CN107909060A - Gymnasium body-building action identification method and device based on deep learning - Google Patents
Gymnasium body-building action identification method and device based on deep learning Download PDFInfo
- Publication number
- CN107909060A CN107909060A CN201711264432.3A CN201711264432A CN107909060A CN 107909060 A CN107909060 A CN 107909060A CN 201711264432 A CN201711264432 A CN 201711264432A CN 107909060 A CN107909060 A CN 107909060A
- Authority
- CN
- China
- Prior art keywords
- action
- deep learning
- user
- motion
- identification method
- 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
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
Gymnasium body-building action identification method based on deep learning, using real-time video as medium, carries out action norm identification to it using based on the method for deep learning, includes the following steps：(1) data acquisition, is specification motion images by specification action record；(2) data mark, and the specification classification of motion is carried out to specification motion images；(3) data are trained, and using the object detection identification framework based on convolutional neural networks in deep learning method Caffe, obtain specification action recognition model；(4) action recognition, shoots user, identifies and the specification classification of motion is consistent in specification action recognition model user action；(5) action scoring, exports specification similarity score and amendment scheme.User's wearing is not required, replaces expensive equipment, cost is controllable；Using video popular and easy to use in intelligent terminal at present as medium, accurate exercise data is generated to user, malfunction is corrected, reaches more preferable body-building effect.
Description
Technical field
The present invention relates to image identification technical field, the gymnasium body-building action recognition side specifically based on deep learning
Method and device.
Background technology
The normalization of action is even more important in body-building, how could correctly take exercise just heavy into one to closing
The problem of wanting.Nonstandard action will cause body-building to be got half the result with twice the effort, and make body-building effect with expected asymmetry, in addition not
The risk acted there are injury gained in sports of specification, and it is also a larger expense for many people to engage fitness.
In order to solve the problems, such as this, more high-end body-building system begins attempt to action recognition, but is that use can be worn mostly
Wear equipment to be aided in, such as motion bracelet, shoes, clothes, armlet, glasses, mobile phone etc. compare, it is necessary to user is actively engaged in
It is cumbersome.Also other non-video sensors are installed on body-building equipment, and coupling is strong, and cost is high.And the body-building of mainstream at present
Room is all traditional equipment mostly, inevitable expensive to be substituted for the said equipment, the budget of remote super ordinary user.
At the same time, a research hotspot of the Human bodys' response as intelligent video analysis field.From 2006 plus take
BP has been broken in the improvement of big propositions and model training method of the University of Toronto professor GeoffreyHinton to deep learning
The bottleneck of neutral net development starts, and deep learning becomes the most popular research direction in machine learning field, and intelligent sound is known
Not, constantly there is new breakthrough in the field such as instant translation, image recognition.
Therefore, applicant is directed to the action norm of body building, has carried out the gymnasium body-building action based on deep learning
The research of identification technology.
The content of the invention
Based on above-mentioned purpose, the present invention provides a kind of gymnasium body-building action identification method based on deep learning and
Device, using video popular and easy to use in intelligent terminal at present as medium, using based on the method for deep learning into
Normative identification is made in action.
Wherein, the gymnasium body-building action identification method of the invention based on deep learning employs following technical side
Case.
Gymnasium body-building action identification method based on deep learning, using real-time video as medium, using based on depth
The method of study carries out action norm identification to it, includes the following steps：
(1) data acquisition, shoots demonstration movement, is specification motion images by specification action record；
(2) data mark, and specification action point is carried out to the specification motion images that step (1) obtains by way of mark
Class；
(3) data are trained, and frame is identified using the object detection based on convolutional neural networks in deep learning method Caffe
Frame, the specification motion images carried out to step (2) after the specification classification of motion are identified training, obtain specification action recognition mould
Type；
(4) action recognition, shoots user, and the specification action recognition model obtained by step (3), to reality
When video in user images scan for, identify and the specification classification of motion is consistent in specification action recognition model user
Action；
(5) action scoring, the user action identified for step (4), by specification action recognition model by itself and phase
Specification motion images comparison operation in the specification classification of motion answered, and export specification similarity score and amendment scheme.
Wherein, the gymnasium body-building action recognition device of the invention based on deep learning, be used for realization it is described based on
The gymnasium body-building action identification method of deep learning, including：
Camera or optical sensor, for the gymnasium body-building action identification method step based on deep learning
Suddenly demonstration movement is shot in (1), and its user is shot in step (4)；
Processing module and processing center, for the gymnasium body-building action identification method step based on deep learning
Suddenly in (3) specification motion images are identified with training and obtains specification action recognition model, and for identification in step (4)
User action, is additionally operable to be compared user action computing in step (5) and exports specification similarity score；
Cloud platform, is obtained for storing in the gymnasium body-building action identification method step (5) based on deep learning
The specification similarity score and amendment scheme obtained；
Intelligent terminal, for downloading the specification similarity score and amendment scheme from the cloud platform.
Compared with prior art, the beneficial effects of the present invention are：
User's wearing is not required, replaces expensive equipment, cost is controllable；With at present in intelligent terminal it is popular and conveniently
The video used carries out action norm identification as medium, using the method based on deep learning, identification model editable,
And accurate body-building action recognition can be provided as a result, generating accurate exercise data to user, correct malfunction, reduced strong
Experienced tedious work is taught others by his own example, or even substitutes fitness, reaches more preferable body-building effect.
With reference to explanation the drawings and specific embodiments, the present invention is described further.
Brief description of the drawings
Fig. 1 is the structure diagram of the gymnasium body-building action recognition device based on deep learning of the present invention.
Embodiment
Present invention firstly provides a kind of gymnasium body-building action identification method based on deep learning, with real-time video
As medium, action norm identification is carried out to it using based on the method for deep learning, is included the following steps：
(1) data acquisition, shoots demonstration movement, is specification motion images by specification action record；
In this step, affiliated movement is judged with motion characteristic according to the human posture that specification motion images are extracted
Project, such as：Running, dumbbell etc., and it is stored as different feature templates library files respectively by sports events；
(2) data mark, and specification action point is carried out to the specification motion images that step (1) obtains by way of mark
Class；
In this step, it is to be operated for feature templates library file；
(3) data are trained, and frame is identified using the object detection based on convolutional neural networks in deep learning method Caffe
Frame, the specification motion images carried out to step (2) after the specification classification of motion are identified training, obtain specification action recognition mould
Type；
(4) action recognition, shoots user, and the specification action recognition model obtained by step (3), to reality
When video in user images scan for, identify and the specification classification of motion is consistent in specification action recognition model user
Action；
(5) action scoring, the user action identified for step (4), by specification action recognition model by itself and phase
Specification motion images comparison operation in the specification classification of motion answered, and export specification similarity score and amendment scheme.
Further, the step (1), the mode of shooting is to including but not limited to the bodies such as height, gender, figure
The different several professional fitnesses of body characteristics carry out multi-angled shooting.
Further, the step (2), the label parameters of the specification classification of motion include but not limited to head in image,
The body parts such as face, palm, upper arm, underarm, trunk, thigh, shank, foot carry out color mark or frame choosing mark.
Further, the step (3), specification action recognition model learn and remember including but not limited to body
Whether position relative position in the picture, move, the specification motion value such as mobile range, translational speed, travel frequency.
Further, the step (4) according to specification action recognition model in user action include but not limited to body
Body region relative position in the picture, whether move, the user movement value such as mobile range, translational speed, travel frequency carries out
Identification.
Further, specification motion value of the step (5) in specification action recognition model sets different numerical value
The threshold value of deviation, is used as output specification similarity score and amendment side by the comparison result of user's motion value and threshold value
The reference data of case.
Further, the step (2), the label parameters of the specification classification of motion include but not limited to height, gender,
The physical traits such as figure and shooting angle.
Further, the step (3), divide specification motion images by the label parameters of the specification classification of motion
Group training, obtains corresponding specification action recognition model packet.
Further, the step (4), using physical trait pair corresponding with the label parameters of the specification classification of motion
User carries out customer parameter definition, and is grouped using specification action recognition model immediate with the user's parameter to regarding in real time
User images in frequency are identified.
It is to a kind of detailed process that the user images in real-time video are identified：Background Recognition is carried out first, by people
Body carries out realizing Division identification with scene background；Then segmentation identification is carried out to human body according to body part, then code requirement
Action recognition model carries out each body part action parameter statistics and action matches.
By taking swing arm as an example, user images are directly inputted to trained specification action recognition model, if useful in image
The human body figure at family, specification action recognition model can Direct Recognition go out the body part of user position in the picture and class
Not, the gender for judging user, age face can also be identified, so as to match the packet of specification action recognition model for it；
Then identify body part change in location situation in the picture, according to change in location situation directly judge its whether swing arm,
The data such as swing arm amplitude size, speed are how many, frequency is how many, by it with the specification motion value in specification action recognition model and
Threshold value compares, and the corresponding corresponding explanation lack of standardization of each threshold value and fraction, if small Mr. Yu's thresholding, output corresponds to
The amendment scheme of correcting method lack of standardization and explanation, and reciprocal fraction is deducted, it is final finally to meet that the fraction of thresholding is added
Specification similarity score.
Specific identification process carries out computing by computer, and it is as follows to be directed to main algorithm：
Step 1：Algorithm processing module (or GPU work stations) obtains current camera video stream, and to each frame picture
It is identified；
Step 2：Whether the result parameter of identification data training before, judgement currently have users and (are denoted as Di)
In the presence of, if it is pure background (being denoted as C), if C, then nobody is in silent status Static in movement, if with the presence of Di,
Then have users, then export rectangle position or other shapes position of the user in picture, which includes people
Body, but background is eliminated most possibly.
Step 3：The each position key point position acquisition of human body (such as wrist, the knee of predefined when will above train
Joint, elbow joint), it is Xi, Yi that these, which press position of the XY coordinate systems in figure, and wherein i is the index at each position, is sat herein
Mark system is using original picture in its entirety upper left corner starting point as origin.
Step 4：Rectangle position (Xf1, Yf1), (Xf2, Yf2), (Xf3, Yf3), (Xf4, the Yf4) of face are exported, and
Database is searched for, comparison database face (can use template matches, can also use Hong Kong Chinese University's DeepI models, also may be used
To be contrasted with characteristic point), so as to obtain the information such as the corresponding user's gender of recognition of face, age.
Step 5：In video streaming tracking step three obtain each position key point position and when the current frame between, one
In section time T, each position key point time series (Xi, Yi, Ti) can be obtained.
Step 6：Combination between key point can form human body module, such as wrist and elbow joint line,
Line slope k is (XWrist-XElbow joint)/(YWrist-YElbow joint), then line is
(y-YWrist)=k (x-XWrist), wherein x={ XWrist, XElbow joint}
Wherein y={ yWrist, yElbow joint, the line segment can obtain forearm current data after combination.
Step 7：By key point change in location, the parameter of swing arm amplitude, speed and frequency can be obtained, such as wrist is left
There are N frame pictures in elbow joint in regular hour T, and initial shifts number is K=0, then judge (X wrists, t1) whether etc.
Whether it is equal to (Y wrists, t2) in (X wrists, t2) and (Y wrists, t1), thinks that wrist is not moved if condition is met, if being unsatisfactory for
Then wrist changes K=K+1, final frequency Fr=K/T to condition；For swing arm amplitude, between two time points, left hand wrist
Point with left hand elbow joint point line angle be k1, k2, then k1 and k2 be utilized respectively antitrigonometric function acquisition angle be
Angle1, angle2, then amplitude is abs (anglel-angle2)；Due to image and actually having certain difference, so with correcting
Parameter wa is corrected, and wa is obtained by observation and experience, can be any real number, and correction amplitude is
A=wa*abs (angle1-angle2)
Or it is expressed as with wrist distance change
A wrists=wf*sqrt (((X wrists, t1)-(X wrists, t2)) ^2+ ((Y wrists, t1)-(Y wrists, t2)) ^2)
Step 9：Action matching, simplest method is to obtain each position key point positions of human body N, and lane database is deposited
There is the standard key of coach or predefined point position, judge the distance between identical key point position difference
Di=sqrt ((Xi-X templates) ^2+ (Yi-Y templates) ^2)
If Di<TiTi is current key point positional distance thresholding, then it is assumed that the node meets action norm, otherwise the section
Point does not meet action norm, and correcting suggestion by the specification of Input of Data provides scoring；
During the motion, swing arm amplitude can also match, when step 8 identifies that two moment k1 and k2 is changed,
That is k1！=k2, then it is assumed that, in swing arm, T1 moment key point is (x1, y1, t1) for it, and T2 moment key point is (x2, y2, t1),
According to
D=sqrt (((X1, t1)-(X2, t2)) ^2+ ((Y2, t1)-(Y2, t2)) ^2)
If D<T then thinks that key point position is correct during swing arm, otherwise incorrect, or judges that amplitude A is correct with database
Whether template A absolute differences meet to be less than thresholding, meet then correct, are unsatisfactory for, export lack of standardization.
As shown in Figure 1, the present invention is also carried according to the above-mentioned gymnasium body-building action identification method based on deep learning
A kind of gymnasium body-building action recognition device based on deep learning has been supplied, has been used for realization being good for based on deep learning
Body room body-building action identification method, including：
Camera or optical sensor 1, for the gymnasium body-building action identification method step based on deep learning
Suddenly demonstration movement is shot in (1), and its user is shot in step (4)；
Processing module and processing center 2, for the gymnasium body-building action identification method step based on deep learning
Suddenly in (3) specification motion images are identified with training and obtains specification action recognition model, and for identification in step (4)
User action, is additionally operable to be compared user action computing in step (5) and exports specification similarity score；
Cloud platform 3, for storing in the gymnasium body-building action identification method step (5) based on deep learning
The specification similarity score and amendment scheme of acquisition；
Intelligent terminal 4, for downloading the specification similarity score and amendment scheme from the cloud platform.
For those skilled in the art, other various phases can be obtained according to revealed structure and principle is invented
The change and deformation answered, and all these change and deformation belongs to protection category of the invention.
Claims (10)
1. the gymnasium body-building action identification method based on deep learning, it is characterized in that, using real-time video as medium, using base
Action norm identification is carried out to it in the method for deep learning, is included the following steps：
(1) data acquisition, shoots demonstration movement, is specification motion images by specification action record；
(2) data mark, and the specification classification of motion is carried out to the specification motion images that step (1) obtains by way of mark；
(3) data are trained, right using the object detection identification framework based on convolutional neural networks in deep learning method Caffe
Training is identified in the specification motion images that step (2) carries out after the specification classification of motion, obtains specification action recognition model；
(4) action recognition, shoots user, and the specification action recognition model obtained by step (3), to real-time video
In user images scan for, identify and the specification classification of motion is consistent in specification action recognition model user action；
(5) action scoring, the user action identified for step (4), by specification action recognition model by its with it is corresponding
Specification motion images comparison operation in the specification classification of motion, and export specification similarity score and amendment scheme.
2. the gymnasium body-building action identification method according to claim 1 based on deep learning, it is characterized in that, it is described
Step (1), the mode of shooting is the several specialties different to including but not limited to the physical traits such as height, gender, figure
Fitness carries out multi-angled shooting.
3. the gymnasium body-building action identification method according to claim 2 based on deep learning, it is characterized in that, it is described
Step (2), the label parameters of the specification classification of motion include but not limited to head in image, face, palm, upper arm, underarm, trunk,
The body parts such as thigh, shank, foot carry out color mark or frame choosing mark.
4. the gymnasium body-building action identification method according to claim 3 based on deep learning, it is characterized in that, it is described
Step (3), specification action recognition model learns and remembers including but not limited to body part relative position in the picture, whether
The specification motion value such as movement, mobile range, translational speed, travel frequency.
5. the gymnasium body-building action identification method according to claim 4 based on deep learning, it is characterized in that, it is described
Step (4) is according to specification action recognition model to including but not limited to the opposite position of body part in the picture in user action
Put, whether move, the user movement value such as mobile range, translational speed, travel frequency is identified.
6. the gymnasium body-building action identification method according to claim 5 based on deep learning, it is characterized in that, it is described
Specification motion value of the step (5) in specification action recognition model sets the threshold value that different numerical value deviate, and is transported by user
Reference data of the comparison result of dynamic value and threshold value as output specification similarity score and amendment scheme.
7. the gymnasium body-building action identification method according to claim 2 based on deep learning, it is characterized in that, it is described
Step (2), the label parameters of the specification classification of motion include but not limited to the physical traits such as height, gender, figure and shooting angle
Degree.
8. the gymnasium body-building action identification method based on deep learning according to claim 3 or 4 or 5 or 6 or 7, its
It is characterized in, the step (3), training is grouped to specification motion images by the label parameters of the specification classification of motion, is obtained
Corresponding specification action recognition model packet.
9. the gymnasium body-building action identification method according to claim 8 based on deep learning, it is characterized in that, it is described
Step (4), customer parameter definition is carried out using physical trait corresponding with the label parameters of the specification classification of motion to user, and is adopted
The user images in real-time video are identified with specification action recognition model immediate with the user's parameter packet.
10. the gymnasium body-building action recognition device based on deep learning, is used for realization claim 1 to 9 any one of them
Gymnasium body-building action identification method based on deep learning, it is characterized in that, including：
Camera or optical sensor, for the gymnasium body-building action identification method step (1) based on deep learning
In demonstration movement is shot, and its user is shot in step (4)；
Processing module and processing center, for the gymnasium body-building action identification method step (3) based on deep learning
In specification motion images be identified with training obtain specification action recognition model, and moved for identification user in step (4)
Make, be additionally operable to be compared user action computing in step (5) and export specification similarity score；
Cloud platform, for storing what is obtained in the gymnasium body-building action identification method step (5) based on deep learning
Specification similarity score and amendment scheme；
Intelligent terminal, for downloading the specification similarity score and amendment scheme from the cloud platform.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711264432.3A CN107909060A (en) | 2017-12-05 | 2017-12-05 | Gymnasium body-building action identification method and device based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711264432.3A CN107909060A (en) | 2017-12-05 | 2017-12-05 | Gymnasium body-building action identification method and device based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107909060A true CN107909060A (en) | 2018-04-13 |
Family
ID=61854296
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711264432.3A Pending CN107909060A (en) | 2017-12-05 | 2017-12-05 | Gymnasium body-building action identification method and device based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107909060A (en) |
Cited By (38)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108596067A (en) * | 2018-04-15 | 2018-09-28 | 中少科普（北京）教育科技有限公司 | A kind of Young Pioneer's salute detection bearing calibration |
| CN108635671A (en) * | 2018-07-06 | 2018-10-12 | 谭希妤 | A kind of artificial intelligence AED automated external defibrillator system and method |
| CN108734104A (en) * | 2018-04-20 | 2018-11-02 | 杭州易舞科技有限公司 | Body-building action error correction method based on deep learning image recognition and system |
| CN108769391A (en) * | 2018-05-14 | 2018-11-06 | 谭诗涵 | The anti-sudden death artificial intelligence life monitoring CPR mobile phone of one kind and method |
| CN108805109A (en) * | 2018-08-07 | 2018-11-13 | 深圳市云康创新网络科技有限公司 | A kind of exercise data capture display system |
| CN108921127A (en) * | 2018-07-19 | 2018-11-30 | 上海小蚁科技有限公司 | Method for testing motion and device, storage medium, terminal |
| CN109034776A (en) * | 2018-07-15 | 2018-12-18 | 谭诗涵 | A kind of financial payment systems and method of the physician guidance first aid of false alarm prevention wechat video |
| CN109011419A (en) * | 2018-08-22 | 2018-12-18 | 重庆电政信息科技有限公司 | A kind of athletic performance training method based on MEMS sensor |
| CN109191588A (en) * | 2018-08-27 | 2019-01-11 | 百度在线网络技术（北京）有限公司 | Move teaching method, device, storage medium and electronic equipment |
| CN109241909A (en) * | 2018-09-06 | 2019-01-18 | 闫维新 | A kind of long-range dance movement capture evaluating system based on intelligent terminal |
| CN109344922A (en) * | 2018-09-06 | 2019-02-15 | 闫维新 | A kind of dance movement evaluating system having motion-captured function |
| CN109409199A (en) * | 2018-08-31 | 2019-03-01 | 百度在线网络技术（北京）有限公司 | Micro- expression training method, device, storage medium and electronic equipment |
| CN109446952A (en) * | 2018-10-16 | 2019-03-08 | 赵笑婷 | A kind of piano measure of supervision, device, computer equipment and storage medium |
| CN109472217A (en) * | 2018-10-19 | 2019-03-15 | 广州慧睿思通信息科技有限公司 | Intelligent training model building method and device, training method and device |
| CN109840478A (en) * | 2019-01-04 | 2019-06-04 | 广东智媒云图科技股份有限公司 | A kind of movement appraisal procedure, device, mobile terminal and readable storage medium storing program for executing |
| CN109919034A (en) * | 2019-01-31 | 2019-06-21 | 厦门大学 | A kind of identification of limb action with correct auxiliary training system and method |
| CN110018646A (en) * | 2019-04-19 | 2019-07-16 | 北京潞电电气设备有限公司 | A kind of power equipment working specification monitoring system |
| CN110070036A (en) * | 2019-04-22 | 2019-07-30 | 北京迈格威科技有限公司 | The method, apparatus and electronic equipment of synkinesia action training |
| CN110135289A (en) * | 2019-04-28 | 2019-08-16 | 北京天地玛珂电液控制系统有限公司 | A kind of underground coal mine intelligent use cloud service platform based on deep learning |
| CN110222977A (en) * | 2019-06-03 | 2019-09-10 | 张学志 | One kind movement sport methods of marking based on computer vision and device |
| CN110321776A (en) * | 2019-04-17 | 2019-10-11 | 郑州轻工业学院 | Methods of Machine Recognition of Faces based on deep learning |
| CN110472613A (en) * | 2019-08-22 | 2019-11-19 | 海信集团有限公司 | Object behavior recognition methods and device |
| CN110610173A (en) * | 2019-10-16 | 2019-12-24 | 电子科技大学 | Badminton motion analysis system and method based on Mobilenet |
| CN110841262A (en) * | 2019-12-06 | 2020-02-28 | 郑州大学体育学院 | Football training system based on wearable equipment |
| CN111259716A (en) * | 2019-10-17 | 2020-06-09 | 浙江工业大学 | Human body running posture identification and analysis method and device based on computer vision |
| CN111553218A (en) * | 2020-04-20 | 2020-08-18 | 南京医科大学 | Intelligent medical skill teaching monitoring system based on human body posture recognition |
| CN111680608A (en) * | 2020-06-03 | 2020-09-18 | 长春博立电子科技有限公司 | Intelligent sports auxiliary training system and training method based on video analysis |
| CN111787375A (en) * | 2020-04-30 | 2020-10-16 | 聚好看科技股份有限公司 | Display device and information display method |
| CN112004159A (en) * | 2020-08-28 | 2020-11-27 | 深圳创维－Rgb电子有限公司 | Video data processing method, device, system and storage medium |
| CN112040301A (en) * | 2020-09-14 | 2020-12-04 | 成都拟合未来科技有限公司 | Interactive exercise equipment action explanation method, system, terminal and medium |
| CN112257568A (en) * | 2020-10-21 | 2021-01-22 | 中国人民解放军国防科技大学 | Intelligent real-time supervision and error correction system and method for individual soldier queue actions |
| CN112272324A (en) * | 2020-10-15 | 2021-01-26 | 聚好看科技股份有限公司 | Follow-up mode control method and display device |
| CN112399234A (en) * | 2019-08-18 | 2021-02-23 | 聚好看科技股份有限公司 | Interface display method and display equipment |
| CN112464882A (en) * | 2020-12-11 | 2021-03-09 | 东莞先知大数据有限公司 | Method, apparatus, medium, and device for recognizing continuous motion |
| CN112766147A (en) * | 2021-01-16 | 2021-05-07 | 大连理工大学 | Error action positioning method based on deep learning |
| CN113128283A (en) * | 2019-12-31 | 2021-07-16 | 沸腾时刻智能科技(深圳)有限公司 | Evaluation method, model construction method, teaching machine, teaching system and electronic equipment |
| CN114330575A (en) * | 2021-12-31 | 2022-04-12 | 深圳千岸科技股份有限公司 | Body-building action discrimination method based on twin neural network and intelligent sand bag |
| US11924513B2 (en) | 2019-08-18 | 2024-03-05 | Juhaokan Technology Co., Ltd. | Display apparatus and method for display user interface |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102368297A (en) * | 2011-09-14 | 2012-03-07 | 北京英福生科技有限公司 | Equipment, system and method for recognizing actions of detected object |
| CN105903157A (en) * | 2016-04-19 | 2016-08-31 | 深圳泰山体育科技股份有限公司 | Electronic coach realization method and system |
| CN106055873A (en) * | 2016-05-20 | 2016-10-26 | 北京旷视科技有限公司 | Fitness auxiliary method and apparatus based on image recognition |
| CN107220608A (en) * | 2017-05-22 | 2017-09-29 | 华南理工大学 | What a kind of basketball action model was rebuild and defended instructs system and method |
| CN107291232A (en) * | 2017-06-20 | 2017-10-24 | 深圳市泽科科技有限公司 | A kind of somatic sensation television game exchange method and system based on deep learning and big data |
-
2017
- 2017-12-05 CN CN201711264432.3A patent/CN107909060A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102368297A (en) * | 2011-09-14 | 2012-03-07 | 北京英福生科技有限公司 | Equipment, system and method for recognizing actions of detected object |
| CN105903157A (en) * | 2016-04-19 | 2016-08-31 | 深圳泰山体育科技股份有限公司 | Electronic coach realization method and system |
| CN106055873A (en) * | 2016-05-20 | 2016-10-26 | 北京旷视科技有限公司 | Fitness auxiliary method and apparatus based on image recognition |
| CN107220608A (en) * | 2017-05-22 | 2017-09-29 | 华南理工大学 | What a kind of basketball action model was rebuild and defended instructs system and method |
| CN107291232A (en) * | 2017-06-20 | 2017-10-24 | 深圳市泽科科技有限公司 | A kind of somatic sensation television game exchange method and system based on deep learning and big data |
Cited By (48)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108596067A (en) * | 2018-04-15 | 2018-09-28 | 中少科普（北京）教育科技有限公司 | A kind of Young Pioneer's salute detection bearing calibration |
| CN108596067B (en) * | 2018-04-15 | 2019-09-10 | 中少科普（北京）教育科技有限公司 | A kind of Young Pioneer's salute detection bearing calibration |
| CN108734104A (en) * | 2018-04-20 | 2018-11-02 | 杭州易舞科技有限公司 | Body-building action error correction method based on deep learning image recognition and system |
| CN108769391A (en) * | 2018-05-14 | 2018-11-06 | 谭诗涵 | The anti-sudden death artificial intelligence life monitoring CPR mobile phone of one kind and method |
| CN108635671A (en) * | 2018-07-06 | 2018-10-12 | 谭希妤 | A kind of artificial intelligence AED automated external defibrillator system and method |
| CN109034776A (en) * | 2018-07-15 | 2018-12-18 | 谭诗涵 | A kind of financial payment systems and method of the physician guidance first aid of false alarm prevention wechat video |
| CN108921127A (en) * | 2018-07-19 | 2018-11-30 | 上海小蚁科技有限公司 | Method for testing motion and device, storage medium, terminal |
| CN108805109A (en) * | 2018-08-07 | 2018-11-13 | 深圳市云康创新网络科技有限公司 | A kind of exercise data capture display system |
| CN108805109B (en) * | 2018-08-07 | 2024-04-23 | 深圳市天玖隆科技有限公司 | Motion data capturing and displaying system |
| CN109011419A (en) * | 2018-08-22 | 2018-12-18 | 重庆电政信息科技有限公司 | A kind of athletic performance training method based on MEMS sensor |
| CN109191588A (en) * | 2018-08-27 | 2019-01-11 | 百度在线网络技术（北京）有限公司 | Move teaching method, device, storage medium and electronic equipment |
| CN109191588B (en) * | 2018-08-27 | 2020-04-07 | 百度在线网络技术（北京）有限公司 | Motion teaching method, motion teaching device, storage medium and electronic equipment |
| CN109409199A (en) * | 2018-08-31 | 2019-03-01 | 百度在线网络技术（北京）有限公司 | Micro- expression training method, device, storage medium and electronic equipment |
| CN109409199B (en) * | 2018-08-31 | 2021-01-12 | 百度在线网络技术（北京）有限公司 | Micro-expression training method and device, storage medium and electronic equipment |
| CN109344922A (en) * | 2018-09-06 | 2019-02-15 | 闫维新 | A kind of dance movement evaluating system having motion-captured function |
| CN109241909A (en) * | 2018-09-06 | 2019-01-18 | 闫维新 | A kind of long-range dance movement capture evaluating system based on intelligent terminal |
| CN109446952A (en) * | 2018-10-16 | 2019-03-08 | 赵笑婷 | A kind of piano measure of supervision, device, computer equipment and storage medium |
| CN109472217B (en) * | 2018-10-19 | 2021-08-31 | 广州慧睿思通信息科技有限公司 | Intelligent exercise training model construction method and device and training method and device |
| CN109472217A (en) * | 2018-10-19 | 2019-03-15 | 广州慧睿思通信息科技有限公司 | Intelligent training model building method and device, training method and device |
| CN109840478A (en) * | 2019-01-04 | 2019-06-04 | 广东智媒云图科技股份有限公司 | A kind of movement appraisal procedure, device, mobile terminal and readable storage medium storing program for executing |
| CN109919034A (en) * | 2019-01-31 | 2019-06-21 | 厦门大学 | A kind of identification of limb action with correct auxiliary training system and method |
| CN110321776A (en) * | 2019-04-17 | 2019-10-11 | 郑州轻工业学院 | Methods of Machine Recognition of Faces based on deep learning |
| CN110018646A (en) * | 2019-04-19 | 2019-07-16 | 北京潞电电气设备有限公司 | A kind of power equipment working specification monitoring system |
| CN110070036A (en) * | 2019-04-22 | 2019-07-30 | 北京迈格威科技有限公司 | The method, apparatus and electronic equipment of synkinesia action training |
| CN110135289A (en) * | 2019-04-28 | 2019-08-16 | 北京天地玛珂电液控制系统有限公司 | A kind of underground coal mine intelligent use cloud service platform based on deep learning |
| CN110222977A (en) * | 2019-06-03 | 2019-09-10 | 张学志 | One kind movement sport methods of marking based on computer vision and device |
| WO2021031809A1 (en) * | 2019-08-18 | 2021-02-25 | 聚好看科技股份有限公司 | Interface display method and display device |
| CN112399234A (en) * | 2019-08-18 | 2021-02-23 | 聚好看科技股份有限公司 | Interface display method and display equipment |
| US11924513B2 (en) | 2019-08-18 | 2024-03-05 | Juhaokan Technology Co., Ltd. | Display apparatus and method for display user interface |
| CN110472613A (en) * | 2019-08-22 | 2019-11-19 | 海信集团有限公司 | Object behavior recognition methods and device |
| CN110472613B (en) * | 2019-08-22 | 2022-05-10 | 海信集团有限公司 | Object behavior identification method and device |
| CN110610173A (en) * | 2019-10-16 | 2019-12-24 | 电子科技大学 | Badminton motion analysis system and method based on Mobilenet |
| CN111259716A (en) * | 2019-10-17 | 2020-06-09 | 浙江工业大学 | Human body running posture identification and analysis method and device based on computer vision |
| CN110841262A (en) * | 2019-12-06 | 2020-02-28 | 郑州大学体育学院 | Football training system based on wearable equipment |
| CN113128283A (en) * | 2019-12-31 | 2021-07-16 | 沸腾时刻智能科技(深圳)有限公司 | Evaluation method, model construction method, teaching machine, teaching system and electronic equipment |
| CN111553218A (en) * | 2020-04-20 | 2020-08-18 | 南京医科大学 | Intelligent medical skill teaching monitoring system based on human body posture recognition |
| CN111787375A (en) * | 2020-04-30 | 2020-10-16 | 聚好看科技股份有限公司 | Display device and information display method |
| CN111680608B (en) * | 2020-06-03 | 2023-08-18 | 长春博立电子科技有限公司 | Intelligent sports auxiliary training system and training method based on video analysis |
| CN111680608A (en) * | 2020-06-03 | 2020-09-18 | 长春博立电子科技有限公司 | Intelligent sports auxiliary training system and training method based on video analysis |
| CN112004159A (en) * | 2020-08-28 | 2020-11-27 | 深圳创维－Rgb电子有限公司 | Video data processing method, device, system and storage medium |
| CN112040301A (en) * | 2020-09-14 | 2020-12-04 | 成都拟合未来科技有限公司 | Interactive exercise equipment action explanation method, system, terminal and medium |
| CN112272324A (en) * | 2020-10-15 | 2021-01-26 | 聚好看科技股份有限公司 | Follow-up mode control method and display device |
| CN112272324B (en) * | 2020-10-15 | 2023-03-14 | 聚好看科技股份有限公司 | Follow-up mode control method and display device |
| CN112257568A (en) * | 2020-10-21 | 2021-01-22 | 中国人民解放军国防科技大学 | Intelligent real-time supervision and error correction system and method for individual soldier queue actions |
| CN112464882A (en) * | 2020-12-11 | 2021-03-09 | 东莞先知大数据有限公司 | Method, apparatus, medium, and device for recognizing continuous motion |
| CN112766147A (en) * | 2021-01-16 | 2021-05-07 | 大连理工大学 | Error action positioning method based on deep learning |
| CN112766147B (en) * | 2021-01-16 | 2022-10-14 | 大连理工大学 | Error action positioning method based on deep learning |
| CN114330575A (en) * | 2021-12-31 | 2022-04-12 | 深圳千岸科技股份有限公司 | Body-building action discrimination method based on twin neural network and intelligent sand bag |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107909060A (en) | Gymnasium body-building action identification method and device based on deep learning | |
| CN106650687B (en) | Posture correction method based on depth information and skeleton information | |
| Qiao et al. | Real-time human gesture grading based on OpenPose | |
| CN110705390A (en) | Body posture recognition method and device based on LSTM and storage medium | |
| Trejo et al. | Recognition of yoga poses through an interactive system with kinect device | |
| Chaudhari et al. | Yog-guru: Real-time yoga pose correction system using deep learning methods | |
| CN113255522B (en) | Personalized motion attitude estimation and analysis method and system based on time consistency | |
| CN110428486B (en) | Virtual interaction fitness method, electronic equipment and storage medium | |
| CN111191599A (en) | Gesture recognition method, device, equipment and storage medium | |
| CN110490109B (en) | Monocular vision-based online human body rehabilitation action recognition method | |
| CN109815776A (en) | Action prompt method and apparatus, storage medium and electronic device | |
| CN115131879B (en) | Action evaluation method and device | |
| CN106815855A (en) | Based on the human body motion tracking method that production and discriminate combine | |
| CN107273857B (en) | Motion action recognition method and device and electronic equipment | |
| CN111383735A (en) | Unmanned body-building analysis method based on artificial intelligence | |
| Liu | Aerobics posture recognition based on neural network and sensors | |
| Yang et al. | Human exercise posture analysis based on pose estimation | |
| Amaliya et al. | Study on hand keypoint framework for sign language recognition | |
| KR102377767B1 (en) | Handwriting and arm movement learning-based sign language translation system and method | |
| Holden | Visual recognition of hand motion | |
| CN115530814A (en) | Child motion rehabilitation training method based on visual posture detection and computer deep learning | |
| Wang | Research on the evaluation of sports training effect based on artificial intelligence technology | |
| KR102324690B1 (en) | Table tennis robot system applying deep learning | |
| CN114550071A (en) | Method, device and medium for automatically identifying and capturing track and field video action key frames | |
| Rodrigues et al. | Supervised classification of motor-rehabilitation body movements with rgb cameras and pose tracking data |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180413 |