CN103336953B - A kind of method passed judgment on based on body sense equipment action - Google Patents
A kind of method passed judgment on based on body sense equipment action Download PDFInfo
- Publication number
- CN103336953B CN103336953B CN201310282049.6A CN201310282049A CN103336953B CN 103336953 B CN103336953 B CN 103336953B CN 201310282049 A CN201310282049 A CN 201310282049A CN 103336953 B CN103336953 B CN 103336953B
- Authority
- CN
- China
- Prior art keywords
- action
- curve
- correlation coefficients
- pearson correlation
- window
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Length Measuring Devices With Unspecified Measuring Means (AREA)
Abstract
The present invention provides a kind of method passed judgment on based on body sense equipment action, comprises the following steps: data gathering; For each skeleton point, the three dimensional space coordinate point of these discrete temporally series arrangement is carried out curve fitting; Ignore to fall the modification of time, obtain within for some time a set of action the curved path of process; Judge the similarity of two cover actions; As long as the comprehensive Pearson correlation coefficients of operation curve is greater than set threshold value, then judge that two cover actions are same set of actions, or it is enough accurate to say that imitator imitates standard cover action. Need not consider that the action that two people do is synchronous same frequency. To a set of action, the speed frequency of substep action need not be considered, as long as action sequence is accomplished in order. Because regardless of speed, the curve that action is spatially formed is the same. Need not considering individual body size difference, as long as action sequence is correct, namely the variation tendency of curve is correct, can obtain the Pearson correlation coefficients of high mark.
Description
Technical field
The present invention relates to a kind of action identification method, particularly relate to a kind of method passed judgment on based on body sense equipment action.
Background technology
One people imitates given body and moves work, how to judge whether he does right, and provides reasonable scoring; Or the action of another people is imitated at a people scene, how to judge whether he does right, and provides reasonable scoring. Summary of the invention
Given a set of standard action, another person imitates and does identical a set of action, judges whether he does right, and in order to solve the problem, the present invention provides a kind of method passed judgment on based on body sense equipment action.
Based on the method that body sense equipment action is passed judgment on, comprise the following steps:
Data gathering: utilizing body sense equipment to gather the three-dimensional space position of human body each position skeleton point in a time series, do a set of action for someone, each skeleton point has a string orderly three dimensional space coordinate;
For each skeleton point, the three dimensional space coordinate point of these discrete temporally series arrangement is carried out curve fitting;
Ignore to fall the modification of time, obtain within for some time a set of action the curved path of process;
Judge the similarity of two cover actions, first the similarity of these two plots changes of the corresponding skeleton point of two people is judged, the curve of matching carries out synchronous discrete point sampling, and (namely two curves are all sampled the point of identical number, so the spacing of two curve samplings is different), calculate the similarity of two plots changes with Pearson correlation coefficients, finally the similarity of all skeleton point paths tendency is carried out weight comprehensive;
As long as the comprehensive Pearson correlation coefficients of operation curve is greater than set threshold value, then judge that two cover actions are same set of actions, or it is enough accurate to say that imitator imitates standard cover action.
As a further improvement on the present invention, calculating the Pearson correlation coefficients of two curves in three-dimensional space, first the corresponding points on curve are projected x, y, z respectively, a dimension space in direction, every bar curve just obtains 3 vectors,
Article two, three dimension curves, obtain 3 groups of vectors,
Each group of vector is carried out the calculating of Pearson correlation coefficients r,
Wherein XiYi (i=1 ..., n) be the observed value component of vector (),Being two averages (vector component averages) respectively, SX, SY are standard deviations.
As a further improvement on the present invention, Pearson correlation coefficients reflection linear variable displacement degree of correlation, Pearson correlation coefficients is more big, show that positive correlation is more strong, the Pearson correlation coefficients of the three groups of vectors calculated is carried out weights comprehensively, just represents the Pearson correlation coefficients between these two three-dimensional space curves.
As a further improvement on the present invention, Pearson correlation coefficients does not reach threshold value, but centre has one section to be that complete doing is over a set of action, judge whether have complete a set of action to exist with sliding window, carry out the coupling of Pearson correlation coefficients with the curve in window and standard operation curve, get the action in the window mated most.
As a further improvement on the present invention, if curve A is operation curve to be matched, curve B is standard operation curve, assume that curve A comes from time series frame Z1, Z2, ..., Zt, with the window that slides of a fixed length, curve A is intercepted, represent and intercept [Zi, Zj] between the segment of the A of curve that forms of time series frame, wherein frame number Zj-Zi between Zi and Zj represents the length of this window, length of window is by artificially determining, this window slides from Zi=1, until Zi=Zt-(Zj-Zi), because Zj equals Zt at that time, often slide a window, curve A segment in window and curve B are carried out likelihood calculating, then it is the most similar with curve B to obtain certain curve A segment, then think this action sequence corresponding to curve A segment be do to effective action, above and action below be considered as invalid.
The invention has the beneficial effects as follows:
The advantage of two cover actions is compared with Pearson correlation coefficients:
Need not consider that the action that two people do is synchronous same frequency. To a set of action, the speed frequency of substep action need not be considered, as long as action sequence is accomplished in order. Because regardless of speed, the curve that action is spatially formed is the same.
Need not considering individual body size difference, as long as action sequence is correct, namely the variation tendency of curve is correct, can obtain the Pearson correlation coefficients of high mark.
Accompanying drawing explanation
Fig. 1 is that the present invention operation curve figure to be matched illustrates;
Fig. 2 is that standard operation curve figure of the present invention illustrates.
Embodiment
The present invention will be further described below.
Data gathering: utilize body sense equipment to gather the three-dimensional space position of human body each position skeleton point in a time series. Doing a set of action for someone, each skeleton point has a string orderly three dimensional space coordinate.
Determination methods:
For each skeleton point, the three dimensional space coordinate point of these discrete temporally series arrangement is carried out curve fitting;
Ignore to fall the modification of time, obtain within for some time a set of action the curved path of process.
Judge the similarity of two cover actions, in fact it is exactly the similarity of these two plots changes first judging the corresponding skeleton point of two people. The present invention's Pearson correlation coefficients calculates the similarity of two plots changes. Finally the similarity of all skeleton point paths tendency is carried out weight comprehensive.
As long as the comprehensive Pearson correlation coefficients of operation curve is greater than set threshold value, then can judge that two cover actions are same set of actions, or it is enough accurate to say that imitator imitates standard cover action.
Algorithm (technology) is explained: first the corresponding points on curve are projected x, y, z, a dimension space in direction by Pearson correlation coefficients respectively that calculate two curves in three-dimensional space, and every bar curve just obtains 3 vectors.
Article two, three dimension curves, obtain 3 groups of vectors.
Each group of vector is carried out the calculating of Pearson correlation coefficients.
Wherein XiYi (i=1 ..., n) be the observed value component of vector (),Being two averages (vector component averages) respectively, SXSY is standard deviation;
Pearson correlation coefficients reflection linear variable displacement degree of correlation. Pearson correlation coefficients is more big, shows that positive correlation is more strong.
The Pearson correlation coefficients of the three groups of vectors calculated is carried out weights comprehensively, just represents the Pearson correlation coefficients between these two three-dimensional space curves.
Pearson correlation coefficients does not reach threshold value and just illustrates that a set of action is not done accurately, and how wrong that has on earth
This just relates to several standards
Centre has one section to be that complete doing is over a set of action, before and after may wrong (not getting out/do not enter other actions after state/complete be taken), can judge whether have complete a set of action to exist with sliding window, carry out the coupling of Pearson correlation coefficients with the curve in window and standard operation curve, get the action in the window mated most.
Slide window and judge whether have complete a set of action to exist, specific as follows: to assume that curve A is operation curve to be matched, curve B is standard operation curve. assume that curve A comes from time series frame Z1, Z2 ..., Zt, with the window that slides of a fixed length, curve A is intercepted, represent the segment of the A of the curve that the time series frame intercepted between [Zi, Zj] is formed, wherein frame number Zj-Zi between Zi and Zj represents the length of this window, and length of window is by artificially determining. This window slides from Zi=1, until Zi=Zt-(Zj-Zi), because Zj equals Zt at that time. Often slide a window, curve A segment in window and curve B are carried out likelihood calculating, then it is the most similar with curve B to obtain certain curve A segment, we then think this action sequence corresponding to curve A segment be do to effective action, above and action below be considered as invalid.
Action is not in place (relates to the amplitude that action is not in place. When amplitude not in place is too big, naturally can be regarded as stroke defect; When amplitude not in place is little, Pearson correlation coefficients can not fluctuating error too big).
Action amplitude not in place too conference cause the error of Pearson correlation coefficients similarity-rough set.
It is too complicated that middle action does not do correct concept, is automatically judged to stroke defect.
Standard action: Fig. 2, action to be matched: Fig. 1, advantage: individual size difference need not be considered, although the individual build of Fig. 1 is taller and bigger, so action amplitude is relatively big, but identical with Fig. 2 curvilinear trend; Need not considering frequency: P1 ', the P2 ' of Fig. 1 and P1, P2 of Fig. 2 are corresponding points, the time difference between these two corresponding points can be different, and namely action does not affect the algorithm of the present invention from the speed that P1 ' moves to P2 '.
Above content is in conjunction with concrete preferred implementation further description made for the present invention, can not assert that specific embodiment of the invention is confined to these explanations. For general technical staff of the technical field of the invention, without departing from the inventive concept of the premise, it is also possible to make some simple deduction or replace, all should be considered as belonging to protection scope of the present invention.
Claims (5)
1. the method passed judgment on based on body sense equipment action, it is characterised in that, comprise the following steps:
Data gathering: utilizing body sense equipment to gather the three-dimensional space position of human body each position skeleton point in a time series, do a set of action for someone, each skeleton point has a string orderly three dimensional space coordinate;
For each skeleton point, the three dimensional space coordinate point of these discrete temporally series arrangement is carried out curve fitting;
Ignore to fall the modification of time, obtain within for some time a set of action the curved path of process;
Given a set of standard action, judge the similarity of a set of action and the standard action collected, first the similarity of these two plots changes of the corresponding skeleton point of two people is judged, the curve of matching is carried out synchronous discrete point sampling, article two, curve is all sampled the point of identical number, so the spacing of two curve samplings is different, calculates the similarity of two plots changes with Pearson correlation coefficients, finally the similarity of all skeleton point paths tendency is carried out weight comprehensive;
As long as the comprehensive Pearson correlation coefficients of operation curve is greater than set threshold value, then judge that two cover actions are same set of actions, or it is enough accurate to say that imitator imitates standard cover action.
2. a kind of method passed judgment on based on body sense equipment action according to claim 1, it is characterised in that: first the corresponding points on curve are projected x by Pearson correlation coefficients respectively that calculate two curves in three-dimensional space, y, one dimension space in z direction, every bar curve just obtains 3 vectors
Article two, three dimension curves, obtain 3 groups of vectors,
Each group of vector is carried out the calculating of Pearson correlation coefficients r,
Wherein XiYi (i=1 ..., n) be the observed value component of vector (),Being two averages (vector component averages) respectively, SX, SY are standard deviations.
3. a kind of method passed judgment on based on body sense equipment action according to claim 2, it is characterized in that: Pearson correlation coefficients reflection linear variable displacement degree of correlation, Pearson correlation coefficients is more big, show that positive correlation is more strong, the Pearson correlation coefficients of the three groups of vectors calculated is carried out weights comprehensively, just represents the Pearson correlation coefficients between these two three-dimensional space curves.
4. a kind of method passed judgment on based on body sense equipment action according to claim 1, it is characterized in that: Pearson correlation coefficients does not reach threshold value, but centre has one section to be that complete doing is over a set of action, judge whether have complete a set of action to exist with sliding window, carry out the coupling of Pearson correlation coefficients with the curve in window and standard operation curve, get the action in the window mated most.
5. a kind of method passed judgment on based on body sense equipment action according to claim 4, it is characterized in that: slide window and judge whether have complete a set of action to exist, specific as follows: to establish curve A to be operation curve to be matched, curve B is standard operation curve, assume that curve A comes from time series frame Z1, Z2, ..., Zt, with the window that slides of a fixed length, curve A is intercepted, represent and intercept [Zi, Zj] between the segment of the A of curve that forms of time series frame, wherein frame number Zj-Zi between Zi and Zj represents the length of this window, length of window is by artificially determining, this window slides from Zi=1, until Zi=Zt-(Zj-Zi), because Zj equals Zt at that time, often slide a window, curve A segment in window and curve B are carried out likelihood calculating, then it is the most similar with curve B to obtain certain curve A segment, then think this action sequence corresponding to curve A segment be do to effective action, above and action below be considered as invalid.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310282049.6A CN103336953B (en) | 2013-07-05 | 2013-07-05 | A kind of method passed judgment on based on body sense equipment action |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310282049.6A CN103336953B (en) | 2013-07-05 | 2013-07-05 | A kind of method passed judgment on based on body sense equipment action |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103336953A CN103336953A (en) | 2013-10-02 |
CN103336953B true CN103336953B (en) | 2016-06-01 |
Family
ID=49245110
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310282049.6A Active CN103336953B (en) | 2013-07-05 | 2013-07-05 | A kind of method passed judgment on based on body sense equipment action |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103336953B (en) |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106021926B (en) * | 2016-05-20 | 2019-06-18 | 北京九艺同兴科技有限公司 | A kind of real-time estimating method of human action sequence |
CN106683689B (en) * | 2017-03-13 | 2018-05-25 | 中南大学 | The long-range assessment method of language affinity and system |
CN107247924A (en) * | 2017-05-19 | 2017-10-13 | 安徽信息工程学院 | Action Compare System and comparison method based on Kinect |
CN108875469A (en) * | 2017-06-14 | 2018-11-23 | 北京旷视科技有限公司 | In vivo detection and identity authentication method, device and computer storage medium |
CN107920203A (en) * | 2017-11-23 | 2018-04-17 | 乐蜜有限公司 | Image-pickup method, device and electronic equipment |
CN108154125B (en) * | 2017-12-26 | 2021-08-24 | 深圳Tcl新技术有限公司 | Action teaching method, terminal and computer readable storage medium |
CN109492659B (en) * | 2018-09-25 | 2021-10-01 | 维灵(杭州)信息技术有限公司 | Method for calculating curve similarity for electrocardio and electroencephalogram waveform comparison |
CN112641441B (en) * | 2020-12-18 | 2024-01-02 | 河南翔宇医疗设备股份有限公司 | Posture evaluation method, system, device and computer readable storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101241600A (en) * | 2008-02-19 | 2008-08-13 | 深圳先进技术研究院 | Chain-shaped bone matching method in movement capturing technology |
CN101308500A (en) * | 2008-05-23 | 2008-11-19 | 浙江大学 | Visual and efficient three-dimensional human body movement data retrieval method based on demonstrated performance |
CN102800126A (en) * | 2012-07-04 | 2012-11-28 | 浙江大学 | Method for recovering real-time three-dimensional body posture based on multimodal fusion |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR100507780B1 (en) * | 2002-12-20 | 2005-08-17 | 한국전자통신연구원 | Apparatus and method for high-speed marker-free motion capture |
KR101804848B1 (en) * | 2011-04-22 | 2017-12-06 | 삼성전자주식회사 | Video Object Detecting Apparatus, Video Object Deforming Apparatus and Method thereof |
-
2013
- 2013-07-05 CN CN201310282049.6A patent/CN103336953B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101241600A (en) * | 2008-02-19 | 2008-08-13 | 深圳先进技术研究院 | Chain-shaped bone matching method in movement capturing technology |
CN101308500A (en) * | 2008-05-23 | 2008-11-19 | 浙江大学 | Visual and efficient three-dimensional human body movement data retrieval method based on demonstrated performance |
CN102800126A (en) * | 2012-07-04 | 2012-11-28 | 浙江大学 | Method for recovering real-time three-dimensional body posture based on multimodal fusion |
Non-Patent Citations (2)
Title |
---|
三维人体动作识别及其在交互舞蹈系统上的应用;邓利群;《万方学位论文》;20121130;第1-110页 * |
基于三维特征提取的人体运动数据分析和检索;向坚等;《计算机应用》;20080501;第28卷(第5期);第1344-1346页 * |
Also Published As
Publication number | Publication date |
---|---|
CN103336953A (en) | 2013-10-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103336953B (en) | A kind of method passed judgment on based on body sense equipment action | |
CN101558996B (en) | Gait recognition method based on orthogonal projection three-dimensional reconstruction of human motion structure | |
CN107358250B (en) | Body gait recognition methods and system based on the fusion of two waveband radar micro-doppler | |
CN104964685B (en) | A kind of decision method of mobile phone athletic posture | |
CN107240122A (en) | Video target tracking method based on space and time continuous correlation filtering | |
CN107506684A (en) | Gait recognition method and device | |
CN105652250B (en) | A kind of maneuvering target tracking technology based on double-deck expectation maximization | |
CN110047095A (en) | Tracking, device and terminal device based on target detection | |
CN103268497A (en) | Gesture detecting method for human face and application of gesture detecting method in human face identification | |
CN102629329B (en) | Personnel indoor positioning method based on adaptive SIFI (scale invariant feature transform) algorithm | |
CN114063153B (en) | Method and device for automatically inverting mechanism solution of seismic source | |
CN105116394A (en) | Object trace point extraction method based on multi-dimensional fine characteristic analysis | |
CN102298097A (en) | Method for estimating thunder impulse signal Time Difference of Arrival (TDOA) | |
CN106384363A (en) | Fast adaptive weight stereo matching algorithm | |
CN103150552B (en) | A kind of driving training management method based on number of people counting | |
CN114152980B (en) | Method and device for rapidly and automatically producing seismic source mechanism solution | |
CN103399346A (en) | Well-to-seismic integration initial wave impedance modeling method | |
CN105652235A (en) | Linear regression algorithm-based WLAN indoor positioning multi-user RSS (Received Signal Strength) fusion method | |
CN104754735A (en) | Construction method of position fingerprint database and positioning method based on position fingerprint database | |
CN102914302A (en) | Method for detecting unmanned aerial vehicle visual navigation landing cooperative target robust | |
CN104517100A (en) | Gesture pre-judging method and system | |
CN103297087A (en) | Arrival time estimation method for ultra-wideband positioning system | |
CN104112062A (en) | Method for obtaining wind resource distribution based on interpolation method | |
CN106487026A (en) | A kind of Coherent Generator Group method for quickly identifying based on wide area measurement information | |
CN106570479B (en) | A kind of pet motions recognition methods of Embedded platform |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: A method of motion evaluation based on somatosensory equipment Effective date of registration: 20210714 Granted publication date: 20160601 Pledgee: Shenzhen hi tech investment small loan Co.,Ltd. Pledgor: SHENZHEN VISTANDARD DIGITAL TECHNOLOGY Co.,Ltd. Registration number: Y2021980006159 |
|
PE01 | Entry into force of the registration of the contract for pledge of patent right |