CN111275032A - Deep squatting detection method, device, equipment and medium based on human body key points - Google Patents
Deep squatting detection method, device, equipment and medium based on human body key points Download PDFInfo
- Publication number
- CN111275032A CN111275032A CN202010377914.5A CN202010377914A CN111275032A CN 111275032 A CN111275032 A CN 111275032A CN 202010377914 A CN202010377914 A CN 202010377914A CN 111275032 A CN111275032 A CN 111275032A
- Authority
- CN
- China
- Prior art keywords
- action
- key point
- included angle
- preset
- tester
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/23—Recognition of whole body movements, e.g. for sport training
-
- 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
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Eye Examination Apparatus (AREA)
Abstract
The invention discloses a deep squatting detection method, a device, equipment and a medium based on human body key points, wherein the method comprises the steps of identifying each original frame image according to the time sequence of the original frame image, acquiring bone key points to be identified corresponding to each tester, determining the action state of the tester based on the bone key points to be identified, acquiring the actual action duration based on the action starting time and the action ending time, and selecting two-foot key points, two-shoulder key points, a hip key point and a neck key point based on the bone key points to be identified to acquire the action detection result of the tester when the actual action duration is less than or equal to the preset standard action duration; when the action detection result is the action standard, the deep-squatting number is counted based on the action ending moment, the efficiency and the accuracy of detecting the deep-squatting action of the tester are improved, and the labor cost is saved.
Description
Technical Field
The invention relates to the field of fitness movement detection, in particular to a deep squatting detection method, device, equipment and medium based on human body key points.
Background
In fitness or strength exercises, the squat motion is a compound, full-body exercise motion that exercises the exerciser's hip and posterior thigh muscles, as well as enhancing the exerciser's joint flexibility and body coordination. Wherein, standard deep squat movements require the back and the waist to be kept straight, and the hip joint is lower than the knee joint. At present, whether the deep squatting action training of a tester meets the training requirement is judged, and the training requirement is mainly obtained through the following two aspects of data: firstly, judging whether the deep squatting action of a tester is standard or not; and secondly, whether the number of standard deep squatting actions of the tester meets the training requirement within the specified time or not. Whether the action of squatting deeply of tester is standard is confirmed mainly through artifical observation at present to the action of squatting deeply to the standard is counted, and the cost of labor is high, and is inefficient, and the error is big, the condition of practising fraud appears easily.
Disclosure of Invention
The invention aims to solve the technical problems that whether the deep squatting action of a tester is standard or not is manually determined, the number of the standard deep squatting actions is counted, the cost is high, the efficiency is low, the error is large, and the cheating condition is easy to occur. Therefore, the deep squatting detection method, the device, the equipment and the medium based on the human body key points are provided, so that the labor cost is reduced, and the judgment efficiency and the judgment accuracy are improved.
The invention is realized by the following technical scheme:
a deep squatting detection method based on human body key points comprises the following steps:
acquiring original frame images, identifying each original frame image according to the time sequence of the original frame images based on a human body posture identification algorithm, and acquiring bone key points to be identified corresponding to each tester;
acquiring an included angle of a knee key point and displacement of a neck key point in the vertical direction based on the bone key point to be identified;
determining whether the action state of the tester is action start or not based on the knee key point included angle and the displacement of the neck key point in the vertical direction;
when the action state is action start, comparing the knee key point included angle with a preset standing included angle threshold value, a preset movement included angle threshold value and a preset squatting included angle threshold value respectively, and comparing the displacement of the neck key point in the vertical direction with a preset displacement to determine whether the action state of the tester is in action;
when the action state is in motion, comparing the knee key point included angle with the preset standing included angle threshold value, and comparing the displacement of the neck key point in the vertical direction with the preset displacement to determine whether the action state of the tester is the action end;
when the action state is action end, taking the moment when the action state is action start as action start moment, taking the moment when the action state is action end as action end moment, and acquiring an actual action duration based on the action start moment and the action end moment;
when the actual action duration is less than or equal to a preset standard action duration, selecting a two-foot key point, a two-shoulder key point, a hip key point and the neck key point based on the bone key point to be identified, and acquiring an action detection result of the tester;
and when the action detection result is an action standard, counting the number of deep squats based on the action ending time.
Further, the determining whether the action state of the tester is action start based on the knee key point included angle and the displacement of the neck key point in the vertical direction includes:
comparing the included angle of the knee key point with a preset standing included angle threshold value, and comparing the displacement of the neck key point in the vertical direction with a preset displacement;
and when the included angle of the knee key point is larger than a preset standing included angle threshold value and the displacement of the neck key point in the vertical direction is smaller than or equal to a preset displacement, determining that the action state of the tester is the action start.
Further, will knee key point contained angle respectively with predetermine stand contained angle threshold value, predetermine motion contained angle threshold value and predetermine the contained angle threshold value of squatting and compare, and will the displacement of neck key point on vertical direction is compared with predetermine the displacement, confirms whether tester's action state is in the action, include:
if the included angle of the knee key point is larger than the preset motion included angle threshold value and smaller than or equal to the preset standing included angle threshold value, and the displacement of the neck key point in the vertical direction is larger than the preset displacement, determining that the action state of the tester in the corresponding original frame image is in descending action;
when the action state is descending, if the included angle of the knee key point is greater than the preset squat included angle threshold and less than or equal to the preset movement included angle threshold, and the displacement of the neck key point in the vertical direction is greater than the preset displacement, determining that the action state of the tester in the corresponding original frame image is squat action;
when the action state is squatting, if knee key point included angle is greater than preset motion included angle threshold value and less than or equal to preset standing included angle threshold value, the displacement of neck key point in vertical direction is greater than preset displacement, then confirm the action state of tester in corresponding original frame image is in the action of rising.
Further, the selecting a pair of feet key points, a pair of shoulders key points, a hip key point and the neck key point based on the bone key points to be identified to obtain the action detection result of the tester includes:
calculating the distance between the key points of the feet based on the key points of the feet, and calculating the distance between the key points of the shoulders based on the key points of the shoulders;
calculating the distance difference between the distance between the two-foot key points and the distance between the two-shoulder key points, and taking the absolute value of the distance difference as a first distance;
calculating a second distance based on the hip keypoints and the neck keypoints;
and when the first distance is smaller than or equal to a first calculation threshold value and the second distance is larger than or equal to a second calculation threshold value, acquiring an action detection result of the action standard.
Further, the acquiring of the original frame images, based on a human body posture recognition algorithm, recognizing each of the original frame images according to a time sequence of the original frame images, and acquiring the bone key points to be recognized corresponding to each tester includes:
acquiring original frame images, and tracking each original frame image according to the time sequence of the original frame images based on a pedestrian detection algorithm to acquire a target to be identified;
carrying out face recognition on each target to be recognized based on a face recognition algorithm to obtain identity information;
and identifying each target to be identified based on a human body posture identification algorithm to obtain the bone key points to be identified corresponding to each tester.
Further, the acquiring of the original frame images, based on a pedestrian detection algorithm, performing target tracking on each of the original frame images according to the time sequence of the original frame images, and acquiring a target to be identified includes:
acquiring original frame images, and preprocessing each pair of original frame images according to the time sequence of the original frame images to acquire effective frame images;
and tracking the target of each effective frame image based on a pedestrian detection algorithm to obtain the target to be identified.
Further, the deep squatting detection method further comprises the following steps:
according to the time sequence, taking the first action starting time as the motion starting time;
based on the motion starting time, counting the number of squats within the preset motion duration, based on the identity information, selecting a corresponding scoring standard to score the number of squats, acquiring the motion score of the tester, and determining whether the motion of the tester reaches the standard or not based on the motion score.
A detection device squats deeply based on human key point includes:
the bone key point acquisition module to be identified is used for acquiring original frame images, identifying each original frame image according to the time sequence of the original frame images based on a human body posture identification algorithm, and acquiring the bone key points to be identified corresponding to each tester;
the bone key point processing module is used for acquiring the included angle of knee key points and the displacement of neck key points in the vertical direction based on the bone key points to be identified;
the action starting judging module is used for determining whether the action state of the tester is action starting or not based on the knee key point included angle and the displacement of the neck key point in the vertical direction;
the motion in-motion judging module is used for comparing the included angle of the knee key point with a preset standing included angle threshold value, a preset motion included angle threshold value and a preset squatting included angle threshold value respectively when the motion state is motion start, comparing the displacement of the neck key point in the vertical direction with a preset displacement, and determining whether the motion state of the tester is in motion or not;
the action ending judging module is used for comparing the included angle of the knee key point with the preset standing included angle threshold value when the action state is in motion, comparing the displacement of the neck key point in the vertical direction with the preset displacement, and determining whether the action state of the tester is the action ending;
the action duration calculation module is used for taking the moment when the action state is the action start moment as the action start moment when the action state is the action end, taking the moment when the action state is the action end as the action end moment, and acquiring the actual action duration based on the action start moment and the action end moment;
the action detection result acquisition module is used for selecting a two-foot key point, a two-shoulder key point, a hip key point and the neck key point based on the bone key point to be identified when the actual action duration is less than or equal to a preset standard action duration, and acquiring an action detection result of the tester;
and the deep squatting number counting module is used for counting the deep squatting number based on the action ending moment when the action detection result is the action standard.
A computer device comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, wherein the processor executes the computer program to realize the deep squatting detection method based on the human body key points.
A computer-readable storage medium storing a computer program which, when executed by a processor, implements the above-described human keypoint-based deep squat detection method.
The invention is based on the deep squat detection method, the device, the equipment and the medium of the human body key points, identifies each original frame image through a human body posture identification algorithm, obtains the bone key points to be identified corresponding to each tester, and based on the bone key points to be identified, determining the action state of the tester doing a deep squatting action, then determining the actual action time length of the tester according to the action state of the tester, when the actual action time length is less than or equal to the preset standard action time length, determining whether the deep-squatting action of the tester is standard according to the key points of the feet, the key points of the shoulders, the key points of the buttocks and the key points of the neck in the key points of the bones to be identified, then the number of the squat actions of the tester is counted to complete the detection of the squat actions of the tester, so that the detection efficiency and accuracy are improved, and the labor cost is saved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principles of the invention. In the drawings:
FIG. 1 is a flow chart of the deep squatting detection method based on human body key points.
Fig. 2 is a specific flowchart of step S30 in fig. 1.
Fig. 3 is a specific flowchart of step S40 in fig. 1.
Fig. 4 is a specific flowchart of step S70 in fig. 1.
Fig. 5 is a specific flowchart of step S10 in fig. 1.
Fig. 6 is a specific flowchart of step S11 in fig. 5.
FIG. 7 is another flowchart of the squat detection method based on human body key points according to the present invention.
FIG. 8 is a schematic structural diagram of the deep squatting detection device based on the key points of the human body.
FIG. 9 is a schematic diagram of the computer apparatus of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to examples and accompanying drawings, and the exemplary embodiments and descriptions thereof are only used for explaining the present invention and are not meant to limit the present invention.
Example 1
The invention provides a deep squatting detection method based on human body key points, which can be applied to different electronic equipment, wherein the electronic equipment comprises but is not limited to various personal computers, notebook computers, smart phones and tablet computers.
As shown in figure 1, the invention provides a deep squatting detection method based on human body key points, which comprises the following steps:
s10: acquiring original frame images, identifying each original frame image according to the time sequence of the original frame images based on a human body posture identification algorithm, and acquiring bone key points to be identified corresponding to each tester.
Wherein, the original frame image refers to a single frame image which is sent by the server and shoots the squat action of the tester.
Specifically, after acquiring the original frame images, the server identifies each original frame image according to the time sequence of the original frame images, and acquires the bone key points to be identified from each original frame image. The skeleton key points to be recognized refer to points which represent a human body frame and are obtained by recognizing a human body in the original frame image through a human body posture recognition algorithm.
In the present embodiment, 15 skeletal key points to be identified are described, which are respectively 1 right shoulder, 2 right elbow, 3 right wrist, 4 left shoulder, 5 left elbow, 6 left wrist, 7 right hip, 8 right knee, 9 right ankle, 10 left hip, 11 left knee, 12 left ankle, 13 vertex, 14 neck and 15 hip, and the 15 skeletal key points to be identified are usually connected by line segments to form a human body frame.
The human body gesture recognition algorithm in this embodiment refers to an algorithm of a human body bone key point in an image, and the human body gesture recognition algorithm in this embodiment includes, but is not limited to, OpenPose and AlphaPose.
S20: based on the bone key points to be identified, the included angle of the knee key points and the displacement of the neck key points in the vertical direction are obtained.
The knee key point included angle v refers to an included angle formed by corresponding bone key points to be identified to the left hip, the left knee and the left ankle, or an included angle formed by corresponding bone key points to be identified to the right hip, the right knee and the right ankle.
Specifically, after acquiring an original frame image, the server establishes a rectangular coordinate system with the upper left corner of the original frame image as the origin of coordinates, and through the rectangular coordinate system, the server selects a corresponding number of original frame images according to a preset number of frames (for example, 5 frames), and acquires the abscissa and the ordinate of a neck key point from each original frame image, and calculates the movement displacement of the neck key point in the other original frame images relative to a reference point with the neck key point in the first frame original frame image as the reference point, and if the movement displacement is smaller than a set movement judgment value, it is determined that the tester does not start a deep squatting action. The movement displacement refers to a distance that a neck key point in the original frame image moves relative to the reference point. The movement judgment value is a value for judging whether the movement of the key point of the neck of the subject is a movement for starting a deep-squatting action.
Further, the displacement of the neck key point in the vertical direction in the present embodiment refers to the displacement of the neck key point moving with respect to the reference point.
S30: and determining whether the action state of the tester is the action start or not based on the knee key point included angle and the displacement of the neck key point in the vertical direction.
Specifically, after acquiring an included angle v of a knee key point and a displacement D of a neck key point in the vertical direction from each frame of original frame image, the server compares the included angle v of the knee key point in each frame of original frame image with a preset standing angle threshold α, compares the displacement D of the neck key point in the vertical direction with a preset displacement D, and when the included angle v of the knee key point is greater than the preset standing angle threshold α and the displacement D of the neck key point in the vertical direction is less than or equal to the preset displacement D, it indicates that the frame of original frame image starts a squat action for a tester, and the action state of the tester is the action start.
S40: when the action state is action start, the included angle of the knee key point is respectively compared with a preset standing included angle threshold value, a preset movement included angle threshold value and a preset squatting included angle threshold value, the displacement of the neck key point in the vertical direction is compared with the preset displacement, and whether the action state of the tester is in action or not is determined.
The preset motion included angle threshold β refers to a preset included angle for judging whether the tester is in a motion stage, and the preset squat included angle threshold γ refers to a preset included angle for judging whether the tester is in a standing stage.
Specifically, after determining that the motion state represented by a certain frame of original frame image is motion start, the subsequently acquired original frame image is continuously detected, when it is detected that the knee key point angle v in a certain original frame image is greater than a preset motion angle threshold β and less than or equal to a preset standing angle threshold α, and the displacement D of the neck key point of the tester in the vertical direction is greater than a preset displacement D, it is determined that the tester is performing a descending motion and the descending motion is standard, when it is determined that the tester is performing a descending motion, the server continuously detects the subsequently acquired original frame image, when it is detected that the knee key point angle v in a certain original frame image is greater than a preset squatting angle threshold γ and less than or equal to a preset motion angle threshold β, and the displacement D of the neck key point of the tester in the vertical direction is greater than a preset displacement D, it is determined that the tester is performing a squatting motion and the squatting motion is standard, when it is determined that the knee key point angle v in a certain original frame image is greater than a preset displacement D, and the knee key point displacement D in the vertical direction is greater than a preset displacement D, and the test neck angle displacement D is greater than a preset displacement D in the vertical direction of the test neck angle threshold 36, and the test procedure is determined that the test key point is greater than the preset displacement D and the vertical angle D is greater than the preset displacement D in the test angle threshold 36.
S50: and when the action state is in motion, comparing the included angle of the knee key point with a preset standing included angle threshold value, and comparing the displacement of the neck key point in the vertical direction with a preset displacement to determine whether the action state of the tester is the action end.
Specifically, after determining that the motion state represented by a certain frame of original frame image is in motion, the server continues to detect the subsequently acquired original frame images, and when detecting that the knee key point included angle v in a certain original frame image is greater than a preset standing angle threshold value α and the displacement D of the neck key point of the tester in the vertical direction is less than or equal to a preset displacement D, it is also standard to represent that the motion of the tester is finished, and when the motion of the tester is finished, it represents that the tester completes a deep-squatting motion, and the motion state of the tester is finished.
S60: and when the action state is action end, taking the moment when the action state is action start as the action start moment and the moment when the action state is action end as the action end moment, and acquiring the actual action time length based on the action start moment and the action end moment.
In particular, by
Calculating the actual action duration, wherein,
the time when the motion is started is referred to as the motion starting time,
the time when the finger movement is finished is referred to,
refers to the actual action duration. The actual action duration in this embodiment refers to the duration actually required for the tester to perform a squat action.
S70: and when the actual action duration is less than or equal to the preset standard action duration, selecting a double-foot key point, a double-shoulder key point, a hip key point and a neck key point based on the bone key point to be identified, and acquiring an action detection result of the tester.
Wherein, the standard action duration is preset
Refers to the preset time required to theoretically complete a standard deep squatting action.
In particular, when
If the test result is positive, the tester completes a deep squatting action within the specified standard action duration, and the test result is qualified. The server will base on the bone key points to be identifiedThe method comprises the steps of selecting two-foot key points, two-shoulder key points, a hip key point and a neck key point, calculating the distance between the two-foot key points according to the two-foot key points, calculating the distance between the two-shoulder key points according to the two-shoulder key points, calculating the distance between the hip key point and the neck key point according to the hip key point and the neck key point, calculating the distance difference between the distance between the two-foot key points and the distance between the two-shoulder key points, and finally judging the action detection result of a tester by judging whether the distance difference and the distance between the hip key point and the neck key point reach preset conditions or not.
Further, when
If the tester needs to continue the test, the test method can be restarted.
S80: and when the action detection result is the action standard, counting the number of deep squats based on the action ending time.
Step S10-step S80, recognizing skeleton key points to be recognized of all testers in each frame of original frame image through a human posture recognition algorithm, obtaining the included angle of knee key points and the displacement of neck key points in the vertical direction, determining whether the action state of the testers in action is standard or not by judging whether the included angle of the knee key points and the displacement of the neck key points in the vertical direction meet action starting conditions, action middle conditions and action ending conditions or not, judging whether the action state of action ending is standard or not when the action state of the testers in action is standard, and if the action state of action ending is standard, indicating that the testers finish a deep squatting action. After the standard is detected in the deep-squatting process, whether the tester completes a deep-squatting action within the set time needs to be detected, if the tester completes a deep-squatting action within the set time, the tester completes a deep-squatting action within the set time is indicated, in order to detect whether the whole deep-squatting of the tester is standard, the distance difference between the distance between the key points of the feet and the distance between the key points of the shoulders of the tester and whether the distance between the key points of the buttocks and the key points of the neck meet the requirement of standard deep-squatting are also needed to be detected, whether the tester is standard deep-squatting is determined, whether the test on the standard of the deep-squatting action of the tester is completed is detected, the detection accuracy is effectively improved, the number of the deep-squatting actions of the action standard is counted based on the moment when the action is finished, the number of the tester is automatically counted, and the labor cost is reduced.
As shown in fig. 2, further, in step S30, determining whether the action state of the tester is action start based on the knee key point included angle and the displacement of the neck key point in the vertical direction, specifically includes the following steps:
s31: and comparing the included angle of the knee key point with a preset standing included angle threshold value, and comparing the displacement of the neck key point in the vertical direction with a preset displacement.
S32: and when the included angle of the knee key point is larger than a preset standing included angle threshold value and the displacement of the neck key point in the vertical direction is smaller than or equal to a preset displacement, determining that the action state of the tester is the action start.
As shown in fig. 3, further, in step S40, comparing the knee key point included angle with a preset standing included angle threshold, a preset exercise included angle threshold and a preset squatting included angle threshold, and comparing the displacement of the neck key point in the vertical direction with a preset displacement, to determine whether the action state of the tester is in action, specifically including the following steps:
s41: if the included angle of the knee key point is larger than the preset motion included angle threshold value and smaller than or equal to the preset standing included angle threshold value, and the displacement of the neck key point in the vertical direction is larger than the preset displacement, the action state of the tester in the corresponding original frame image is determined to be in descending action.
S42: when the action state is descending, if the included angle of the knee key point is larger than a preset squat included angle threshold value and smaller than or equal to a preset movement included angle threshold value, and the displacement of the neck key point in the vertical direction is larger than a preset displacement, the action state of the tester in the corresponding original frame image is determined to be squat action.
Further, if knee key point included angle is less than or equal to the preset threshold value of the included angle of squatting, the tester is considered to squat too downwards, and the test is an irregular action of squatting.
S43: when the action state is squatting, if the included angle of the knee key points is larger than a preset motion included angle threshold value and smaller than or equal to a preset standing included angle threshold value, and the displacement of the neck key points in the vertical direction is larger than a preset displacement, the action state of the tester in the corresponding original frame image is determined to be ascending.
And S41-S42, determining whether the action state of the tester in the movement is standard or not by judging whether the included angle of the knee key point and the displacement of the neck key point in the vertical direction meet the set requirements in the action, and improving the detection accuracy.
As shown in fig. 4, further, in step S70, selecting a two-foot key point, a two-shoulder key point, a hip key point, and a neck key point based on the bone key point to be identified, and obtaining the motion detection result of the tester specifically includes the following steps:
s71: and calculating the distance between the key points of the feet based on the key points of the feet, and calculating the distance between the key points of the shoulders based on the key points of the shoulders.
S72: and calculating the distance difference between the two-foot key point distance and the two-shoulder key point distance, and taking the absolute value of the distance difference as the first distance.
S73: a second distance is calculated based on the hip keypoint and the neck keypoint.
S74: and when the first distance is smaller than or equal to the first calculation threshold and the second distance is larger than or equal to the second calculation threshold, acquiring the action detection result of the action standard.
The first calculation threshold refers to a value set for judging whether the distance difference between the two-foot key point distance and the two-shoulder key point distance meets the requirement or not. The second calculation threshold value is a value set for determining whether the distance between the hip key point and the neck key point satisfies a requirement.
As shown in fig. 5, further, in step S10, acquiring original frame images, identifying each original frame image according to a time sequence of the original frame images based on a human body posture identification algorithm, and acquiring bone key points to be identified corresponding to each tester, specifically including the following steps:
s11: and acquiring original frame images, and tracking the target of each original frame image according to the time sequence of the original frame images based on a pedestrian detection algorithm to acquire the target to be identified.
The pedestrian detection algorithm in the embodiment refers to an algorithm for detecting and tracking a pedestrian in an original frame image, and includes two algorithms of pedestrian detection and pedestrian tracking.
Wherein, the target to be identified refers to the image data of the squat action of each tester of the original frame image.
S12: and carrying out face recognition on each target to be recognized based on a face recognition algorithm to obtain identity information.
The Face recognition algorithm in this embodiment includes, but is not limited to, a Face recognition open source library Dlib, libfacedetection, or a commercial open Face recognition library SDK, a Face + + Face recognition library, and a hundredth AI Face recognition library.
The to-be-identified information refers to data which can represent the identity of the tester and is obtained by carrying out face recognition on each to-be-identified target according to a face recognition algorithm, and the data includes but is not limited to age and gender.
S13: and identifying each target to be identified based on a human body posture identification algorithm, and acquiring the bone key points to be identified corresponding to each tester.
As shown in fig. 6, further, in step S11, acquiring original frame images, and based on a pedestrian detection algorithm, performing target tracking on each original frame image according to the time sequence of the original frame images to acquire a target to be identified, specifically including the following steps:
s111: and acquiring original frame images, preprocessing each pair of original frame images according to the time sequence of the original frame images, and acquiring effective frame images.
The effective frame image refers to an image obtained by performing image processing on an original frame image.
S112: and tracking the target of each effective frame image based on a pedestrian detection algorithm to obtain the target to be identified.
As shown in fig. 7, further, the method further includes the following steps:
s91: and taking the first action starting moment as the motion starting moment according to the time sequence.
Since some images (such as adjustment state, standing according to regulations, etc.) before the tester performs the deep squat action exist in the original frame image, in order to improve the accuracy of subsequent judgment, the first action starting time is taken as the motion starting time. Wherein, the exercise starting moment refers to the moment when the tester starts to do the deep squatting exercise.
S92: and counting the number of squats within the preset exercise duration based on the exercise starting time, grading the number of squats based on the grading standard corresponding to the selected area based on the identity information to obtain the exercise score of the tester, and determining whether the exercise of the tester reaches the standard based on the exercise score.
Wherein, the preset exercise duration refers to the preset exercise duration which can achieve the deep squatting training effect.
Because testers of different ages and different sexes have different physical qualities and the number of deep squats of action standards which can be made in the specified preset exercise time is different, different scoring standards are set for how many deep squats of action standards are made by testers of different ages and sexes. If the exercise training time is 15 minutes, men between the ages of 20 and 30 do 35 or less for 60 minutes, 35 to 40 (including end points) for 80 minutes, 40 or more for 95 minutes, the duration of the male plate supporting action between the ages of 30 and 35 is 30 or less for 60 minutes, 30 to 35 (including end points) for 80 minutes, 35 or more for 95 minutes, women between the ages of 20 and 30 do 30 or less for 60 minutes, 30 to 35 (including end points) for 80 minutes, 35 or more for 95 minutes, the duration of the female plate supporting action between the ages of 30 and 35 is 25 or less for 60 minutes, 25 to 30 (including end points) for 80 minutes, 35 or more for 95 minutes, 80 or more for reaching the standard, and 80 or less for not reaching the standard.
According to the deep squatting detection method based on the human body key points, the effective frame image is obtained by preprocessing the original frame image, so that the data of subsequent image processing is reduced, and the efficiency of the subsequent image processing is improved. The target tracking is carried out on each effective frame image through a pedestrian detection algorithm and a face recognition algorithm, then the face recognition is carried out on the tracked target, so that each tester in the effective frame image is accurately recognized, and a reliable data source is provided for the subsequent process of accurately recognizing the squat of each tester. The method comprises the steps of identifying skeleton key points to be identified of all testers in each frame of original frame image through a human posture identification algorithm, obtaining knee key point included angles and displacement of neck key points in the vertical direction, determining whether the testers are in a standard action state or not by judging whether the knee key point included angles and the displacement of the neck key points in the vertical direction meet action starting conditions, action middle conditions and action ending conditions or not, judging whether the action state of the testers in the action is standard or not when the testers carry the action, and indicating that the testers finish a deep squatting action if the action state of the action ending is standard. After the standard of the deep-squatting process is detected, whether the tester completes a deep-squatting action within the specified time is also detected, if the tester completes a deep-squatting action within the specified time, it means that the tester completes a deep squat within a prescribed time, and in order to detect whether the whole deep squat of the tester is standard, it is also necessary to detect the distance difference between the distance between the key points of the feet and the distance between the key points of the shoulders of the tester, and whether the distance between the key points of the buttocks and the key points of the neck meets the requirement of standard deep squat, so as to determine whether the tester squats deeply or not, finish the detection of whether the tester squats deeply or not, effectively improve the detection accuracy, in order to determine whether the exercise of the tester reaches the standard, the exercise duration of the tester needs to be scored so as to realize automatic statistics and judgment of whether the deep squatting exercise reaches the standard without manual judgment.
Example 2
As shown in fig. 8, the present embodiment is different from embodiment 1 in that a deep squatting detection device based on key points of a human body comprises:
and the bone key point to be identified acquisition module 10 is configured to acquire original frame images, identify each original frame image according to a time sequence of the original frame image based on a human body posture identification algorithm, and acquire a bone key point to be identified corresponding to each tester.
And the bone key point to be identified processing module 20 is configured to obtain a knee key point included angle and a displacement of the neck key point in the vertical direction based on the bone key point to be identified.
And the action starting judging module 30 is used for determining whether the action state of the tester is the action starting or not based on the knee key point included angle and the displacement of the neck key point in the vertical direction.
And the action in-process judging module 40 is used for comparing the knee key point included angle with a preset standing included angle threshold value, a preset movement included angle threshold value and a preset squatting included angle threshold value respectively when the action state is the action start, comparing the displacement of the neck key point in the vertical direction with the preset displacement, and determining whether the action state of the tester is in action.
And the action ending judgment module 50 is configured to compare the knee key point included angle with a preset standing included angle threshold value when the action state is in motion, compare the displacement of the neck key point in the vertical direction with a preset displacement, and determine whether the action state of the tester is an action ending.
And an action duration calculation module 60, configured to, when the action state is an action end, take a time when the action state is the action start as an action start time, and take a time when the action state is the action end as an action end time, and obtain an actual action duration based on the action start time and the action end time.
And an action detection result obtaining module 70, configured to select a two-foot key point, a two-shoulder key point, a hip key point, and a neck key point based on the bone key point to be identified, and obtain an action detection result of the tester, when the actual action duration is less than or equal to a preset standard action duration.
And a deep-squatting number counting module 80, configured to count the number of deep squats based on the action ending time when the action detection result is the action standard.
Further, the action start judging module 30 includes an action start parameter comparing unit and an action start result acquiring unit.
And the action starting parameter comparison unit is used for comparing the included angle of the knee key point with a preset standing included angle threshold value and comparing the displacement of the neck key point in the vertical direction with a preset displacement.
And the action starting result acquisition unit is used for determining that the action state of the tester is action starting when the included angle of the knee key point is greater than a preset standing included angle threshold value and the displacement of the neck key point in the vertical direction is less than or equal to a preset displacement.
Further, the in-motion determination module 40 includes a descending motion state determination unit, a squatting motion state determination unit, and an ascending motion state determination unit.
And the descending action state judging unit is used for determining that the action state of the tester in the corresponding original frame image is in descending action if the included angle of the knee key point is greater than a preset motion included angle threshold value and less than or equal to a preset standing included angle threshold value and the displacement of the neck key point in the vertical direction is greater than a preset displacement.
And the squat action state judging unit is used for determining that the action state of the tester in the corresponding original frame image is in the squat action if the knee key point included angle is greater than a preset squat included angle threshold value and less than or equal to a preset movement included angle threshold value and the displacement of the neck key point in the vertical direction is greater than a preset displacement when the action state is in the descent.
And the ascending action state judging unit is used for determining that the action state of the tester in the corresponding original frame image is ascending action if the knee key point included angle is greater than a preset movement included angle threshold value and less than or equal to a preset standing included angle threshold value and the displacement of the neck key point in the vertical direction is greater than a preset displacement when the action state is squatting.
Further, the motion detection result acquisition module 70 includes a key point parameter acquisition unit, a first distance calculation unit, a second distance calculation unit, and a motion detection result acquisition unit.
And the key point parameter acquisition unit is used for calculating the distance between the key points of the feet based on the key points of the feet and calculating the distance between the key points of the shoulders based on the key points of the shoulders.
And the first distance calculation unit is used for calculating the distance difference between the two-foot key point distance and the two-shoulder key point distance and taking the absolute value of the distance difference as the first distance.
A second distance calculating unit for calculating a second distance based on the hip key point and the neck key point.
And the action detection result acquisition unit is used for acquiring an action detection result of the action standard when the first distance is less than or equal to a first calculation threshold value and the second distance is greater than or equal to a second calculation threshold value.
Further, the module 10 for acquiring a bone key point to be recognized includes an acquiring unit of a target to be recognized, an acquiring unit of identity information, and an acquiring unit of a bone key point to be recognized.
And the target to be recognized acquisition unit is used for acquiring the original frame images, and tracking the target of each original frame image according to the time sequence of the original frame images based on a pedestrian detection algorithm to acquire the target to be recognized.
And the identity information acquisition unit is used for carrying out face recognition on each target to be recognized based on a face recognition algorithm to acquire identity information.
And the bone key point acquisition unit to be recognized is used for recognizing each target to be recognized based on a human body posture recognition algorithm and acquiring the bone key points to be recognized corresponding to each tester.
Further, the target to be identified acquisition unit comprises an original frame image processing unit and a target tracking unit
And the original frame image processing unit is used for acquiring original frame images, preprocessing each pair of original frame images according to the time sequence of the original frame images and acquiring effective frame images.
And the target tracking unit is used for tracking the target of each effective frame image based on a pedestrian detection algorithm to obtain the target to be identified.
Furthermore, the deep squatting detection device based on the human body key points further comprises a movement moment acquisition module, a movement duration calculation module and a movement standard reaching judgment module.
And the movement moment acquisition module is used for taking the first action starting moment as the movement starting moment according to the time sequence.
And the exercise standard-reaching judging module is used for counting the number of deep squats within the preset exercise duration based on the exercise starting time, selecting a corresponding scoring standard to score the number of deep squats based on the identity information, acquiring the exercise score of the tester, and determining whether the exercise of the tester reaches the standard based on the exercise score.
For specific definition of deep-squatting detection based on human body key points, reference may be made to the above definition of the deep-squatting detection method based on human body key points, and details are not repeated here. All modules in the deep squatting detection based on the human body key points can be completely or partially realized through software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
Example 3
The embodiment provides a computer device, which may be a server, and the internal structure diagram of the computer device may be as shown in fig. 9. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a computer readable storage medium, an internal memory. The computer readable storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the computer-readable storage medium. The database of the computer device is used for storing data involved in the method of deep squat detection based on human body key points. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a deep squat detection method based on human body key points.
The present embodiment provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for deep-squatting detection based on human body key points in the foregoing embodiments when executing the computer program, for example, steps S10-S80 shown in fig. 1 or steps shown in fig. 2 to 7, and are not repeated herein to avoid repetition. Alternatively, the processor, when executing the computer program, implements the functionality of the modules/units of the apparatus based on deep squat detection of human body key points in the above described embodiments, e.g. the functionality of modules 10 to 80 shown in fig. 8. To avoid repetition, further description is omitted here.
Example 4
The present embodiment provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the information pushing method in the foregoing embodiments are implemented, for example, steps S10 to S80 shown in fig. 1 or steps shown in fig. 2 to fig. 7, which are not described herein again to avoid repetition. Alternatively, the processor implements the functions of each module/unit in the embodiment of the information pushing apparatus when executing the computer program, for example, the functions of the modules 10 to 80 shown in fig. 8. To avoid repetition, further description is omitted here.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are merely exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A deep squatting detection method based on human body key points is characterized by comprising the following steps:
acquiring original frame images, identifying each original frame image according to the time sequence of the original frame images based on a human body posture identification algorithm, and acquiring bone key points to be identified corresponding to each tester;
acquiring an included angle of a knee key point and displacement of a neck key point in the vertical direction based on the bone key point to be identified;
determining whether the action state of the tester is action start or not based on the knee key point included angle and the displacement of the neck key point in the vertical direction;
when the action state is action start, comparing the knee key point included angle with a preset standing included angle threshold value, a preset movement included angle threshold value and a preset squatting included angle threshold value respectively, and comparing the displacement of the neck key point in the vertical direction with a preset displacement to determine whether the action state of the tester is in action;
when the action state is in motion, comparing the knee key point included angle with the preset standing included angle threshold value, and comparing the displacement of the neck key point in the vertical direction with the preset displacement to determine whether the action state of the tester is the action end;
when the action state is action end, taking the moment when the action state is action start as action start moment, taking the moment when the action state is action end as action end moment, and acquiring an actual action duration based on the action start moment and the action end moment;
when the actual action duration is less than or equal to a preset standard action duration, selecting a two-foot key point, a two-shoulder key point, a hip key point and the neck key point based on the bone key point to be identified, and acquiring an action detection result of the tester;
and when the action detection result is an action standard, counting the number of deep squats based on the action ending time.
2. The deep squatting detection method based on the human body key points as claimed in claim 1, wherein the determining whether the action state of the tester is action start based on the knee key point included angle and the displacement of the neck key point in the vertical direction comprises:
comparing the included angle of the knee key point with a preset standing included angle threshold value, and comparing the displacement of the neck key point in the vertical direction with a preset displacement;
and when the included angle of the knee key point is larger than a preset standing included angle threshold value and the displacement of the neck key point in the vertical direction is smaller than or equal to a preset displacement, determining that the action state of the tester is the action start.
3. The deep squatting detection method based on the key points of the human body as claimed in claim 1, wherein the comparing the knee key point included angle with a preset standing included angle threshold value, a preset exercise included angle threshold value and a preset squatting included angle threshold value respectively, and comparing the displacement of the neck key point in the vertical direction with a preset displacement to determine whether the action state of the tester is in action comprises:
if the included angle of the knee key point is larger than the preset motion included angle threshold value and smaller than or equal to the preset standing included angle threshold value, and the displacement of the neck key point in the vertical direction is larger than the preset displacement, determining that the action state of the tester in the corresponding original frame image is in descending action;
when the action state is descending, if the included angle of the knee key point is greater than the preset squat included angle threshold and less than or equal to the preset movement included angle threshold, and the displacement of the neck key point in the vertical direction is greater than the preset displacement, determining that the action state of the tester in the corresponding original frame image is squat action;
when the action state is squatting, if knee key point included angle is greater than preset motion included angle threshold value and less than or equal to preset standing included angle threshold value, the displacement of neck key point in vertical direction is greater than preset displacement, then confirm the action state of tester in corresponding original frame image is in the action of rising.
4. The deep squatting detection method based on the human body key points as claimed in claim 1, wherein the step of selecting the key points of feet, shoulders, buttocks and neck based on the bone key points to be identified to obtain the action detection result of the tester comprises the steps of:
calculating the distance between the key points of the feet based on the key points of the feet, and calculating the distance between the key points of the shoulders based on the key points of the shoulders;
calculating the distance difference between the distance between the two-foot key points and the distance between the two-shoulder key points, and taking the absolute value of the distance difference as a first distance;
calculating a second distance based on the hip keypoints and the neck keypoints;
and when the first distance is smaller than or equal to a first calculation threshold value and the second distance is larger than or equal to a second calculation threshold value, acquiring an action detection result of the action standard.
5. The human body key point-based deep squatting detection method as claimed in claim 1, wherein the obtaining of original frame images, identifying each original frame image according to the time sequence of the original frame images based on a human body posture identification algorithm, and obtaining the bone key points to be identified corresponding to each tester comprises:
acquiring original frame images, and tracking each original frame image according to the time sequence of the original frame images based on a pedestrian detection algorithm to acquire a target to be identified;
carrying out face recognition on each target to be recognized based on a face recognition algorithm to obtain identity information;
and identifying each target to be identified based on a human body posture identification algorithm to obtain the bone key points to be identified corresponding to each tester.
6. The deep squatting detection method based on the human body key points as claimed in claim 5, wherein the obtaining of original frame images, based on a pedestrian detection algorithm, performing target tracking on each original frame image according to the time sequence of the original frame images to obtain a target to be identified comprises:
acquiring original frame images, and preprocessing each pair of original frame images according to the time sequence of the original frame images to acquire effective frame images;
and tracking the target of each effective frame image based on a pedestrian detection algorithm to obtain the target to be identified.
7. The deep squat detection method based on human body key points as claimed in claim 5, wherein the deep squat detection method further comprises:
according to the time sequence, taking the first action starting time as the motion starting time;
based on the motion starting time, counting the number of squats within the preset motion duration, based on the identity information, selecting a corresponding scoring standard to score the number of squats, acquiring the motion score of the tester, and determining whether the motion of the tester reaches the standard or not based on the motion score.
8. The utility model provides a detection device squats deeply based on human key point which characterized in that includes:
the bone key point acquisition module to be identified is used for acquiring original frame images, identifying each original frame image according to the time sequence of the original frame images based on a human body posture identification algorithm, and acquiring the bone key points to be identified corresponding to each tester;
the bone key point processing module is used for acquiring the included angle of knee key points and the displacement of neck key points in the vertical direction based on the bone key points to be identified;
the action starting judging module is used for determining whether the action state of the tester is action starting or not based on the knee key point included angle and the displacement of the neck key point in the vertical direction;
the motion in-motion judging module is used for comparing the included angle of the knee key point with a preset standing included angle threshold value, a preset motion included angle threshold value and a preset squatting included angle threshold value respectively when the motion state is motion start, comparing the displacement of the neck key point in the vertical direction with a preset displacement, and determining whether the motion state of the tester is in motion or not;
the action ending judging module is used for comparing the included angle of the knee key point with the preset standing included angle threshold value when the action state is in motion, comparing the displacement of the neck key point in the vertical direction with the preset displacement, and determining whether the action state of the tester is the action ending;
the action duration calculation module is used for taking the moment when the action state is the action start moment as the action start moment when the action state is the action end, taking the moment when the action state is the action end as the action end moment, and acquiring the actual action duration based on the action start moment and the action end moment;
the action detection result acquisition module is used for selecting a two-foot key point, a two-shoulder key point, a hip key point and the neck key point based on the bone key point to be identified when the actual action duration is less than or equal to a preset standard action duration, and acquiring an action detection result of the tester;
and the deep squatting number counting module is used for counting the deep squatting number based on the action ending moment when the action detection result is the action standard.
9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor when executing the computer program implements the human keypoint-based squat detection method as claimed in any one of claims 1 to 7.
10. A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, implements the human keypoint-based deep squat detection method as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010377914.5A CN111275032B (en) | 2020-05-07 | 2020-05-07 | Deep squatting detection method, device, equipment and medium based on human body key points |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010377914.5A CN111275032B (en) | 2020-05-07 | 2020-05-07 | Deep squatting detection method, device, equipment and medium based on human body key points |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111275032A true CN111275032A (en) | 2020-06-12 |
| CN111275032B CN111275032B (en) | 2020-09-15 |
Family
ID=70999513
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010377914.5A Active CN111275032B (en) | 2020-05-07 | 2020-05-07 | Deep squatting detection method, device, equipment and medium based on human body key points |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111275032B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112149602A (en) * | 2020-09-30 | 2020-12-29 | 广州华多网络科技有限公司 | Action counting method and device, electronic equipment and storage medium |
| CN112163479A (en) * | 2020-09-16 | 2021-01-01 | 广州华多网络科技有限公司 | Motion detection method, motion detection device, computer equipment and computer-readable storage medium |
| CN112184664A (en) * | 2020-09-27 | 2021-01-05 | 杭州依图医疗技术有限公司 | Vertebra detection method and computer equipment |
| CN112966597A (en) * | 2021-03-04 | 2021-06-15 | 山东云缦智能科技有限公司 | Human motion action counting method based on skeleton key points |
| CN113255623A (en) * | 2021-07-14 | 2021-08-13 | 北京壹体科技有限公司 | System and method for intelligently identifying push-up action posture completion condition |
| CN113378758A (en) * | 2021-06-24 | 2021-09-10 | 首都师范大学 | Action recognition method and device, electronic equipment and storage medium |
| CN113486757A (en) * | 2021-06-29 | 2021-10-08 | 北京科技大学 | Multi-person linear running test timing method based on human skeleton key point detection |
| CN113902084A (en) * | 2020-07-06 | 2022-01-07 | 阿里体育有限公司 | Motion counting method and device, electronic equipment and computer storage medium |
| CN114191804A (en) * | 2021-12-08 | 2022-03-18 | 上海影谱科技有限公司 | Deep-learning-based method and device for judging whether deep-squatting posture is standard or not |
| CN114596451A (en) * | 2022-04-01 | 2022-06-07 | 此刻启动(北京)智能科技有限公司 | Body fitness testing method and device based on AI vision and storage medium |
| CN114783059A (en) * | 2022-04-20 | 2022-07-22 | 浙江东昊信息工程有限公司 | Temple incense and worship participation management method and system based on depth camera |
| CN117316377A (en) * | 2023-10-20 | 2023-12-29 | 深圳咕嘟熊教育科技有限责任公司 | Infant outdoor exercises health data acquisition system and intelligent terminal |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101964047A (en) * | 2009-07-22 | 2011-02-02 | 深圳泰山在线科技有限公司 | Multiple trace point-based human body action recognition method |
| CN108256433A (en) * | 2017-12-22 | 2018-07-06 | 银河水滴科技（北京）有限公司 | A kind of athletic posture appraisal procedure and system |
| CN109192267A (en) * | 2018-08-09 | 2019-01-11 | 深圳狗尾草智能科技有限公司 | Virtual robot is accompanied in movement |
| CN109214231A (en) * | 2017-06-29 | 2019-01-15 | 深圳泰山体育科技股份有限公司 | Physical education auxiliary system and method based on human body attitude identification |
| CN109815907A (en) * | 2019-01-25 | 2019-05-28 | 深圳市象形字科技股份有限公司 | A kind of sit-ups attitude detection and guidance method based on computer vision technique |
| CN109934111A (en) * | 2019-02-12 | 2019-06-25 | 清华大学深圳研究生院 | A kind of body-building Attitude estimation method and system based on key point |
| US20190366154A1 (en) * | 2018-05-31 | 2019-12-05 | Microsoft Technology Licensing, Llc | Physical activity training assistant |
| CN110559645A (en) * | 2019-07-18 | 2019-12-13 | 华为技术有限公司 | Application operation method and electronic equipment |
| CN110969114A (en) * | 2019-11-28 | 2020-04-07 | 四川省骨科医院 | Human body action function detection system, detection method and detector |
-
2020
- 2020-05-07 CN CN202010377914.5A patent/CN111275032B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101964047A (en) * | 2009-07-22 | 2011-02-02 | 深圳泰山在线科技有限公司 | Multiple trace point-based human body action recognition method |
| CN109214231A (en) * | 2017-06-29 | 2019-01-15 | 深圳泰山体育科技股份有限公司 | Physical education auxiliary system and method based on human body attitude identification |
| CN108256433A (en) * | 2017-12-22 | 2018-07-06 | 银河水滴科技（北京）有限公司 | A kind of athletic posture appraisal procedure and system |
| US20190366154A1 (en) * | 2018-05-31 | 2019-12-05 | Microsoft Technology Licensing, Llc | Physical activity training assistant |
| CN109192267A (en) * | 2018-08-09 | 2019-01-11 | 深圳狗尾草智能科技有限公司 | Virtual robot is accompanied in movement |
| CN109815907A (en) * | 2019-01-25 | 2019-05-28 | 深圳市象形字科技股份有限公司 | A kind of sit-ups attitude detection and guidance method based on computer vision technique |
| CN109934111A (en) * | 2019-02-12 | 2019-06-25 | 清华大学深圳研究生院 | A kind of body-building Attitude estimation method and system based on key point |
| CN110559645A (en) * | 2019-07-18 | 2019-12-13 | 华为技术有限公司 | Application operation method and electronic equipment |
| CN110969114A (en) * | 2019-11-28 | 2020-04-07 | 四川省骨科医院 | Human body action function detection system, detection method and detector |
Non-Patent Citations (2)
| Title |
|---|
| NANTANA RUNGSAWASDISAP 等: "Squat Movement Recognition Using Hidden Markov Models", 《2018 INTERNATIONAL WORKSHOP ON ADVANCED IMAGE TECHNOLOGY (IWAIT)》 * |
| 喻露 等: "基于人体骨架的非标准深蹲姿势检测方法", 《计算机应用》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113902084A (en) * | 2020-07-06 | 2022-01-07 | 阿里体育有限公司 | Motion counting method and device, electronic equipment and computer storage medium |
| CN112163479A (en) * | 2020-09-16 | 2021-01-01 | 广州华多网络科技有限公司 | Motion detection method, motion detection device, computer equipment and computer-readable storage medium |
| CN112184664A (en) * | 2020-09-27 | 2021-01-05 | 杭州依图医疗技术有限公司 | Vertebra detection method and computer equipment |
| CN112184664B (en) * | 2020-09-27 | 2023-05-26 | 杭州依图医疗技术有限公司 | Vertebra detection method and computer equipment |
| CN112149602A (en) * | 2020-09-30 | 2020-12-29 | 广州华多网络科技有限公司 | Action counting method and device, electronic equipment and storage medium |
| CN112966597A (en) * | 2021-03-04 | 2021-06-15 | 山东云缦智能科技有限公司 | Human motion action counting method based on skeleton key points |
| CN113378758A (en) * | 2021-06-24 | 2021-09-10 | 首都师范大学 | Action recognition method and device, electronic equipment and storage medium |
| CN113486757A (en) * | 2021-06-29 | 2021-10-08 | 北京科技大学 | Multi-person linear running test timing method based on human skeleton key point detection |
| CN113255623A (en) * | 2021-07-14 | 2021-08-13 | 北京壹体科技有限公司 | System and method for intelligently identifying push-up action posture completion condition |
| CN114191804A (en) * | 2021-12-08 | 2022-03-18 | 上海影谱科技有限公司 | Deep-learning-based method and device for judging whether deep-squatting posture is standard or not |
| CN114596451A (en) * | 2022-04-01 | 2022-06-07 | 此刻启动(北京)智能科技有限公司 | Body fitness testing method and device based on AI vision and storage medium |
| CN114783059A (en) * | 2022-04-20 | 2022-07-22 | 浙江东昊信息工程有限公司 | Temple incense and worship participation management method and system based on depth camera |
| CN117316377A (en) * | 2023-10-20 | 2023-12-29 | 深圳咕嘟熊教育科技有限责任公司 | Infant outdoor exercises health data acquisition system and intelligent terminal |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111275032B (en) | 2020-09-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111275032B (en) | Deep squatting detection method, device, equipment and medium based on human body key points | |
| CN111275031B (en) | Flat plate support detection method, device, equipment and medium based on human body key points | |
| CN110495889B (en) | Posture evaluation method, electronic device, computer device, and storage medium | |
| CN108205654B (en) | Action detection method and device based on video | |
| CN110969114A (en) | Human body action function detection system, detection method and detector | |
| CN111062239A (en) | Human body target detection method and device, computer equipment and storage medium | |
| CN111753747B (en) | Violent motion detection method based on monocular camera and three-dimensional attitude estimation | |
| CN112633196A (en) | Human body posture detection method and device and computer equipment | |
| JP2021064367A (en) | Motion recognition method, motion recognition device, and electronic device | |
| US20220270387A1 (en) | Modeling method and modeling device for human body model, electronic device, and storage medium | |
| CN113392741A (en) | Video clip extraction method and device, electronic equipment and storage medium | |
| CN112568898A (en) | Method, device and equipment for automatically evaluating injury risk and correcting motion of human body motion based on visual image | |
| CN113191200A (en) | Push-up test counting method, device, equipment and medium | |
| CN113557546B (en) | Method, device, equipment and storage medium for detecting associated objects in image | |
| CN114092863A (en) | Human body motion evaluation method for multi-view video image | |
| CN113392743A (en) | Abnormal action detection method, abnormal action detection device, electronic equipment and computer storage medium | |
| CN117037278A (en) | Jump floor error assessment method and apparatus | |
| CN111353345B (en) | Method, apparatus, system, electronic device, and storage medium for providing training feedback | |
| CN113842622B (en) | Motion teaching method, device, system, electronic equipment and storage medium | |
| CN114821803A (en) | Human motion scoring method and device, computer equipment and storage medium | |
| CN115240856A (en) | Exercise health assessment method, system and equipment based on exercise posture | |
| WO2022110453A1 (en) | Running posture recognition method and apparatus, and computer device | |
| CN111369532A (en) | Method and device for processing mammary gland X-ray image | |
| CN112233769A (en) | Recovery system after suffering from illness based on data acquisition | |
| CN111460868A (en) | Action recognition error correction method, system, electronic device and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |