CN113255450A - Human motion rhythm comparison system and method based on attitude estimation - Google Patents
Human motion rhythm comparison system and method based on attitude estimation Download PDFInfo
- Publication number
- CN113255450A CN113255450A CN202110447115.5A CN202110447115A CN113255450A CN 113255450 A CN113255450 A CN 113255450A CN 202110447115 A CN202110447115 A CN 202110447115A CN 113255450 A CN113255450 A CN 113255450A
- Authority
- CN
- China
- Prior art keywords
- sporter
- comparison
- information sequence
- angle information
- rhythm
- 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
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
- G16H20/30—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to physical therapies or activities, e.g. physiotherapy, acupressure or exercising
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/23—Recognition of whole body movements, e.g. for sport training
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Psychiatry (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- General Engineering & Computer Science (AREA)
- Physical Education & Sports Medicine (AREA)
- Epidemiology (AREA)
- Medical Informatics (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Image Analysis (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a human motion rhythm comparison system and method based on attitude estimation, wherein the system comprises: the image acquisition module is used for acquiring the body posture image of the sporter in real time; the key point detection module is used for detecting the input motion videos of the sporter and the coach and outputting two-dimensional coordinate information sequences of joint points of the sporter and the coach; the key point information preprocessing module is used for converting the input two-dimensional coordinate information sequence of the joint point of the sporter and the two-dimensional coordinate information sequence of the joint point of the coach into an angle information sequence of the joint point; the standard action module is used for establishing a standard action template; the key frame extraction module is used for extracting key frames of the sporter and the coach and recording the time interval of the key frames for comparison; the comparison module is used for comparing the exercise information of the sporter with the exercise information of the coach.
Description
Technical Field
The invention belongs to the technical field of motion assistance, and particularly relates to a human motion rhythm comparison system and method based on attitude estimation.
Background
According to the data of the present situation of nutrition and health of Chinese residents published by the ministry of health in 2004, about 1.6 hundred million people suffer from hypertension of residents aged 18 years and older in China; the estimated number of people with overweight is 2.0 hundred million; the investigation result also shows that the physical activity is closely related to the occurrence of obesity, diabetes and the like.
Sufficient amount of exercise has a positive effect on the health of the body. A plurality of sport apps and sport videos emerge at present, and people can do targeted sports without going out of home. The download and viewing volume is high, and therefore the visible motion is also valued by people. The household sports is influenced by factors such as the field, and the effect of the sports can be influenced if the action is in place and cannot be known at the first time. With the improvement of living standard and the development of science and technology, the requirements of people on the quality of life also become high. The development of human pose estimation makes intelligent motion assistance possible.
Most of the related exercise assisting products on the market at present collect data of an exerciser by a sensor, wear professional equipment or place markers on the body to track exercise information of all parts of a human body, limit the exercise of the exerciser to a certain extent, and are difficult to be widely applied. On the other hand, in the field of computer vision research, most of the time is to photograph the motion of the athlete by a depth camera such as Kinect or photograph video by a general camera to predict the information of each skeletal key point of the athlete by a posture estimation method. Currently, there are only two techniques for obtaining the sports information of the players, but obtaining the sports information of the players is not enough to identify whether the movements of the players are wrong, and can not directly point out the critical wrong places like professional coaches.
Disclosure of Invention
For home sports, a flexible and efficient guidance evaluation system is established, so that a sporter can be helped to adjust action details in time, and the exercise efficiency is improved. Based on the defects in the prior art, the invention provides a human motion rhythm comparison system and method based on posture estimation.
The invention adopts the following technical scheme:
a human motion rhythm comparison system based on posture estimation comprises:
the rhythm comparison preparation module is used for acquiring a video, extracting human skeleton information from the video, and preprocessing and outputting the extracted information;
the rhythm comparison module is used for carrying out rhythm comparison on the input human body key point information and the coach key point information during exercise and outputting a result of whether the comparison is in accordance with the comparison result;
a rhythm comparison feedback module: the system is used for feeding back the input rhythm comparison result to a system user so as to better improve the exercise efficiency.
As a preferred solution, the rhythm comparison preparation module includes:
the acquisition submodule is used for outputting the motion videos of the sporter and the coach;
the joint point detection processing submodule is used for detecting an input motion video of a sporter and a coach, extracting joint point information of a human body in the video by adopting an Openpos algorithm, converting an extracted two-dimensional coordinate information sequence of the joint point of the sporter and a two-dimensional coordinate information sequence of the joint point of the coach into a joint point angle information sequence, filling missing values and removing and filling abnormal values of the joint point angle information sequence, filling the missing values and the abnormal values by utilizing the continuity of human motion and adopting a linear interpolation method, and outputting the joint point angle information sequences of the sporter and the coach after preprocessing;
further, the method for converting the angle of the joint point comprises the following steps: let Plelbow、Plshoulder、Plbreast、PlwristTwo-dimensional coordinates of the left wrist joint, the left shoulder joint, the left spine joint and the left elbow joint, AlshoulderAnd AlelbowRespectively a left shoulder joint angle and a left elbow joint angle. Then
The standard action submodule is used for carrying out similarity comparison processing on a plurality of input coach joint angle information sequences to obtain a new joint angle information sequence as a template, and the joint angle information sequence of the sporter is compared with the standard action template;
further, the standard action template establishing method comprises the following steps: and (3) selecting 15 groups of coach motion data, and filtering the data to obtain 15 groups of smooth motion data because data jumping does not occur due to the continuity of human motion. And selecting a group of data as an initial action template, calculating the DTW distance between the next group of data and the template, discarding if the DTW distance exceeds a threshold, remapping according to the matching path if the DTW distance does not exceed the threshold, distributing a weight of 0.6 to the template, distributing a weight of 0.4 to the next group of data, and calculating a weighted value of a corresponding point to obtain a new template. Repeating DTW distance calculation, and obtaining a standard action template after 15 groups of data are calculated.
As a preferred embodiment, the rhythm comparison module includes:
the key frame extraction submodule is used for segmenting the input joint point angle information sequence of the sporter and the complete motion angle information sequence of the standard motion template, and extracting segmentation points from extreme points in the joint angle sequence to be used as key frames;
the motion starting judgment sub-module is used for judging whether the motion of the sporter starts or not, reminding the sporter to start the motion if the motion does not start, and performing subsequent comparison if the motion starting judgment is obtained;
and the comparison submodule is used for carrying out similarity comparison rhythm comparison between the input joint angle information sequence of the sporter and the joint angle information sequence of the standard action template, comparing the obtained joint angle information sequence of the sporter with the joint angle information sequence of the standard action template when the sporter starts to exercise, and outputting a comparison result to remind.
Further, the similarity comparison is to search the sporter and the standard action template from the starting frame to the ending frame in a matching traversal manner, the searched path is represented by W, and the kth element of W is defined as Wk(i, j), an angular information sequence Y of the sporter is definedu(u is more than or equal to 1 and less than or equal to m) and key point angle information J of the coachv(v is more than or equal to 1 and less than or equal to n);
W＝w1,w2,……wk 1≤K<m+n-1；
the search path must satisfy the following constraints:
the characteristics of the bounding: starting a path from (1,1) to (m, n) and ensuring that each point on the sequences Y and J is traversed;
continuity: m and n can only be increased by 0 or 1 in sequence, namely the point after (m, n) is required to be one of (m +1, n), (m, n +1) and (m +1, n + 1);
monotonicity: the paths need to ensure that the time sequence is monotonically non-decreasing.
Further, the alignment method comprises:
the similarity comparison is used for comparing the similarity of the complete joint point angle information sequence of the sporter and the joint point information sequence of the standard action template after the sporter finishes the movement, and outputting the overall evaluation of the movement completion degree of the sporter;
and the rhythm comparison is used for comparing the time between the key frames and the key frames of the sporter with the time interval of the key frames of the standard action template, so that the advance and the delay of the comparison between the sporter and the standard action template and the speed and the slow of the action rhythm can be judged.
Further, the alignment process is as follows:
the start of the exercise is detected, and the athlete can not put action comparison in the preparation process, so the start of the exercise needs to be detected, and the first ten frames of the joint angle information of the standard action template are taken as the initial matching frame. The match between the sporter and the ten frames in the action is within the threshold value, namely the start of the movement, and then the comparison of the movement similarity is started;
detecting motion similarity, comparing the joint angle information sequence of the sporter with a standard motion template by adopting an open DTW algorithm, judging that the motion is qualified if the angle information sequence is smaller than a threshold value, then performing smoothing treatment on the information sequence to eliminate burr interference, performing rhythm comparison, and directly reminding the sporter that the motion does not accord with the standard motion if the angle information sequence is larger than the threshold value;
detecting the movement rhythm, wherein the movement action similarity accords with a standard action template to enter rhythm comparison, and the moment of a standard action key frame is recorded as TbeginT1T2...TendThe key frame time interval is recorded as Δ T1＝T1-Tbegin,ΔT2＝T2-T1...Tn＝Tn-Tn-1And reading an extreme value in the change of the action angle of the sporter, sequentially comparing the time interval with the time interval of the standard action, reminding the sporter to pay attention to the accelerated rhythm if the time interval of the sporter is greater than the time interval of the standard action, and reminding the sporter to pay attention to the slowed rhythm if the time interval of the sporter is less than the time interval of the standard action.
As a better scheme, the human motion rhythm comparison method based on posture estimation includes the following steps:
s1, obtaining the motion videos of the sporter and the coach;
s2, extracting two-dimensional coordinate information of human body joint points in the video by using an Openpos algorithm to respectively obtain
Preprocessing a two-dimensional coordinate information sequence of a joint point of a sporter and a two-dimensional coordinate information sequence of a joint point of a coach movement, converting two-dimensional coordinates of the joint point into angle information, and filling a missing value and an abnormal value by adopting a linear interpolation method;
s3, establishing a standard action template, comparing the similarity of the joint angle information sequences of a plurality of coaches by adopting a DTW algorithm, and recombining to form a new joint angle information sequence which is the standard action template;
s4, extracting key frames, namely taking frames where extreme points in the joint point angle information sequence are located as the key frames, and performing rhythm comparison on actions by dividing the complete action sequence;
s5, comparing the joint point angle information sequence of the sporter with a standard action template to judge whether the sporter starts to move or not;
s6, comparing the joint point angle information sequence of the sporter with a standard action template to judge whether the sporter movement meets the standard;
s7, comparing the joint point angle information sequence of the sporter with a standard action template to judge the movement rhythm problem of the sporter;
and S8, judging to carry out corresponding reminding according to the motion judgment data.
The comparing method according to step S6 includes:
comparing the similarity, namely comparing the similarity of the joint angle information sequence of the sporter with the joint information sequence of the standard action template by adopting an open DTW algorithm, outputting the overall evaluation of the action completion degree of the sporter, and comparing the rhythm if the action similarity is within a matched threshold value;
and rhythm comparison, which is used for comparing the time between the key frame and the key frame of the sporter with the time interval of the key frame of the standard action template, can judge the advance and delay of the comparison between the sporter and the standard action template, the action rhythm is fast and slow, the movement action similarity accords with the standard action template, the rhythm comparison is carried out, the moment of the key frame of the standard action is recorded, the time interval of the key frame is recorded, the extreme value in the movement angle change of the sporter is read, the time interval is sequentially compared with the time interval of the standard action, if the time interval of the sporter is greater than the time interval of the standard action, the sporter is reminded to pay attention to quickening the rhythm, and if the time interval of the sporter is less than the time interval of the standard action, the sporter is reminded to delay the rhythm.
According to the alignment step, the method comprises the following steps:
the start of the movement is detected, and the preparation process of the sporter cannot put action comparison, so the start of the movement needs to be detected, and the first ten frames of the joint angle information of the standard action template are taken as the initial matching frame. The match between the sporter and the ten frames in the action is within the threshold value, namely the start of the movement, and the movement comparison is started;
and detecting the motion similarity, comparing the joint angle information sequence of the sporter with a standard motion template by adopting an open DTW algorithm, judging that the motion is qualified if the joint angle information sequence is smaller than a threshold value, smoothing the information sequence to eliminate burr interference, performing rhythm comparison, and directly reminding the sporter that the motion does not accord with the standard motion if the joint angle information sequence is larger than the threshold value.
Drawings
FIG. 1 is a schematic diagram of the system configuration
FIG. 2 is a schematic block diagram of a system
FIG. 3 is a flow chart of the system architecture
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Fig. 1 is a schematic structural diagram of the present invention:
the system comprises a computer and a camera, wherein the computer comprises a standard action database of a coach and an intelligent motion comparison system.
The specific mode for realizing the household exercise assisting system is as follows: the camera is used for collecting the limb actions of the sporter in real time and is connected with the computer through the USB interface, and the collected limb actions of the sporter are transmitted to the intelligent sports comparison system through videos. A joint point detection processing submodule in the intelligent comparison system extracts joint point two-dimensional coordinate information of a sporter from a video by using an openposition algorithm, combines the joint point two-dimensional coordinate information into a joint point two-dimensional coordinate information sequence, fills missing values and removes and fills abnormal values of the joint point two-dimensional coordinate information sequence, and converts the joint point two-dimensional coordinate information sequence into a joint point angle information sequence for outputting. And the comparison submodule compares the input joint angle information sequence of the sporter with a standard action template in a coach standard action database. The standard action template is established by comparing and recombining the joint angle information sequences of a plurality of groups of coach actions. The standard action template also extracts key frames in a key frame extraction module for comparison of motion rhythms. The comparison submodule judges the motion start first, and then carries out similarity comparison when judging the motion start of the sporter, and further compares whether the rhythms are matched or not when the sporter is qualified; and if the judgment result is not yes, carrying out corresponding prompt.
As shown in fig. 2, the auxiliary system is shown in a schematic block diagram, and includes a camera, a key point detection processing module, a standard action module, a key frame extraction module, and a comparison module, where:
the camera is used for collecting the limb actions of the sporter in real time and outputting a movement video of the sporter;
the key point detection processing module is used for detecting an input motion video of a sporter and a coach, extracting joint point information of a human body in the video by adopting an Openpos algorithm, converting an extracted two-dimensional coordinate information sequence of the joint point of the sporter and a two-dimensional coordinate information sequence of the joint point of the coach into a joint point angle information sequence, filling missing values and removing and filling abnormal values of the joint point angle information sequence, filling the missing values and the abnormal values by utilizing the continuity of human motion and adopting a linear interpolation method, and outputting the joint point angle information sequences of the sporter and the coach after preprocessing;
the standard action module is used for carrying out similarity comparison processing on a plurality of input coach joint angle information sequences to obtain a new joint angle information sequence as a template, and the joint angle information sequence of the sporter is compared with the standard action template;
the key frame extraction module is used for segmenting the input joint point angle information sequence of the sporter and the complete motion angle information sequence of the standard motion template, and extracting segmentation points from extreme points in the joint angle sequence to be used as key frames;
the comparison module is used for comparing the movement of the sporter with the movement of a coach, and completing three comparison steps of initial judgment, similarity judgment and rhythm comparison.
As shown in fig. 3, the auxiliary system flowchart also provides a human motion rhythm comparison method based on posture estimation, which includes the following steps:
s1, obtaining the motion videos of the sporter and the coach;
s2, extracting two-dimensional coordinate information of human body joint points in the video by using an Openpos algorithm to respectively obtain
Preprocessing a two-dimensional coordinate information sequence of a joint point of a sporter and a two-dimensional coordinate information sequence of a joint point of a coach movement, converting two-dimensional coordinates of the joint point into angle information, and filling a missing value and an abnormal value by adopting a linear interpolation method;
s3, establishing a standard action template, comparing the similarity of the joint angle information sequences of a plurality of coaches by adopting a DTW algorithm, and recombining to form a new joint angle information sequence which is the standard action template;
s4, extracting key frames, namely taking frames where extreme points in the joint point angle information sequence are located as the key frames, and performing rhythm comparison on actions by dividing the complete action sequence;
s5, comparing the joint point angle information sequence of the sporter with a standard action template to judge whether the sporter starts to move or not;
s6, comparing the joint point angle information sequence of the sporter with a standard action template to judge whether the sporter movement meets the standard;
s7, comparing the joint point angle information sequence of the sporter with a standard action template to judge the movement rhythm problem of the sporter;
and S8, judging to carry out corresponding reminding according to the motion judgment data.
The comparing method according to step S6 includes:
comparing the similarity, namely comparing the similarity of the joint angle information sequence of the sporter with the joint information sequence of the standard action template by adopting an open DTW algorithm, outputting the overall evaluation of the action completion degree of the sporter, and comparing the rhythm if the action similarity is within a matched threshold value;
and rhythm comparison, which is used for comparing the time between the key frame and the key frame of the sporter with the time interval of the key frame of the standard action template, can judge the advance and delay of the comparison between the sporter and the standard action template, the action rhythm is fast and slow, the movement action similarity accords with the standard action template, the rhythm comparison is carried out, the moment of the key frame of the standard action is recorded, the time interval of the key frame is recorded, the extreme value in the movement angle change of the sporter is read, the time interval is sequentially compared with the time interval of the standard action, if the time interval of the sporter is greater than the time interval of the standard action, the sporter is reminded to pay attention to quickening the rhythm, and if the time interval of the sporter is less than the time interval of the standard action, the sporter is reminded to delay the rhythm.
According to the alignment step, the method comprises the following steps:
the start of the movement is detected, and the preparation process of the sporter cannot put action comparison, so the start of the movement needs to be detected, and the first ten frames of the joint angle information of the standard action template are taken as the initial matching frame. The match between the sporter and the ten frames in the action is within the threshold value, namely the start of the movement, and the movement comparison is started;
and detecting the motion similarity, comparing the joint angle information sequence of the sporter with a standard motion template by adopting an open DTW algorithm, judging that the motion is qualified if the joint angle information sequence is smaller than a threshold value, smoothing the information sequence to eliminate burr interference, performing rhythm comparison, and directly reminding the sporter that the motion does not accord with the standard motion if the joint angle information sequence is larger than the threshold value.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.
Claims (8)
1. A human motion rhythm comparison system based on posture estimation is characterized by comprising:
the rhythm comparison preparation module is used for acquiring a video, extracting human skeleton information from the video, and preprocessing and outputting the extracted information;
the rhythm comparison module is used for carrying out rhythm comparison on the input human body key point information and the coach key point information during exercise and outputting a result of whether the comparison is in accordance with the comparison result;
a rhythm comparison feedback module: the system is used for feeding back the input rhythm comparison result to a system user so as to better improve the exercise efficiency.
2. The system according to claim 1, wherein the system comprises: the rhythm comparison preparation module comprises:
the acquisition submodule is used for outputting the motion videos of the sporter and the coach;
the joint point detection processing submodule is used for detecting an input motion video of a sporter and a coach, extracting joint point information of a human body in the video by adopting an Openpos algorithm, converting an extracted two-dimensional coordinate information sequence of the joint point of the sporter and a two-dimensional coordinate information sequence of the joint point of the coach into a joint point angle information sequence, filling missing values and removing and filling abnormal values of the joint point angle information sequence, filling the missing values and the abnormal values by utilizing the continuity of human motion and adopting a linear interpolation method, and outputting the joint point angle information sequences of the sporter and the coach after preprocessing;
and the standard action submodule is used for carrying out similarity comparison processing on the input multiple groups of trainer joint point angle information sequences to obtain a new group of joint point angle information sequences as a template, and the joint point angle information sequences of the sporter are compared with the standard action template.
3. The system according to claim 1, wherein the system comprises: the rhythm comparison module comprises:
the key frame extraction submodule is used for segmenting the input joint point angle information sequence of the sporter and the complete motion angle information sequence of the standard motion template, and extracting segmentation points from extreme points in the joint angle sequence to be used as key frames;
the motion starting judgment sub-module is used for judging whether the motion of the sporter starts or not, reminding the sporter to start the motion if the motion does not start, and performing subsequent comparison if the motion starting judgment is obtained;
and the comparison submodule is used for carrying out similarity comparison rhythm comparison between the input joint angle information sequence of the sporter and the joint angle information sequence of the standard action template, comparing the obtained joint angle information sequence of the sporter with the joint angle information sequence of the standard action template when the sporter starts to exercise, and outputting a comparison result to remind.
4. The system according to claim 3, wherein the system comprises: and comparing the similarity, namely comparing the similarity of the joint angle information sequence of the sporter with the joint information sequence of the standard action template by adopting an open DTW algorithm, outputting the overall evaluation of the action completion degree of the sporter, and comparing the rhythm if the action similarity is within a matched threshold value.
5. The system according to claim 3, wherein the system comprises: and rhythm comparison, which is used for comparing the time between the key frame and the key frame of the sporter with the time interval of the key frame of the standard action template, can judge the advance and delay of the comparison between the sporter and the standard action template, the action rhythm is fast and slow, the movement action similarity accords with the standard action template, the rhythm comparison is carried out, the moment of the key frame of the standard action is recorded, the time interval of the key frame is recorded, the extreme value in the movement angle change of the sporter is read, the time interval is sequentially compared with the time interval of the standard action, if the time interval of the sporter is greater than the time interval of the standard action, the sporter is reminded to pay attention to quickening the rhythm, and if the time interval of the sporter is less than the time interval of the standard action, the sporter is reminded to delay the rhythm.
6. A human motion rhythm comparison method based on attitude estimation is characterized by comprising the following steps:
s1, obtaining the motion videos of the sporter and the coach;
s2, extracting two-dimensional coordinate information of a human joint point in a video by using an Openpos algorithm, respectively obtaining a two-dimensional coordinate information sequence of a joint point of a sporter and a two-dimensional coordinate information sequence of a joint point of a coach movement, preprocessing, converting two-dimensional coordinates of the joint point into angle information, and filling a missing value and an abnormal value by using a linear interpolation method;
s3, establishing a standard action template, comparing the similarity of the joint angle information sequences of a plurality of coaches by adopting a DTW algorithm, and recombining to form a new joint angle information sequence which is the standard action template;
s4, extracting key frames, namely taking frames where extreme points in the joint point angle information sequence are located as the key frames, and performing rhythm comparison on actions by dividing the complete action sequence;
s5, comparing the joint point angle information sequence of the sporter with a standard action template to judge whether the sporter starts to move or not;
s6, comparing the joint point angle information sequence of the sporter with a standard action template to judge whether the sporter movement meets the standard;
s7, comparing the joint point angle information sequence of the sporter with a standard action template to judge the movement rhythm problem of the sporter;
and S8, judging to carry out corresponding reminding according to the motion judgment data.
7. The method of claim 6, wherein the human motion rhythm comparison method based on pose estimation comprises: the comparison mode comprises the following steps:
comparing the similarity, namely comparing the similarity of the joint angle information sequence of the sporter with the joint information sequence of the standard action template by adopting an open DTW algorithm, outputting the overall evaluation of the action completion degree of the sporter, and comparing the rhythm if the action similarity is within a matched threshold value;
and the rhythm comparison is used for comparing the time between the key frames and the key frames of the sporter with the time interval of the key frames of the standard action template, so that the advance and the delay of the comparison between the sporter and the standard action template and the speed and the slow of the action rhythm can be judged.
8. The method of claim 6, wherein the human motion rhythm comparison method based on pose estimation comprises: the step of aligning comprises:
the start of the movement is detected, and the preparation process of the sporter cannot put action comparison, so the start of the movement needs to be detected, and the first ten frames of the joint angle information of the standard action template are taken as the initial matching frame. The match between the sporter and the ten frames in the action is within the threshold value, namely the start of the movement, and the movement comparison is started;
and detecting the motion similarity, comparing the joint angle information sequence of the sporter with a standard motion template by adopting an open DTW algorithm, judging that the motion is qualified if the joint angle information sequence is smaller than a threshold value, smoothing the information sequence to eliminate burr interference, performing rhythm comparison, and directly reminding the sporter that the motion does not accord with the standard motion if the joint angle information sequence is larger than the threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110447115.5A CN113255450A (en) | 2021-04-25 | 2021-04-25 | Human motion rhythm comparison system and method based on attitude estimation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110447115.5A CN113255450A (en) | 2021-04-25 | 2021-04-25 | Human motion rhythm comparison system and method based on attitude estimation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113255450A true CN113255450A (en) | 2021-08-13 |
Family
ID=77221547
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110447115.5A Pending CN113255450A (en) | 2021-04-25 | 2021-04-25 | Human motion rhythm comparison system and method based on attitude estimation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113255450A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116805433A (en) * | 2023-06-27 | 2023-09-26 | 北京奥康达体育科技有限公司 | Human motion trail data analysis system |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4375080A (en) * | 1980-06-04 | 1983-02-22 | Barry Patrick D | Recording and evaluating instrument and method for teacher evaluation |
| US20040125124A1 (en) * | 2000-07-24 | 2004-07-01 | Hyeokman Kim | Techniques for constructing and browsing a hierarchical video structure |
| US20130207890A1 (en) * | 2010-10-22 | 2013-08-15 | Joshua Michael Young | Methods devices and systems for creating control signals |
| US20130231942A1 (en) * | 2011-12-08 | 2013-09-05 | Neurodar, Llc | Apparatus, system, and method for therapy based speech enhancement and brain reconfiguration |
| CN109858457A (en) * | 2019-02-18 | 2019-06-07 | 齐鲁工业大学 | Cheating movement based on OpenPose assists in identifying method and system |
| CN110728220A (en) * | 2019-09-30 | 2020-01-24 | 上海大学 | Gymnastics auxiliary training method based on human body action skeleton information |
| CN111353347A (en) * | 2018-12-21 | 2020-06-30 | 上海形趣信息科技有限公司 | Motion recognition error correction method, electronic device, and storage medium |
| CN111563487A (en) * | 2020-07-14 | 2020-08-21 | 平安国际智慧城市科技股份有限公司 | Dance scoring method based on gesture recognition model and related equipment |
-
2021
- 2021-04-25 CN CN202110447115.5A patent/CN113255450A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4375080A (en) * | 1980-06-04 | 1983-02-22 | Barry Patrick D | Recording and evaluating instrument and method for teacher evaluation |
| US20040125124A1 (en) * | 2000-07-24 | 2004-07-01 | Hyeokman Kim | Techniques for constructing and browsing a hierarchical video structure |
| US20130207890A1 (en) * | 2010-10-22 | 2013-08-15 | Joshua Michael Young | Methods devices and systems for creating control signals |
| US20130231942A1 (en) * | 2011-12-08 | 2013-09-05 | Neurodar, Llc | Apparatus, system, and method for therapy based speech enhancement and brain reconfiguration |
| CN111353347A (en) * | 2018-12-21 | 2020-06-30 | 上海形趣信息科技有限公司 | Motion recognition error correction method, electronic device, and storage medium |
| CN109858457A (en) * | 2019-02-18 | 2019-06-07 | 齐鲁工业大学 | Cheating movement based on OpenPose assists in identifying method and system |
| CN110728220A (en) * | 2019-09-30 | 2020-01-24 | 上海大学 | Gymnastics auxiliary training method based on human body action skeleton information |
| CN111563487A (en) * | 2020-07-14 | 2020-08-21 | 平安国际智慧城市科技股份有限公司 | Dance scoring method based on gesture recognition model and related equipment |
Non-Patent Citations (2)
| Title |
|---|
| THIERRY PINHEIRO MOREIRA ET AL.: "First-person action recognition through Visual Rhythm texture description", 《 2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP)》 * |
| 李丹丹: "数字化芭蕾基训教学演示的设计与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116805433A (en) * | 2023-06-27 | 2023-09-26 | 北京奥康达体育科技有限公司 | Human motion trail data analysis system |
| CN116805433B (en) * | 2023-06-27 | 2024-02-13 | 北京奥康达体育科技有限公司 | Human motion trail data analysis system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111144217B (en) | Motion evaluation method based on human body three-dimensional joint point detection | |
| CN108256433B (en) | Motion attitude assessment method and system | |
| CN108564596B (en) | Intelligent comparison analysis system and method for golf swing video | |
| CN110263720B (en) | Action recognition method based on depth image and skeleton information | |
| CN111437583B (en) | Badminton basic action auxiliary training system based on Kinect | |
| Chaudhari et al. | Yog-guru: Real-time yoga pose correction system using deep learning methods | |
| Yang et al. | Unik: A unified framework for real-world skeleton-based action recognition | |
| Chen et al. | Yoga posture recognition for self-training | |
| CN110448870B (en) | Human body posture training method | |
| Sabo et al. | Estimating parkinsonism severity in natural gait videos of older adults with dementia | |
| US20100208038A1 (en) | Method and system for gesture recognition | |
| CN110738154A (en) | pedestrian falling detection method based on human body posture estimation | |
| CN109758756B (en) | Gymnastics video analysis method and system based on 3D camera | |
| CN105664462A (en) | Auxiliary training system based on human body posture estimation algorithm | |
| CN114550027A (en) | Vision-based motion video fine analysis method and device | |
| Yang et al. | Human exercise posture analysis based on pose estimation | |
| CN113975775B (en) | Wearable inertial body feeling ping-pong exercise training system and working method thereof | |
| CN113255450A (en) | Human motion rhythm comparison system and method based on attitude estimation | |
| Bandini et al. | A wearable vision-based system for detecting hand-object interactions in individuals with cervical spinal cord injury: First results in the home environment | |
| Yadav et al. | YogaTube: a video benchmark for Yoga action recognition | |
| CN114550071B (en) | Method, device and medium for automatically identifying and capturing track and field video action key frames | |
| CN117333932A (en) | Method, equipment and medium for identifying sarcopenia based on machine vision | |
| Tsai et al. | Temporal-variation skeleton point correction algorithm for improved accuracy of human action recognition | |
| Kumar et al. | Self-learning yoga pose with accuracy detection using deep learning | |
| Zhou et al. | Kinematics parameter analysis of long jump based on human posture vision |
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 |