CN110414473B - Data glove gesture recognition algorithm based on mathematical statistics - Google Patents

Data glove gesture recognition algorithm based on mathematical statistics Download PDF

Info

Publication number
CN110414473B
CN110414473B CN201910721419.9A CN201910721419A CN110414473B CN 110414473 B CN110414473 B CN 110414473B CN 201910721419 A CN201910721419 A CN 201910721419A CN 110414473 B CN110414473 B CN 110414473B
Authority
CN
China
Prior art keywords
gamma
screening
standard deviation
gesture
data glove
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910721419.9A
Other languages
Chinese (zh)
Other versions
CN110414473A (en
Inventor
张效娟
毛亚平
程思
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.)
Qinghai Normal University
Original Assignee
Qinghai Normal University
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 Qinghai Normal University filed Critical Qinghai Normal University
Priority to CN201910721419.9A priority Critical patent/CN110414473B/en
Publication of CN110414473A publication Critical patent/CN110414473A/en
Application granted granted Critical
Publication of CN110414473B publication Critical patent/CN110414473B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/20—Movements or behaviour, e.g. gesture recognition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/20—Movements or behaviour, e.g. gesture recognition
    • G06V40/28—Recognition of hand or arm movements, e.g. recognition of deaf sign language

Abstract

The invention discloses a data glove gesture recognition algorithm based on mathematical statistics, and relates to the technical field of artificial intelligence: the method comprises the steps that the worn data glove uses a sensor. The numerical values in the text documents are read and saved in the groups a1 and b1, a2 and b2,.., a20 and b20, ai ═ in (ai,1, ai,2, ai,3,. ae, ai,24), bi ═ in (bi,1, bi,2, bi,3,. bi, 24). The method judges the deviation degree and linear relation between the randomly acquired gesture and the gesture in the template library by calculating the standard deviation and the correlation coefficient, and finds out the gesture which is most similar to the randomly acquired gesture in the template library. Through twice screening, the precision of the screening value is gradually improved, so that the gesture recognition is very accurate. And the gesture recognition based on the data glove can eliminate the interference of the external environment.

Description

Data glove gesture recognition algorithm based on mathematical statistics
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a data glove gesture recognition algorithm based on mathematical statistics.
Background
Gesture recognition techniques are often used in sign language recognition and human-computer interaction. The process of gesture recognition is a process of classifying the trajectory points in the model parameter space of the hand into a certain subset of the space, static gestures correspond to a point in the model parameter space, and dynamic gestures correspond to a trajectory in the model parameter space, so that the recognition methods are different. The existing gesture recognition technology mainly comprises a template matching method, a neural network method, a hidden Markov model method and the like. Wherein the neural network is a vision-based gesture recognition method. Hidden markov models are suitable for dynamic gesture recognition, while artificial neural networks are more suitable for static gesture recognition. However, in these technologies, due to the inherent characteristics of gestures such as timeliness and spatiality, and the fact that human hands are complex deformable bodies, and the visual method itself has instability, the difficulty of gesture recognition based on vision is increased, so that the recognition rate of gestures is not high, the effect and performance of gesture recognition are affected to a great extent, the interactivity is not good, and the purpose of gesture efficient recognition is not achieved.
Specifically, a commonly used gesture recognition method in the prior art is as follows:
Figure GDA0003482928340000011
by using a gesture mathematical model, the bending states of ten finger joints are represented by ten tuples, (K1, K2, …, K10), and the gesture is described by taking the value of an element item as-1, 1, or 0. When the template is created, the concept of template training is adopted, gesture samples of a user are collected according to standard gestures, and the gestures are automatically calibrated. The user is prompted to hold or open the hand, the values of the data glove sensors are compared, the maximum value and the minimum value of each sensor until the current time are recorded, the response values are corrected by the following formula, different gestures are defined, and therefore the template set is suitable for a large number of users.
Range；＝Maxq-Min
Rect；＝(Angle-Min)*4095/Range；
Wherein Range represents the dynamic variation Range of the No. 1 sensor response, Rect; the correction value of the interval mapping from the 4ngle response value of the number one sensor to 0,4095 is pointed out.
But in this method: because the finger joints of different users have different bending and stretching degrees, the number of suitable groups of the template set is still limited during matching, relatively perfect matching cannot be realized, the accurate degree cannot be achieved, and errors still exist; and the independence is poor, the flexibility is not high, and the gesture recognition rate is not high.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a data glove gesture recognition algorithm based on mathematical statistics, which solves the problem that relatively perfect matching cannot be realized due to different degrees of flexion and extension of finger joints of different users.
In order to achieve the purpose, the invention is realized by the following technical scheme: a data glove gesture recognition algorithm based on mathematical statistics comprises the following steps:
the first step is as follows: getdata () method is used for the worn data glove, the same gesture is divided into a natural state and an extended state to collect the angle values of the sensors respectively, and the values are stored in two text documents respectively. Reading the numerical value in the text document and storing the numerical value in the group a1And b1，a2And b2，…，a20And b20In (a)i＝(ai,1,ai,2,ai,3,…,ai,24),bi＝(bi,1,bi,2,bi,3,…,bi,24)。
The second step is that: the operator wears the data glove, and obtains the bending angle value of each sensor in real time through the sensor.
The third step: calculating standard deviation, and proving that the smaller the standard deviation is, the smaller the deviation degree between the data and the median thereof is, and explaining that the data is equal to (c)1,c,...,c24) Cross over ai＝(ai,1,ai,2,ai,3,...,ai,24) And bi＝(bi,1,bi,2,bi,3,...,bi,24) Gesture made and aiAnd biThe more similar the gestures are represented.
The fourth step: after the standard deviation calculation is completed, the standard deviation is screened, i.e. the smallest standard deviation is screened.
The fifth step: and performing a second round of screening, wherein correlation coefficients are calculated before screening.
And a sixth step: after the calculation of the correlation coefficient (γ) is completed, from γ to (γ)1,…,γnm) And screening out the maximum correlation coefficient.
Preferably, in the third step, the calculation method for calculating the standard deviation is as follows:
(1) for any i (1< ═ i < ═ 20), the median was calculated:
Figure GDA0003482928340000031
(2) calculating the standard deviation:
Figure GDA0003482928340000032
preferably, the screening of the minimum standard deviation in the fourth step is as follows:
(1) if there is a certain uiSo that (u)iAnd r) is 0, then u is retainedi；
If (u)iR) ≠ 0 for all uiConsider whether u is presentiSo that (u)i,r)＝1；
(2) If there is a certain uiSo that (u)iAnd r) 1 then holds ui；
If (u)iR) ≠ 1 for all uiConsider whether u is presentiSo that (u)iR) is 2; and so on.
Preferably, the fifth step correlation coefficient is calculated as follows:
(1) calculating the standard deviation:
Figure GDA0003482928340000041
Figure GDA0003482928340000042
…
Figure GDA0003482928340000043
and
Figure GDA0003482928340000044
Figure GDA0003482928340000045
…
Figure GDA0003482928340000046
(4) and (3) calculating covariance:
Figure GDA0003482928340000051
Figure GDA0003482928340000052
…
Figure GDA0003482928340000053
(5) calculating a correlation coefficient:
Figure GDA0003482928340000054
Figure GDA0003482928340000055
…
Figure GDA0003482928340000056
preferably, the screening process in the sixth step is as follows, and the second round of screening is based on the unit of 0.01, is relatively accurate and has strong feasibility.
(1) If there is some gammaiSo that gamma isiWhen 1, then γ remainsi；
If gamma isiNot equal to 1, for all γiConsider whether or not gamma is presentiSo that gamma isi＝0.99；
(2) If there is some gammaiSo that gamma isiWhen the ratio is 0.99, gamma is retainedi；
If gamma isiNot equal to 0.99 for all gammaiConsider whether or not gamma is presentiSo that gamma isi0.98; and so on.
The invention provides a data glove gesture recognition algorithm based on mathematical statistics. The method judges the deviation degree and linear relation between the randomly acquired gesture and the gesture in the template library by calculating the standard deviation and the correlation coefficient, and finds out the gesture which is most similar to the randomly acquired gesture in the template library. Through twice screening, the precision of the screening value is gradually improved, so that the gesture recognition is very accurate. And the gesture recognition based on the data glove can eliminate the interference of the external environment.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments.
Examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
Please refer to fig. 1, which is a flowchart illustrating the present invention.
The process of the scheme is as follows:
the first step is as follows: takes twenty kinds of single-hand mathematical gestures of hearing-impaired people as the standardBy wearing data gloves, the sensor. Reading the numerical value in the text document and storing the numerical value in the group a1And b1，a2And b2，…，a20And b20In (a)i＝ (ai,1,ai,2,ai,3,…,ai,24),bi＝(bi,1,bi,2,bi,3,…,bi,24)。
The second step is that: an operator wears the data glove, obtains the bending angle value of each sensor in real time through a sensor1,c2，…c24)。
The third step: the standard deviation was calculated according to the following formula. The variance is the mean of the squares of the differences between the respective data and their median, and the standard deviation is the square root of the variance. The smaller the standard deviation, the smaller the degree of offset between the data and its median is proved to be, indicating that data ═ c1,c2,...,c24) Cross over ai＝(ai,1,ai,2,ai,3,...,ai,24) And bi＝(bi,1,bi,2,bi,3,...,bi,24) Gesture made and aiAnd biThe more similar the gestures are represented.
(1) For any i (1< ═ i < ═ 20), the median was calculated:
Figure GDA0003482928340000071
(2) calculating the standard deviation:
Figure GDA0003482928340000072
the fourth step: after the standard deviation calculation is completed, the screening is performed according to the standard deviation, and the screening process is as follows, namely, the smallest standard deviation is screened.
(1) If there is a certain uiSo that (u)iAnd r) is 0, then u is retainedi；
If (u)iR) ≠ 0 for all uiConsider whether u is presentiSo that (u)i,r)＝1；
(2) If there is a certain uiSo that (u)iAnd r) 1 then holds ui；
If (u)iR) ≠ 1 for all uiConsider whether u is presentiSo that (u)iR) is 2; and so on.
The fifth step: since the first round of screening is performed by taking '1' as a unit, the screening is not very accurate, if the standard deviation values of several gestures are equal, a second round of screening is required, before screening, the correlation coefficient is calculated, and the calculation method is as follows. Wherein nm is the number of gestures that satisfy the condition in the first round of screening. Covariance is a measure of the cooperative variation of two random variables. The correlation coefficient is an index for measuring the linear correlation degree between two random variables, and when the correlation coefficient is "1", the two random variables are proved to have a linear relation with the probability of "1". When the correlation coefficient is larger, the degree of linear correlation is better, and conversely, the degree of linear correlation is poorer.
(1) Calculating the standard deviation:
Figure GDA0003482928340000081
Figure GDA0003482928340000082
......
Figure GDA0003482928340000083
and
Figure GDA0003482928340000084
Figure GDA0003482928340000085
......
Figure GDA0003482928340000086
(2) and (3) calculating covariance:
Figure GDA0003482928340000087
Figure GDA0003482928340000088
......
Figure GDA0003482928340000089
(3) calculating a correlation coefficient:
Figure GDA0003482928340000091
Figure GDA0003482928340000092
......
Figure GDA0003482928340000093
and finally: after the calculation of the correlation coefficient γ is completed, from γ ═ (γ)1,···,γnm) Screening out the largest correlation coefficientThe process is shown below, and the second round of screening is in units of 0.01, is relatively accurate and has strong feasibility.
(1) If there is some gammaiSo that gamma isiWhen 1, then γ remainsi；
If gamma isiNot equal to 1, for all γiConsider whether or not gamma is presentiSo that gamma isi＝0.99；
(2) If there is some gammaiSo that gamma isiWhen the ratio is 0.99, gamma is retainedi；
If gamma isiNot equal to 0.99 for all gammaiConsider whether or not gamma is presentiSo that gamma isi0.98; and so on.
Through twice screening, the precision of the screening value is gradually improved, so that the gesture recognition is very accurate.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.

Claims (3)

1. A data glove gesture recognition method based on mathematical statistics is characterized in that: the method comprises the following steps:
the first step is as follows: using a sensor.getdata () method for the worn data glove, dividing the same gesture into a natural state and an extended state to collect the angle values of the sensor, respectively storing the values in two text documents, reading the values in the text documents, and storing the values in a plurality of groups a1And b1，a2And b2，…，a20And b20In (a)i＝(ai,1,ai,2,ai,3,…,ai,24),bi＝(bi,1,bi,2,bi,3,…,bi,24)；
The second step is that: the operator wears the data glove, with the sensor mentioned in the first stepThe method includes acquiring bending angle values corresponding to the sensors in real time, and storing the bending angle values in a data array (c ═ c)1,c2，…c24)；
The third step: calculating the standard deviation in the following manner:
(1) for any i (1< ═ i < ═ 20), the average is calculated:
Figure FDA0003482928330000011
(2) calculating the standard deviation:
Figure FDA0003482928330000012
the fourth step: after the standard deviation calculation is finished, screening according to the standard deviation, namely screening out the minimum standard deviation;
the fifth step: and performing a second round of screening, wherein before screening, a correlation coefficient is calculated in the following way:
(1) calculating the standard deviation:
Figure FDA0003482928330000013
Figure FDA0003482928330000021
…
Figure FDA0003482928330000022
and
Figure FDA0003482928330000023
Figure FDA0003482928330000024
…
Figure FDA0003482928330000025
(2) and (3) calculating covariance:
Figure FDA0003482928330000036
Figure FDA0003482928330000031
…
Figure FDA0003482928330000032
(3) calculating a correlation coefficient:
Figure FDA0003482928330000033
Figure FDA0003482928330000034
…
Figure FDA0003482928330000035
and a sixth step: after the calculation of the correlation coefficient γ is completed, from γ ═ (γ)1,···,γnm) And screening out the maximum correlation coefficient.
2. The method of claim 1, wherein the method comprises a step of identifying a gesture on a glove based on mathematical statistics,
the method is characterized in that: the mode of screening the minimum standard deviation in the fourth step is as follows:
(1) if there is a certain uiSo that (u)iAnd r) is 0, then u is retainedi；
If (u)iR) ≠ 0 for all uiConsider whether u is presentiSo that (u)i,r)＝1；
(2) If there is a certain uiSo that (u)iAnd r) 1 then holds ui；
If (u)iR) ≠ 1 for all uiConsider whether u is presentiSo that (u)iR) is 2; and so on.
3. The method for recognizing the hand gesture of the data glove based on the mathematical statistics as claimed in claim 1, wherein: the screening process in the sixth step is as follows, the second round of screening is in units of "0.01",
(1) if there is some gammaiSo that gamma isiWhen 1, then γ remainsi；
If gamma isiNot equal to 1, for all γiConsider whether or not gamma is presentiSo that gamma isi＝0.99；
(2) If there is some gammaiSo that gamma isiWhen the ratio is 0.99, gamma is retainedi；
If gamma isiNot equal to 0.99 for all gammaiConsider whether or not gamma is presentiSo that gamma isi0.98; and so on.
CN201910721419.9A 2019-08-06 2019-08-06 Data glove gesture recognition algorithm based on mathematical statistics Active CN110414473B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910721419.9A CN110414473B (en) 2019-08-06 2019-08-06 Data glove gesture recognition algorithm based on mathematical statistics

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910721419.9A CN110414473B (en) 2019-08-06 2019-08-06 Data glove gesture recognition algorithm based on mathematical statistics

Publications (2)

Publication Number Publication Date
CN110414473A CN110414473A (en) 2019-11-05
CN110414473B true CN110414473B (en) 2022-02-25

Family

ID=68366080

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910721419.9A Active CN110414473B (en) 2019-08-06 2019-08-06 Data glove gesture recognition algorithm based on mathematical statistics

Country Status (1)

Country Link
CN (1) CN110414473B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105868715A (en) * 2016-03-29 2016-08-17 苏州科达科技股份有限公司 Hand gesture identifying method, apparatus and hand gesture learning system
CN106097835A (en) * 2016-06-03 2016-11-09 西安理工大学 A kind of deaf mute exchanges the method for intelligent assistance system and exchange
CN107678550A (en) * 2017-10-17 2018-02-09 哈尔滨理工大学 A kind of sign language gesture recognition system based on data glove
CN108268132A (en) * 2017-12-26 2018-07-10 北京航空航天大学 A kind of gesture identification method and human-computer interaction device based on gloves acquisition
CN109543644A (en) * 2018-06-28 2019-03-29 济南大学 A kind of recognition methods of multi-modal gesture
CN109885166A (en) * 2019-02-25 2019-06-14 浙江工商大学 Intelligent sign language translation gloves and its gesture identification method

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9619035B2 (en) * 2011-03-04 2017-04-11 Microsoft Technology Licensing, Llc Gesture detection and recognition
TW201544993A (en) * 2014-05-28 2015-12-01 Pegatron Corp Gesture control method, gesture control module, and wearable device having the same

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105868715A (en) * 2016-03-29 2016-08-17 苏州科达科技股份有限公司 Hand gesture identifying method, apparatus and hand gesture learning system
CN106097835A (en) * 2016-06-03 2016-11-09 西安理工大学 A kind of deaf mute exchanges the method for intelligent assistance system and exchange
CN107678550A (en) * 2017-10-17 2018-02-09 哈尔滨理工大学 A kind of sign language gesture recognition system based on data glove
CN108268132A (en) * 2017-12-26 2018-07-10 北京航空航天大学 A kind of gesture identification method and human-computer interaction device based on gloves acquisition
CN109543644A (en) * 2018-06-28 2019-03-29 济南大学 A kind of recognition methods of multi-modal gesture
CN109885166A (en) * 2019-02-25 2019-06-14 浙江工商大学 Intelligent sign language translation gloves and its gesture identification method

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Comparison of hand gesture inputs of leap motion controller & data glove in to a soft finger;P. D. S. H. Gunawardane等;《2017 IEEE International Symposium on Robotics and Intelligent Sensors》;20171007;第62-68页 *
Robust Hand Gesture Recognition for Human Machine Interaction System;Vijay Kumari Thakur等;《Journal of Global Research in Computer Science》;20140331;第5卷(第3期);第14-19页 *
一种视觉信息融合数据手套设计研究;王赋攀 等;《计算机研究与发展》;20181215;第2764-2774页 *

Also Published As

Publication number Publication date
CN110414473A (en) 2019-11-05

Similar Documents

Publication Publication Date Title
CN108182433A (en) A kind of meter reading recognition methods and system
CN109815826A (en) The generation method and device of face character model
CN105868716A (en) Method for human face recognition based on face geometrical features
CN110472481B (en) Sleeping gesture detection method, device and equipment
WO2022012364A1 (en) Electromyographic signal processing method and apparatus, and exoskeleton robot control method and apparatus
CN105760834A (en) Face feature point locating method
CN108197564B (en) Evaluation system and method for picture clock test
CN109101938A (en) A kind of multi-tag age estimation method based on convolutional neural networks
CN105956570B (en) Smiling face&#39;s recognition methods based on lip feature and deep learning
CN104268524A (en) Convolutional neural network image recognition method based on dynamic adjustment of training targets
CN111700718A (en) Holding posture identifying method, holding posture identifying device, artificial limb and readable storage medium
CN107590460A (en) Face classification method, apparatus and intelligent terminal
CN110163113B (en) Human behavior similarity calculation method and device
CN110007764B (en) Gesture skeleton recognition method, device and system and storage medium
WO2015176417A1 (en) Feature grouping normalization method for cognitive state recognition
CN110414473B (en) Data glove gesture recognition algorithm based on mathematical statistics
WO2022021948A1 (en) Action recognition method and apparatus, computer device, and storage medium
CN106529480A (en) Finger tip detection and gesture identification method and system based on depth information
CN110991294B (en) Face action unit recognition method and system capable of being quickly constructed
CN109508735A (en) A kind of soft margin support vector machine classification method based on neurodynamics
CN114237427B (en) High-sensitivity touch pressure detection method and system
CN115937910A (en) Palm print image identification method based on small sample measurement network
EP2889724B1 (en) System and method for selecting features for identifying human activities in a human-computer interacting environment
CN111553249B (en) H-B grading-based accurate facial paralysis degree evaluation method and device under CV
CN108154182A (en) Jewelry method for measuring similarity based on pairs of comparing cell

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