, respectively obtains each touch area inside-paint circle operation on the touch tablet Center-of-mass coordinate and the direction of motion, and respectively according in each touch area center-of-mass coordinate and the direction of motion identify it is corresponding Draw the type of circle operation and output.
6. the dynamic gesture identification method according to claim 1 based on touch tablet, which is characterized in that in the step 5 In, characteristic vector sequence is the set of the direction of motion vector in direction of motion sequence vector with the same direction;
The specific steps for identifying the type of the gesture according to all characteristic vector sequences include:
Step 501: all direction of motion vector field homoemorphisms in each characteristic vector sequence adding up respectively, obtain each spy The mould for levying sequence vector is long;
Step 502: the mould according to all characteristic vector sequences is long, and the total mould for obtaining the gesture is long;
Step 503: the ratio between total mould of the mould length and the gesture that calculate separately each characteristic vector sequence is long obtains every The probability that a characteristic vector sequence occurs in the gesture, and the probability of all characteristic vector sequences is obtained according to all probability Sequence;
The probability that ith feature sequence vector occurs in the gesture are as follows:
Wherein, giMould for ith feature sequence vector is long, p (gi) it is that ith feature sequence vector occurs in the gesture Probability, M is characterized the sum of sequence vector,It is long for total mould of the gesture;
Step 504: traversal probability sequence, and according to the side of the corresponding characteristic vector sequence of maximum value in probability sequence To determining the type of the gesture.
7. the dynamic gesture identification method according to claim 6 based on touch tablet, which is characterized in that in the step Before 501 further include:
Step 500: each characteristic vector sequence being filtered according to the long threshold value of preset dynamic analog.
8. a kind of dynamic hand gesture recognition device based on touch tablet, which is characterized in that including data acquisition module, data processing mould Block, judgment module, the first identification module, extraction module and the second identification module;
The data acquisition module, for obtaining the corresponding coordinate point sequence of the gesture that is being acted on touch tablet with finger;
The data processing module is moved for the coordinate points pair according to each adjacent two moment in coordinate point sequence Direction vector sequence;
The judgment module, for judging whether the type of the gesture is picture circle operation according to direction of motion sequence vector:
First identification module, when for judging that the type of the gesture operates for picture circle when the judgment module, according to fortune The direction of motion vector at all two neighboring moment in dynamic direction vector sequence obtains picture circle operation on the touch tablet Center-of-mass coordinate and the direction of motion, and the type and output that picture circle operates are identified according to center-of-mass coordinate and the direction of motion;
The extraction module, when for judging that the type of the gesture operates for non-picture circle when the judgment module, to the hand Finger leaves the touch tablet, and multiple characteristic vector sequences are extracted from direction of motion sequence vector;
Second identification module, for identifying type and the output of the gesture according to all characteristic vector sequences.
9. a kind of dynamic hand gesture recognition device based on touch tablet, which is characterized in that including processor, memory and be stored in institute The computer program on the processor is stated in memory and be may operate at, is realized when the computer program is run as right is wanted Seek method and step described in any one of 1 to 7 claim.
10. a kind of dynamic hand gesture recognition system based on touch tablet, which is characterized in that including touch tablet, and be arranged described The dynamic hand gesture recognition device as claimed in claim 9 based on touch tablet in touch tablet, the touch tablet are based on described The dynamic hand gesture recognition device of touch tablet is electrically connected.
CN201910579656.6A 2019-06-28 2019-06-28 Dynamic gesture recognition method, device and system based on touch pad Active CN110362264B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910579656.6A CN110362264B (en) 2019-06-28 2019-06-28 Dynamic gesture recognition method, device and system based on touch pad

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910579656.6A CN110362264B (en) 2019-06-28 2019-06-28 Dynamic gesture recognition method, device and system based on touch pad

Publications (2)

Publication Number Publication Date
CN110362264A true CN110362264A (en) 2019-10-22
CN110362264B CN110362264B (en) 2022-06-10

Family

ID=68216146

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910579656.6A Active CN110362264B (en) 2019-06-28 2019-06-28 Dynamic gesture recognition method, device and system based on touch pad

Country Status (1)

Country Link
CN (1) CN110362264B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110908581A (en) * 2019-11-20 2020-03-24 网易(杭州)网络有限公司 Gesture recognition method and device, computer storage medium and electronic equipment
CN110956124A (en) * 2019-11-27 2020-04-03 云南电网有限责任公司电力科学研究院 Display device control method based on gestures and display device
CN112114666A (en) * 2020-08-25 2020-12-22 武汉海微科技有限公司 Dynamic gesture recognition algorithm based on touch panel

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104850264A (en) * 2015-05-29 2015-08-19 北京拉酷网络科技有限公司 Detecting and control method and system of diversified gesture operations on touch panel
CN105354532A (en) * 2015-09-25 2016-02-24 天津大学 Hand motion frame data based gesture identification method
CN106598231A (en) * 2016-11-22 2017-04-26 深圳市元征科技股份有限公司 Gesture identification method and apparatus
CN106648400A (en) * 2015-11-03 2017-05-10 华为终端(东莞)有限公司 Touch-data reporting method and electronic device
US20170286131A1 (en) * 2010-01-26 2017-10-05 Apple Inc. Gesture Recognizers with Delegates for Controlling and Modifying Gesture Recognition

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170286131A1 (en) * 2010-01-26 2017-10-05 Apple Inc. Gesture Recognizers with Delegates for Controlling and Modifying Gesture Recognition
CN104850264A (en) * 2015-05-29 2015-08-19 北京拉酷网络科技有限公司 Detecting and control method and system of diversified gesture operations on touch panel
CN105354532A (en) * 2015-09-25 2016-02-24 天津大学 Hand motion frame data based gesture identification method
CN106648400A (en) * 2015-11-03 2017-05-10 华为终端(东莞)有限公司 Touch-data reporting method and electronic device
CN106598231A (en) * 2016-11-22 2017-04-26 深圳市元征科技股份有限公司 Gesture identification method and apparatus

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110908581A (en) * 2019-11-20 2020-03-24 网易(杭州)网络有限公司 Gesture recognition method and device, computer storage medium and electronic equipment
CN110908581B (en) * 2019-11-20 2021-04-23 网易(杭州)网络有限公司 Gesture recognition method and device, computer storage medium and electronic equipment
CN110956124A (en) * 2019-11-27 2020-04-03 云南电网有限责任公司电力科学研究院 Display device control method based on gestures and display device
CN112114666A (en) * 2020-08-25 2020-12-22 武汉海微科技有限公司 Dynamic gesture recognition algorithm based on touch panel

Also Published As

Publication number Publication date
CN110362264B (en) 2022-06-10

Similar Documents

Publication Publication Date Title
CN110362264A (en) A kind of dynamic gesture identification method based on touch tablet, device and system
CN104731307B (en) A kind of body-sensing action identification method and human-computer interaction device
CN110531875A (en) Write generation method, terminal device and the storage medium of the vigour of style in writing
CN104239761A (en) Continuous identity authentication method based on touch screen slip behavior characteristics
CN103353935A (en) 3D dynamic gesture identification method for intelligent home system
US20130234957A1 (en) Information processing apparatus and information processing method
CN103440049B (en) A kind of input unit and input method
CN105956560A (en) Vehicle model identification method based on pooling multi-scale depth convolution characteristics
CN104966016B (en) Mobile terminal child user cooperatively judges and the method for limitation operating right
CN107357414B (en) Click action recognition method and device
CN109815876A (en) Gesture identification method based on address events stream feature
CN108596944A (en) A kind of method, apparatus and terminal device of extraction moving target
CN108814618A (en) A kind of recognition methods of motion state, device and terminal device
Hasan et al. Features fitting using multivariate gaussian distribution for hand gesture recognition
CN104517100A (en) Gesture pre-judging method and system
CN109117742A (en) Gestures detection model treatment method, apparatus, equipment and storage medium
CN103336967A (en) Hand motion trail detection method and apparatus
Joo et al. Real-time depth-based hand detection and tracking
Panwar Hand gesture based interface for aiding visually impaired
CN107220634B (en) Based on the gesture identification method for improving D-P algorithm and multi-template matching
CN106197470A (en) A kind of data processing method and electronic equipment
CN105956604A (en) Action identification method based on two layers of space-time neighborhood characteristics
CN110929766B (en) Self-adaptive pedestrian mobile phone attitude identification method based on Gaussian mixture clustering algorithm
CN114693943A (en) Non-maximum suppression acceleration method, system and equipment for target detection
CN110926467B (en) Self-adaptive pedestrian mobile phone attitude identification method based on mean value clustering algorithm

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
GR01 Patent grant
GR01 Patent grant
CP03 Change of name, title or address
CP03 Change of name, title or address

Address after: Room 572, 18th Floor, Building B4, Phase I, Longshan Innovation Park, Wuhan Future Science and Technology City, No. 999 Gaoxin Avenue, Donghu New Technology Development Zone, Wuhan City, Hubei Province, 430000

Patentee after: Wuhan Haiwei Technology Co.,Ltd.

Country or region after: China

Address before: Room 1588-2, 15th floor, building B3, zone 2, National Geospatial Information Industry base, No.3, wudayuan 4th Road, Donghu New Technology Development Zone, Wuhan City, Hubei Province, 430040

Patentee before: WUHAN HAIWEI TECHNOLOGY Co.,Ltd.

Country or region before: China