CN103543826A - Method for recognizing gesture based on acceleration sensor - Google Patents
Method for recognizing gesture based on acceleration sensor Download PDFInfo
- Publication number
- CN103543826A CN103543826A CN201310337137.1A CN201310337137A CN103543826A CN 103543826 A CN103543826 A CN 103543826A CN 201310337137 A CN201310337137 A CN 201310337137A CN 103543826 A CN103543826 A CN 103543826A
- Authority
- CN
- China
- Prior art keywords
- gesture
- data
- template
- gesture data
- acceleration
- 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.)
- Pending
Links
Images
Landscapes
- User Interface Of Digital Computer (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method for recognizing a gesture based on an acceleration sensor. The method for recognizing the gesture based on the acceleration sensor comprises the steps that firstly, signals of the three-axis acceleration sensor is collected, a smooth denoising filter is designed and used for preprocessing the signals, and detection of the boundary of a gesture signal is automatically achieved; with the combination of a non-gesture filtering strategy, non-gesture data are filtered out before template matching; by the adoption of dynamic planning and a DTW algorithm, matching calculation is conducted on gesture data to be recognized and stored template data, and therefore a template gesture which is the most similar to the gesture to be recognized is found; through the strategies such as curve path limitation conducted by a slope, path area limitation and setting of a distorted threshold value, the template matching calculation amount is reduced, and the recognizing cost is reduced; through the template self-adaption strategy, a sample template base is updated automatically, and the gesture recognizing accuracy is improved. According to the method for recognizing the gesture based on the acceleration sensor, on the premise that energy consumption of a terminal is not obviously increased, better user experience is provided through high recognition efficiency and high recognition accuracy, and freer man-machine interaction is facilitated.
Description
Technical field
The present invention is a kind of gesture identification method based on acceleration transducer, belongs to human-computer interaction technique field.
Background technology
In recent years along with the development of computer technology and mechanics of communication, user requires more and more higher to computing machine degree easy to use and man-machine interaction, especially in virtual reality and wearable computing field, traditional man-machine interaction means more and more demonstrate their limitation as mouse and keyboard.Gesture, as conventional exchange way intuitively naturally in a kind of daily life, can be expressed specific meaning at many special scenes, than keyboard and the mouse in traditional man-machine interaction, gesture can with more freely, naturally form and machine mutual.The research method of gesture identification mainly contains three kinds at present: the method based on vision, the method based on electromyographic signal and the method based on inertial sensor.Gesture identification method based on image is high to equipment requirement, fund input is large, and is limited in particular place use, can be subject to the impact of background, shooting angle and occlusion issue, and real-time performance is poor.Research method based on electromyographic signal obtains relevant dynamic information by obtaining hand muscle electric signal, to identify gesture.Signals collecting must be close to skin and there being higher requirement aspect electrode configuration, only limit at present laboratory study.In recent years, along with the development of sensor technology and the improvement of manufacture craft thereof, the gesture identification based on acceleration transducer is risen just day by day.Acceleration transducer not only has the advantages such as size is little, precision is high, low in energy consumption, has more the advantage that is not subject to athletic ground and environmental restraint, and constantly spreads to and take on the intelligent terminal that mobile phone is representative, makes intelligent terminal possess perception from strength to strength.The feature of the perception means that intelligent terminal is abundant and " carry-on at any time ", makes the gesture research based on motion sensor more convenient, has more wide application prospect.
The accuracy of identification of the gesture identification based on acceleration transducer improves constantly, recognition efficiency constantly improves, but still there is following deficiency: 1, most research work are all the sensor devices based on special-purpose, in the application of gesture identification, user need to carry specialized equipment, is not suitable for applying on a large scale.2, in the process of gesture identification, need the operation bidirectionals such as user control button to represent beginning and the end of gesture, to user, bring extra operation burden, reduced user's Experience Degree.3, unilateral refinement degree, expense is excessive, does not consider that intelligent terminal is still resource-constrained equipment, especially battery flying power.4, in the processing procedure of gesture data, effectively do not distinguish gesture data and non-gesture data, before recognizer, non-gesture data is not rejected, reduce the efficiency of gesture identification, be unfavorable for the permanent use of intelligent terminal.5, due to gesture diversity and the difference of carrying out gesture individuality, same recognition methods, the discrimination in different crowd is different, need be according to the otherness of the diversity of gesture and gesture individuality, take effective recognition strategy, improve efficiency and the precision of gesture identification.
Summary of the invention
Object of the present invention: design is applicable to the Gesture Recognition Algorithm of resource-constrained intelligent terminal, adopt gesture Boundary Detection strategy, automatically judge beginning and the end of gesture, and in conjunction with non-gesture filtering policy, effectively filter non-gesture data, adopt the thought of dynamic programming and DWT (Dynamic Time Warping, dynamic time warping) algorithm, design Gesture Recognition Algorithm, the efficiency of raising gesture identification; Adopt template place adaptive strategy, improve the precision of gesture identification.
The technical scheme that the present invention adopts is in order to achieve the above object: a kind of Gesture Recognition Algorithm based on acceleration transducer, and its step is as follows:
Step (1), gesture data collection: the three-dimensional acceleration information of utilizing the acceleration transducer collection dynamic gesture on intelligent terminal;
Step (2), gesture data smoothing denoising: design smoothing filter carries out steady denoising to the 3-axis acceleration data of obtaining in gesture process, rejects due to the shake of hand in gesture process and the noise of the precision of sensor own;
Step (3), gesture Boundary Detection: take gesture motion decision plan, automatically detect beginning and the end of gesture, do not need the extra manually appointment of user, improve user experience;
Step (4), non-gesture data filter: the gesture data through gesture Boundary Detection is sent into non-gesture filtration stage.Although non-gesture strobe utility can not be judged concrete gesture, can be that gesture data is also non-gesture data with higher accuracy judgement data stream;
Step (5), gesture data quantize: before gesture identification coupling is calculated, need to quantize gesture data, reduce gesture data sequence length, improve counting yield; Very practical for resource-constrained intelligent terminal;
Step (6), DTW algorithmic match are calculated: utilize DTW algorithm, gesture data to be identified is mated to calculating with the sample template data of storage, thereby find the sample template gesture nearest with gesture to be identified, complete the object of gesture identification;
Step (7), template self-adaptation: gesture is due to polysemy and diversity, different users is when carrying out same gesture, gesture data sequence can show obvious difference, need to be according to the variation of user's gesture custom, sample template is taked to efficient adaptive strategy, select optimal sample template.
Further, the detailed process that gesture data in described step (1) gathers is as follows: the activity of mankind's gesture is generally in 10Hz left and right, according to Nyquist law, when sample frequency is greater than 2 times of highest frequency in signal, the digital signal after sampling has intactly retained the information of original signal.The sample frequency of the acceleration transducer that the present invention adopts is 50Hz.
Further, the detailed process of the gesture data smoothing denoising in described step (2) is as follows: adopt SMA (Simp1e Moving Average, simple Moving Average) method of wave filter is carried out denoising reposefully to the acceleration information obtaining, and elimination is due to shake and the caused noise of the precision of sensor own of inevitable hand in gesture process.The magnitude relationship of data sequence length n in SMA wave filter is to level and smooth effect.N is too little, steadily DeGrain; N is excessive, steadily effective, but easily causes gesture information to be lost.With reference to experience, the length n of data sequence goes 5 to 15.
Further, the detailed process of the gesture Boundary Detection in described step (3) is as follows: by the strategy of threshold value, automatically detect beginning and the end of gesture, and do not need user manually to specify, thereby improve user experience.If the gesture data energy value gathering has surpassed a certain threshold value, think that gesture motion starts; After gesture starts, if the gesture data energy value gathering does not surpass a certain reservation threshold in a certain official hour section, think that gesture finishes.During gesture motion based on threshold value is judged, the setting of threshold values has largely determined the accuracy of judging.If threshold values is too low, the casual trickle action of user all can be judged to be the beginning of gesture, and threshold values is too high, can miss the judgement that starts of normal gesture.The present invention utilizes a stronger threshold values to detect the motion of gesture, judges on this basis beginning and the end of gesture by two less threshold values forward and backward.
Further, the detailed process that non-gesture data in described step (4) filters is as follows: utilize the non-gesture limiting based on border to filter the main maximal value of investigating gesture acceleration, minimum value whether in the scope of a certain restriction, and whether the lasting duration of gesture is in a certain limited range.If acceleration magnitude and time span, in limited range, are thought gesture data; Otherwise be judged to be non-gesture data, directly abandon.Judge that range boundary value is definite when training sample template data.
Further, the detailed process that gesture data in described step (5) quantizes is as follows: quantizing process carries out in two steps: the one, resample, so both can retain the information of gesture motion, greatly reduce again the length of gesture data sequence, thereby alleviated follow-up template matches, calculate.The 2nd, by for the acceleration information of floating point values is converted to shaping data, thereby avoided Floating-point Computation.This is very practical for resource-constrained intelligent terminal.
Further, the specific implementation process that the DTW algorithmic match of described step (6) is calculated is as follows: because gesture exists diversity, even repeat same gesture, each obtained gesture data also can show otherness on room and time.The thought of DTW algorithm based on dynamic programming, adopts Time alignment, eliminates gesture data to be identified and the temporal otherness of sample template gesture data; Utilize DTW Cumulative Distance to measure the otherness on gesture data to be identified and sample template gesture data space, according to the size of DTW Cumulative Distance, identify concrete gesture.
Further, the white specific implementation process adapting to of the template of described step (7) is as follows: DTW matching algorithm mates calculating by the data of collection with the sample template of training, and the quality of sample template must be related to the precision of coupling.Gesture is due to its polysemy and diversity, and different users is when carrying out same gesture, and gesture data sequence can show obvious difference; Even same user carries out same gesture, there is obvious difference at every turn, need to be according to the variation of user's gesture custom, sample template is taked to efficient adaptive strategy, select optimal sample template, thereby reduce because of the impact of user behavior custom on gesture recognition system, improve the precision of gesture identification.
Accompanying drawing explanation
The gesture identification method framework of Fig. 1 based on acceleration transducer;
Fig. 2 smoothing denoising rear right gesture 3-axis acceleration measured value figure that swings one's arm;
Fig. 3 gesture 3-axis acceleration energy value of swinging one's arm on the right side, changing value and weighted sum value figure;
Gesture border start-stop detection figure swings one's arm on the right side that Fig. 4 is cut apart based on energy threshold.
Embodiment
Below in conjunction with accompanying drawing and embodiment, the present invention is described in further detail:
The invention process process comprises five Main Stage: to acceleration transducer signals pre-service; Online acceleration signal sequence is cut apart and isolated corresponding gesture motion sequence; Whether differentiate gesture motion sequence to be tested is that gesture motion data are also non-gesture motion sequences; Utilize DTW algorithm and template matches principle to identify concrete gesture; Gesture template self-adaptation is adjusted.
Stage one, to acceleration transducer signals pre-service, because user moves or the inevitably impact of shake and the precision of sensor own of hand in gesture process, the gesture motion data of acceleration transducer collection are unavoidably subject to noise, utilize the method for SMA (Simple Moving Average, simple Moving Average) wave filter to carry out steady denoising to the acceleration information obtaining.The derivation formula of SMA wave filter is as follows:
Y
i＝(X
i+X
i-1+ΛX
i-(n-1))/n n＝1,2,3,4，Λ
In above formula, n represents the length of data sequence.The magnitude relationship of n is to level and smooth effect.N is too little, steadily DeGrain; N is excessive, steadily effective, but easily causes gesture information to be lost.With reference to experience, depending on different situations, n generally gets 5 to 15.
When calculating, can directly use following formula:
Y
now＝Y
previous-X
i-n/n+X
i/n
The data of the raw data of gesture 1 after SMA filtrator is steady as shown in the figure.
In the stage two, cut apart acceleration signal sequence and isolate gesture acceleration signal sequence.By the strategy of threshold value, automatically detect beginning and the end of gesture, and do not need user manually to specify, thereby improve user experience.If the gesture data energy value gathering has surpassed a certain threshold value, think that gesture motion starts; After gesture starts, if the gesture data gathering does not surpass a certain reservation threshold in a certain official hour section, think that gesture finishes.Decision plan formula based on threshold value is as follows:
Wherein C, for regulating parameter, determines that current 3-axis acceleration energy value and current 3-axis acceleration value are with respect to the importance of last measurement variation amount, ε
bfor the threshold values of setting.During gesture motion based on threshold value is judged, the setting of threshold values has largely determined the accuracy of judging.If threshold values is too low, the casual trickle action of user all can be judged to be the beginning of gesture, and threshold values is too high, can miss the judgement that starts of normal gesture.The present invention utilizes a stronger threshold values to detect the motion of gesture, judges on this basis beginning and the end of gesture by two less threshold values forward and backward.
Stage three, the filtration to non-gesture data.Whether the maximal value of the main investigation of the non-gesture filtration gesture acceleration limiting based on border, minimum value be at a certain limited range [A
min, A
max] within, and whether the lasting duration of gesture is at a certain restriction [T
min, T
max] within, be shown below:
T∈[T
min，T
max]
Through isolated gesture acceleration information of stage two, if acceleration magnitude and time span, within limited range, are thought gesture data; Otherwise, be judged to be non-gesture data, directly abandon.For fear of being non-gesture by gesture identification, border limits and should suitably relax.Can when training sample template data, determine.
In the stage four, utilize DTW algorithm and template matches principle to identify concrete gesture.Gesture data to be identified is mated to calculating with the sample template data of storage, thereby find the sample template gesture nearest with gesture to be identified, complete the object of gesture identification.Before carrying out template matches calculating, need training in advance sample template data.Suppose that gesture data to be identified is T[N]={ T
1, T
2, Λ T
n, total N eigenvector, the gesture data of certain reference template is R[M]={ R
1, R
2, Λ R
m, total M eigenvector, T
jand R
jall by three-dimensional accekeration, formed, comprise X, Y, Z axis three components, namely T
i={ T
i.x, T
i.y, T
i.z, R
i={ R
i.x, R
i.y, R
i.z.DTW matching algorithm is exactly to want hunting time warping function j=w (i), makes gesture vector R[M to be identified] time shaft i be non-linearly mapped to gesture sample template T[N] time shaft j upper, and meet:
D[T in formula (i), R (w (i))] be the distance measure between i frame test vector T (i) and j pin gesture template vector R (j).Dist is the coupling path between two vectors under optimal situation.
In order to make dynamic route search problem have more practical significance, must must meet constraint condition to Time alignment function:
1, Experience about Monotonicity of Functions restriction.The gesture data gathering due to acceleration transducer timing in time, warping function must guarantee that the path of coupling is without prejudice to the time sequencing of index sequence, namely warping function must meet monotonicity restriction:
w(i+1)≥w(i)
2, continuity restriction.In order to guarantee the just definite of identification, warping function does not allow to skip any one match point.
3, in order to make path inclination within reason, can be by Slope Constraint within the scope of 0-2, if path is by lattice point (n
i, m
i), so previous node (n
i-1, m
i-1) may be only one of following three axle situations: (n
i-1, m
i), (n
i-1, m
i-1), (n
i-1, m
i-2), the cumulative distance in path:
D[(n
i，m
j)]＝d[R(n
i)，T(m
i)]+D[(n
i-1，m
i-1)]
(n wherein
i-1, m
i-1) by following formula, determined:
D[(n
i-1，m
i-1)]＝min{D[n
i-1，m
i]，D[(n
i-1，m
i-1)]，D[(n
i-1，m
i-2)]}
4,, in the coupling computation process of gesture identification, the DTW Cumulative Distance between sign language sample template gesture to be identified constantly increases along with difference between the two.If gesture to be identified and gesture sample template belong to same class gesture, DTW Cumulative Distance is little; If belong to non-similar gesture, DTW Cumulative Distance is large.Gesture recognition system gesture the most to be identified is classified as same class gesture with the sample template gesture with minimum DTW Cumulative Distance, complete the identification of gesture, by threshold values is set, once current calculating DTW Cumulative Distance surpasses distortion threshold value, represent that final recognition result is not the represented gesture of this template, can stop immediately the continuation of this template to calculate, thereby reduce a large amount of computing costs, improve the efficiency of algorithm.The setting of distortion threshold value, is directly connected to efficiency and the precision of Gesture Recognition Algorithm, if distortion threshold value is too high, to non-similar gesture without cutting effect, cannot play the effect that reduces calculated amount; If distortion threshold value is too low, will the similar sample calligraphy or painting model of cutting gesture, greatly reduce the precision of algorithm.The present invention adopts average DTW Cumulative Distance to set distortion threshold value.When gesture to be identified mates with sample calligraphy or painting model gesture the n frame that calculates gesture to be identified, average DTW Cumulative Distance is:
Stage five, the self-adaptation adjustment of gesture sample template.Gesture Recognition Algorithm need to mate calculating in gesture sample template by the gesture data collecting, and the quality of sample template must be related to the precision of coupling.Due to diversity and the polysemy of gesture, different users carries out same gesture, even same user carries out same gesture, gesture data sequence has obvious difference.The present invention takes two kinds of simple template adaptive strategies: 1, system selects the gesture data of DTW matching distance minimum on the same day as the sample template of such gesture automatically for each gesture gathers a sample template every day automatically; 2,, when gesture identification mistake, there is user to specify the sample template of corresponding gesture.
Claims (7)
1. the gesture identification method based on acceleration transducer, is characterized in that the method step is as follows:
Step (1), gesture data collection: the three-dimensional acceleration information of utilizing the acceleration transducer collection dynamic gesture on intelligent terminal;
Step (2), gesture data smoothing filter: design smoothing filter carries out steady denoising to the 3-axis acceleration data of obtaining in gesture process, rejects due to the shake of hand in gesture process and the noise of the precision of sensor own;
Step (3), gesture Boundary Detection: take gesture motion decision plan, automatically detect beginning and the end of gesture, do not need the extra manually appointment of user, improve user experience;
Step (4), non-gesture data filter: the gesture data through gesture Boundary Detection is sent into non-gesture filtration stage.Although non-gesture strobe utility can not be judged concrete gesture, can be that gesture data is also non-gesture data with higher accuracy judgement data stream;
Step (5), gesture data quantize: before gesture identification coupling is calculated, need to quantize gesture data, reduce gesture data sequence length, improve counting yield; Very practical for resource-constrained intelligent terminal;
Step (6), DTW algorithmic match are calculated: utilize DTW algorithm, gesture data to be identified is mated to calculating with the sample template data of storage, thereby find the sample template gesture nearest with gesture to be identified, complete the object of gesture identification;
Step (7), template self-adaptation: gesture is due to polysemy and diversity, different users is when carrying out same gesture, gesture data sequence can show obvious difference, need to be according to user in the variation of gesture custom, sample template is taked to efficient adaptive strategy, select optimal sample template.
2. the motion recognition methods based on acceleration sensor according to claim 1, it is characterized in that: the detailed process of the gesture data smoothing denoising in described step (2) is as follows: adopt SMA (Simple Moving Average, simple Moving Average) method of wave filter is carried out steady denoising to the gesture acceleration information obtaining, and elimination is due to shake and the caused noise of the precision of sensor own of inevitable hand in gesture process.
3. the motion recognition methods based on acceleration transducer according to claim 1, it is characterized in that: the detailed process of the gesture Boundary Detection in described step (3) is as follows: by the strategy of threshold value, automatically detect beginning and the end of gesture, and do not need user manually to specify, thereby improve user experience.If the gesture data energy value gathering has surpassed a certain threshold value, think that gesture motion starts; After gesture starts, if the gesture data energy value gathering does not surpass a certain reservation threshold in a certain official hour section, think that gesture finishes.
4. the motion recognition methods based on acceleration transducer according to claim 1, it is characterized in that: the detailed process that the non-gesture data in described step (4) filters is as follows: utilize the non-gesture limiting based on border to filter the main maximal value of investigating gesture acceleration, minimum value whether in the scope of a certain restriction, and whether the lasting duration of gesture is in a certain limited range.If acceleration magnitude and time span, in limited range, are thought gesture data; Otherwise be judged to be non-gesture data, directly abandon.
5. the motion recognition methods based on acceleration transducer according to claim 1, it is characterized in that: the detailed process that the gesture data in described step (5) quantizes is as follows: quantizing process carries out in two steps: the one, resample, so both can retain the information of gesture motion, greatly reduced again in the length of gesture data sequence, thereby alleviated follow-up template matches, calculated.The 2nd, by for the acceleration information of floating point values is converted to shaping data, thereby avoided Floating-point Computation.
6. the motion recognition methods based on acceleration transducer according to claim 1, it is characterized in that: it is as follows that the DTW matching algorithm in described step (6) calculates detailed process: because gesture exists diversity, even repeat same gesture, each obtained gesture data also can show otherness on room and time.The thought of DTW algorithm based on dynamic programming, adopts Time alignment, eliminates gesture data to be identified and the temporal otherness of sample template gesture data; Utilize DTW Cumulative Distance to measure the otherness on gesture data to be identified and sample template gesture data space, according to the size of DTW Cumulative Distance, identify concrete gesture.By slope, limit curved path, limit region, path and set the strategies such as distortion threshold value, alleviate the calculated amount of template matches, reduce the expense of gesture identification.
7. the motion recognition methods based on acceleration transducer according to claim 1, it is characterized in that: the template self-adaptation detailed process in described step (7) is as follows: 1, system selects the gesture data of DTW matching distance minimum on the same day as the sample template of such gesture automatically for each gesture gathers a sample template every day automatically; 2,, when gesture identification mistake, there is user to specify the sample template of corresponding gesture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310337137.1A CN103543826A (en) | 2013-07-30 | 2013-07-30 | Method for recognizing gesture based on acceleration sensor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310337137.1A CN103543826A (en) | 2013-07-30 | 2013-07-30 | Method for recognizing gesture based on acceleration sensor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103543826A true CN103543826A (en) | 2014-01-29 |
Family
ID=49967363
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310337137.1A Pending CN103543826A (en) | 2013-07-30 | 2013-07-30 | Method for recognizing gesture based on acceleration sensor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103543826A (en) |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103984416A (en) * | 2014-06-10 | 2014-08-13 | 北京邮电大学 | Gesture recognition method based on acceleration sensor |
| CN104267835A (en) * | 2014-09-12 | 2015-01-07 | 西安闻泰电子科技有限公司 | Self-adaption gesture recognition method |
| CN105024817A (en) * | 2014-04-29 | 2015-11-04 | 中国移动通信集团湖南有限公司 | Identity authentication method and identity authentication device |
| CN105100376A (en) * | 2014-05-16 | 2015-11-25 | 中国移动通信集团湖南有限公司 | Identity authentication method and apparatus |
| CN105184325A (en) * | 2015-09-23 | 2015-12-23 | 歌尔声学股份有限公司 | Human body action recognition method and mobile intelligent terminal |
| CN105511643A (en) * | 2014-09-30 | 2016-04-20 | 深圳矽递科技有限公司 | Gesture recognition method based on three-axis acceleration sensor |
| CN105677039A (en) * | 2016-02-16 | 2016-06-15 | 北京博研智通科技有限公司 | Method, device and wearable device for gesture-based driving status detection |
| CN105786182A (en) * | 2016-02-26 | 2016-07-20 | 深圳还是威健康科技有限公司 | Method and device for controlling periphery devices based on gesture |
| CN105892301A (en) * | 2016-05-19 | 2016-08-24 | 深圳还是威健康科技有限公司 | Smart home control method and device |
| CN106339104A (en) * | 2016-08-24 | 2017-01-18 | 广州市香港科大霍英东研究院 | Text input method and device for smart watch |
| CN106484088A (en) * | 2015-09-02 | 2017-03-08 | 腾讯科技（深圳）有限公司 | The control method of intelligent mobile terminal, device and intelligent mobile terminal |
| WO2017050140A1 (en) * | 2015-09-23 | 2017-03-30 | 歌尔股份有限公司 | Method for recognizing a human motion, method for recognizing a user action and smart terminal |
| CN106774894A (en) * | 2016-12-16 | 2017-05-31 | 重庆大学 | Interactive teaching methods and interactive system based on gesture |
| CN106774995A (en) * | 2016-12-14 | 2017-05-31 | 吉林大学 | A kind of three-dimensional style of brushwork recognition methods based on localization by ultrasonic |
| CN107153454A (en) * | 2016-03-02 | 2017-09-12 | 曦恩体感科技股份有限公司 | The method and mobile device calibrated with sensor time |
| CN107346207A (en) * | 2017-06-30 | 2017-11-14 | 广州幻境科技有限公司 | A kind of dynamic gesture cutting recognition methods based on HMM |
| CN107463257A (en) * | 2017-08-03 | 2017-12-12 | 微景天下（北京）科技有限公司 | A kind of man-machine interaction method and device of Virtual Reality system |
| CN107766930A (en) * | 2017-09-06 | 2018-03-06 | 华东师范大学 | Based on the fuzzy equivalent ROM distance calculating methods for dividing group's SOM neurons of DTW clusters |
| CN108108015A (en) * | 2017-11-20 | 2018-06-01 | 电子科技大学 | A kind of action gesture recognition methods based on mobile phone gyroscope and dynamic time warping |
| CN108319421A (en) * | 2018-01-29 | 2018-07-24 | 维沃移动通信有限公司 | A kind of display triggering method and mobile terminal |
| CN108664277A (en) * | 2018-05-15 | 2018-10-16 | 精伦电子股份有限公司 | Awakening method and device are lifted based on 3-axis acceleration sensor |
| CN110046585A (en) * | 2019-04-19 | 2019-07-23 | 西北工业大学 | A kind of gesture identification method based on environment light |
| CN110624217A (en) * | 2019-09-23 | 2019-12-31 | 孙孟雯 | Rehabilitation glove based on multi-sensor fusion and implementation method thereof |
| CN110638421A (en) * | 2014-09-23 | 2020-01-03 | 飞比特公司 | Method, system, and apparatus for displaying visibility changes in response to user gestures |
| US10551211B2 (en) | 2013-05-08 | 2020-02-04 | Cm Hk Limited | Methods and devices with sensor time calibration |
| WO2020140266A1 (en) * | 2019-01-04 | 2020-07-09 | 深圳大学 | Interaction method and interactive system for smart watch |
| CN111665934A (en) * | 2020-04-30 | 2020-09-15 | 哈尔滨理工大学 | Gesture recognition system and method based on ZYNQ software and hardware coprocessing |
| CN113567794A (en) * | 2021-09-24 | 2021-10-29 | 国网江苏省电力有限公司营销服务中心 | Electric bicycle indoor charging identification method and system based on dynamic time warping |
| CN117572963A (en) * | 2023-11-06 | 2024-02-20 | 深圳市腾进达信息技术有限公司 | Method for controlling operation of intelligent wearable device based on motion capture technology |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101694692A (en) * | 2009-10-22 | 2010-04-14 | 浙江大学 | Gesture identification method based on acceleration transducer |
| CN101853380A (en) * | 2010-06-18 | 2010-10-06 | 华南理工大学 | Rotary characteristic extraction method based on gesture identification of acceleration sensor |
-
2013
- 2013-07-30 CN CN201310337137.1A patent/CN103543826A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101694692A (en) * | 2009-10-22 | 2010-04-14 | 浙江大学 | Gesture identification method based on acceleration transducer |
| CN101853380A (en) * | 2010-06-18 | 2010-10-06 | 华南理工大学 | Rotary characteristic extraction method based on gesture identification of acceleration sensor |
Non-Patent Citations (1)
| Title |
|---|
| 陈文: "基于加速度传感器的智能终端手势识别关键技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (46)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10551211B2 (en) | 2013-05-08 | 2020-02-04 | Cm Hk Limited | Methods and devices with sensor time calibration |
| CN105024817A (en) * | 2014-04-29 | 2015-11-04 | 中国移动通信集团湖南有限公司 | Identity authentication method and identity authentication device |
| CN105024817B (en) * | 2014-04-29 | 2018-11-02 | 中国移动通信集团湖南有限公司 | A kind of identity identifying method and device |
| CN105100376A (en) * | 2014-05-16 | 2015-11-25 | 中国移动通信集团湖南有限公司 | Identity authentication method and apparatus |
| CN103984416B (en) * | 2014-06-10 | 2017-02-08 | 北京邮电大学 | Gesture recognition method based on acceleration sensor |
| CN103984416A (en) * | 2014-06-10 | 2014-08-13 | 北京邮电大学 | Gesture recognition method based on acceleration sensor |
| CN104267835B (en) * | 2014-09-12 | 2017-08-08 | 西安闻泰电子科技有限公司 | Adaptive gesture identification method |
| CN104267835A (en) * | 2014-09-12 | 2015-01-07 | 西安闻泰电子科技有限公司 | Self-adaption gesture recognition method |
| CN110638421B (en) * | 2014-09-23 | 2022-07-01 | 飞比特有限公司 | Method, system, and apparatus for displaying visibility changes in response to user gestures |
| CN110638421A (en) * | 2014-09-23 | 2020-01-03 | 飞比特公司 | Method, system, and apparatus for displaying visibility changes in response to user gestures |
| CN105511643A (en) * | 2014-09-30 | 2016-04-20 | 深圳矽递科技有限公司 | Gesture recognition method based on three-axis acceleration sensor |
| CN105511643B (en) * | 2014-09-30 | 2018-08-31 | 深圳矽递科技股份有限公司 | A kind of gesture identification method based on 3-axis acceleration sensor |
| CN106484088A (en) * | 2015-09-02 | 2017-03-08 | 腾讯科技（深圳）有限公司 | The control method of intelligent mobile terminal, device and intelligent mobile terminal |
| US10339371B2 (en) | 2015-09-23 | 2019-07-02 | Goertek Inc. | Method for recognizing a human motion, method for recognizing a user action and smart terminal |
| WO2017050140A1 (en) * | 2015-09-23 | 2017-03-30 | 歌尔股份有限公司 | Method for recognizing a human motion, method for recognizing a user action and smart terminal |
| CN105184325A (en) * | 2015-09-23 | 2015-12-23 | 歌尔声学股份有限公司 | Human body action recognition method and mobile intelligent terminal |
| CN105677039A (en) * | 2016-02-16 | 2016-06-15 | 北京博研智通科技有限公司 | Method, device and wearable device for gesture-based driving status detection |
| CN105677039B (en) * | 2016-02-16 | 2020-06-09 | 北京博研智通科技有限公司 | Method and device for detecting driving state based on gesture and wearable device |
| CN105786182B (en) * | 2016-02-26 | 2018-12-07 | 深圳还是威健康科技有限公司 | A kind of method and device based on gesture control surrounding devices |
| CN105786182A (en) * | 2016-02-26 | 2016-07-20 | 深圳还是威健康科技有限公司 | Method and device for controlling periphery devices based on gesture |
| CN107153454A (en) * | 2016-03-02 | 2017-09-12 | 曦恩体感科技股份有限公司 | The method and mobile device calibrated with sensor time |
| CN107153454B (en) * | 2016-03-02 | 2020-10-09 | 曦恩体感科技股份有限公司 | Method and mobile device with sensor time calibration |
| CN105892301A (en) * | 2016-05-19 | 2016-08-24 | 深圳还是威健康科技有限公司 | Smart home control method and device |
| WO2018036023A1 (en) * | 2016-08-24 | 2018-03-01 | 广州市香港科大霍英东研究院 | Text input method and device for smart watch |
| CN106339104A (en) * | 2016-08-24 | 2017-01-18 | 广州市香港科大霍英东研究院 | Text input method and device for smart watch |
| CN106339104B (en) * | 2016-08-24 | 2019-02-15 | 广州市香港科大霍英东研究院 | The text entry method and device of smartwatch |
| CN106774995B (en) * | 2016-12-14 | 2019-05-03 | 吉林大学 | A kind of three-dimensional style of brushwork recognition methods based on localization by ultrasonic |
| CN106774995A (en) * | 2016-12-14 | 2017-05-31 | 吉林大学 | A kind of three-dimensional style of brushwork recognition methods based on localization by ultrasonic |
| CN106774894A (en) * | 2016-12-16 | 2017-05-31 | 重庆大学 | Interactive teaching methods and interactive system based on gesture |
| CN107346207B (en) * | 2017-06-30 | 2019-12-20 | 广州幻境科技有限公司 | Dynamic gesture segmentation recognition method based on hidden Markov model |
| CN107346207A (en) * | 2017-06-30 | 2017-11-14 | 广州幻境科技有限公司 | A kind of dynamic gesture cutting recognition methods based on HMM |
| CN107463257B (en) * | 2017-08-03 | 2020-08-21 | 微景天下（北京）科技有限公司 | Human-computer interaction method and device of virtual reality VR system |
| CN107463257A (en) * | 2017-08-03 | 2017-12-12 | 微景天下（北京）科技有限公司 | A kind of man-machine interaction method and device of Virtual Reality system |
| CN107766930A (en) * | 2017-09-06 | 2018-03-06 | 华东师范大学 | Based on the fuzzy equivalent ROM distance calculating methods for dividing group's SOM neurons of DTW clusters |
| CN107766930B (en) * | 2017-09-06 | 2021-03-26 | 华东师范大学 | Equivalent ROM distance calculation method based on DTW cluster fuzzy clustering SOM neurons |
| CN108108015A (en) * | 2017-11-20 | 2018-06-01 | 电子科技大学 | A kind of action gesture recognition methods based on mobile phone gyroscope and dynamic time warping |
| CN108319421A (en) * | 2018-01-29 | 2018-07-24 | 维沃移动通信有限公司 | A kind of display triggering method and mobile terminal |
| CN108664277A (en) * | 2018-05-15 | 2018-10-16 | 精伦电子股份有限公司 | Awakening method and device are lifted based on 3-axis acceleration sensor |
| CN108664277B (en) * | 2018-05-15 | 2021-11-30 | 精伦电子股份有限公司 | Lifting awakening method and device based on triaxial acceleration sensor |
| US11270109B2 (en) | 2019-01-04 | 2022-03-08 | Shenzhen Unnersity | Interactive method and interactive system for smart watch |
| WO2020140266A1 (en) * | 2019-01-04 | 2020-07-09 | 深圳大学 | Interaction method and interactive system for smart watch |
| CN110046585A (en) * | 2019-04-19 | 2019-07-23 | 西北工业大学 | A kind of gesture identification method based on environment light |
| CN110624217A (en) * | 2019-09-23 | 2019-12-31 | 孙孟雯 | Rehabilitation glove based on multi-sensor fusion and implementation method thereof |
| CN111665934A (en) * | 2020-04-30 | 2020-09-15 | 哈尔滨理工大学 | Gesture recognition system and method based on ZYNQ software and hardware coprocessing |
| CN113567794A (en) * | 2021-09-24 | 2021-10-29 | 国网江苏省电力有限公司营销服务中心 | Electric bicycle indoor charging identification method and system based on dynamic time warping |
| CN117572963A (en) * | 2023-11-06 | 2024-02-20 | 深圳市腾进达信息技术有限公司 | Method for controlling operation of intelligent wearable device based on motion capture technology |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103543826A (en) | Method for recognizing gesture based on acceleration sensor | |
| WO2018040757A1 (en) | Wearable device and method of using same to monitor motion state | |
| CN105976809B (en) | Identification method and system based on speech and facial expression bimodal emotion fusion | |
| CN105184325B (en) | Mobile intelligent terminal | |
| US20170357848A1 (en) | Method for recognizing a human motion, method for recognizing a user action and smart terminal | |
| US10061389B2 (en) | Gesture recognition system and gesture recognition method | |
| CN105912142B (en) | A kind of note step and Activity recognition method based on acceleration sensor | |
| CN103413113A (en) | Intelligent emotional interaction method for service robot | |
| CN106919958B (en) | Human body finger action recognition method based on smart watch | |
| CN108703824B (en) | Bionic hand control system and control method based on myoelectricity bracelet | |
| CN105919591A (en) | Surface myoelectrical signal based sign language recognition vocal system and method | |
| CN107122704A (en) | A kind of gait recognition method based on motion sensor | |
| CN108985157A (en) | A kind of gesture identification method and device | |
| CN105989836A (en) | Voice acquisition method, device and terminal equipment | |
| CN109840480B (en) | Interaction method and interaction system of smart watch | |
| CN105069745A (en) | face-changing system based on common image sensor and enhanced augmented reality technology and method | |
| CN116343284A (en) | Attention mechanism-based multi-feature outdoor environment emotion recognition method | |
| CN112587901A (en) | Swimming gesture recognition method, device, system and storage medium | |
| CN108108015A (en) | A kind of action gesture recognition methods based on mobile phone gyroscope and dynamic time warping | |
| CN111973177A (en) | Sleep assisting system and method based on portable electroencephalogram equipment | |
| CN116400831B (en) | Handwriting pressure induction compensation algorithm of passive electromagnetic touch screen | |
| CN111291804A (en) | Multi-sensor time series analysis model based on attention mechanism | |
| CN110727349A (en) | Man-machine interaction method and AR glasses based on bone conduction interaction | |
| CN113724838B (en) | Emotion identification system based on big data | |
| CN110751060B (en) | Portable motion mode real-time identification system based on multi-source signals |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20140129 |
|
| WD01 | Invention patent application deemed withdrawn after publication |