CN115588231A - Live broadcast follow-up video detection method, device, equipment and storage medium - Google Patents

Live broadcast follow-up video detection method, device, equipment and storage medium Download PDF

Info

Publication number
CN115588231A
CN115588231A CN202211093077.9A CN202211093077A CN115588231A CN 115588231 A CN115588231 A CN 115588231A CN 202211093077 A CN202211093077 A CN 202211093077A CN 115588231 A CN115588231 A CN 115588231A
Authority
CN
China
Prior art keywords
video
standard
follow
live
angle
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
Application number
CN202211093077.9A
Other languages
Chinese (zh)
Inventor
贾泽华
卓钰博
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ping An Technology Shenzhen Co Ltd
Original Assignee
Ping An Technology Shenzhen Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Ping An Technology Shenzhen Co Ltd filed Critical Ping An Technology Shenzhen Co Ltd
Priority to CN202211093077.9A priority Critical patent/CN115588231A/en
Publication of CN115588231A publication Critical patent/CN115588231A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
    • G06V20/42—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items of sport video content
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/48—Matching video sequences
    • 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)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Health & Medical Sciences (AREA)
  • Psychiatry (AREA)
  • Social Psychology (AREA)
  • Human Computer Interaction (AREA)
  • Image Analysis (AREA)

Abstract

The invention relates to artificial intelligence and provides a live broadcast follow-up video detection method, a live broadcast follow-up video detection device, live broadcast follow-up video detection equipment and a storage medium. The method includes the steps of obtaining a live motion video and a follow-up video, calculating playing frequency according to first video duration of the live motion video and second video duration of the follow-up video, extracting a standard video sequence based on the playing frequency, extracting a follow-up action sequence, detecting a standard key point of each standard video frame and a user key point of each follow-up frame, identifying a standard action angle of each standard video frame based on the standard key points, identifying a follow-up action angle of each follow-up frame based on the user key points, and accurately generating a follow-up rating according to the standard action angle and the follow-up action angle. In addition, the invention also relates to a block chain technology, and the following training rating can be stored in the block chain.

Description

Live broadcast follow-up video detection method, device, equipment and storage medium
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a live broadcast follow-up video detection method, a live broadcast follow-up video detection device, live broadcast follow-up video detection equipment and a storage medium.
Background
At present, in a live sports platform, due to the fact that playing frequency control cannot be carried out on a live sports video, when training actions of a user cannot follow coaching actions in the live sports video, the problem that a follow-up video of the user cannot be matched with the live sports video successfully exists, and the situation of the user's sports in the follow-up video cannot be accurately evaluated.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a live-follow video detection method, apparatus, device and storage medium, which can solve the technical problem that the motion condition of the user in the follow video cannot be accurately estimated.
On one hand, the invention provides a live follow-up video detection method, which comprises the following steps:
acquiring a live motion video, and acquiring a follow-up video based on the live motion video;
calculating the playing frequency of the live motion video according to the first video time length of the live motion video and the second video time length of the follow-up video;
extracting a standard video sequence from the live motion video based on the playing frequency, and extracting a follow-up action sequence from the follow-up video;
detecting a standard key point of each standard video frame in the standard video sequence, and detecting a user key point of each follow-up frame in the follow-up action sequence;
identifying a standard action angle of each standard video frame based on the standard key points, and identifying a follow-up action angle of each follow-up frame based on the user key points;
and generating a follow-up rating of the follow-up video according to the standard action angle and the follow-up action angle.
According to a preferred embodiment of the present invention, the calculating the play frequency of the live motion video according to the first video duration of the live motion video and the second video duration of the follow-through video includes:
identifying a starting generation time point of a starting video frame in the live broadcast motion video, and identifying an ending generation time point of an ending video frame in the live broadcast motion video;
calculating a difference value between the generation ending time point and the generation starting time point to obtain the first video time length;
and calculating the ratio of the second video time length to the first video time length to obtain the playing frequency.
According to a preferred embodiment of the present invention, the extracting a standard video sequence from the live motion video based on the play frequency includes:
comparing the playing frequency with a preset frequency;
if the playing frequency is greater than the preset frequency, converting the live broadcast motion video based on the playing frequency to obtain a standard motion video, and generating a standard video sequence according to a standard video frame in the standard motion video and a frame position of the standard video frame in the standard motion video; or
And if the playing frequency is less than or equal to the preset frequency, extracting the standard video frame from the live motion video based on the playing frequency to serve as the standard video sequence.
According to a preferred embodiment of the present invention, the detecting the standard keypoints of each standard video frame in the standard video sequence comprises:
for each standard video frame, acquiring the pixel value and the pixel position of each pixel point in the standard video frame;
detecting the pixel values and the pixel positions based on a pre-trained human body detection model to obtain a detection category corresponding to each pixel point and a detection probability of the pixel point belonging to the detection category;
screening a target pixel point from the plurality of pixel points based on the detection probability;
and determining the area formed by adjacent pixel points corresponding to the same detection type in the target pixel points as the standard key point based on the pixel position and the detection type.
According to a preferred embodiment of the present invention, the identifying the standard action angle of each standard video frame based on the standard key points comprises:
constructing a plane rectangular coordinate system by taking any standard key point as an origin and taking an image edge parallel to any standard video frame as a coordinate axis;
identifying a key point coordinate value corresponding to the standard key point according to the pixel position corresponding to the standard key point and the plane rectangular coordinate system;
acquiring a connection key point pair of any standard key point from a plurality of standard key points, wherein the connection key point pair comprises a first connection point and a second connection point;
determining a connecting edge formed by any standard key point and the first connecting point as a first connecting edge, and calculating an included angle between the first connecting edge and the coordinate axis based on the key point coordinate value of the first connecting point to obtain a first angle;
determining a connecting edge formed by any standard key point and the second connecting point as a second connecting edge, and calculating an included angle between the second connecting edge and the coordinate axis based on the key point coordinate value of the second connecting point to obtain a second angle;
and generating the standard action angle based on the first angle, the second angle and a preset integer value.
According to the preferred embodiment of the present invention, the calculation formula of the standard action angle is:
θ＝β-α±(2kπ)；
wherein θ represents the standard operation angle, β represents the second angle, α represents the first angle, k is the predetermined integer value, and θ >0.
According to a preferred embodiment of the present invention, the generating a follow-up rating of the follow-up video according to the standard action angle and the follow-up action angle includes:
acquiring a key point weight threshold corresponding to any standard key point;
generating the contact ratio of each follow-up frame according to the standard action angle, the follow-up action angle and the corresponding key point weight threshold, wherein the calculation formula of the contact ratio is as follows:
Figure BDA0003837788210000031
wherein S represents the degree of coincidence, W = [ W = 1 ,w 2 ,…,w n ]Representing keypoint weight thresholds corresponding to a plurality of said canonical keypoint, M = [ M = 1 ,m 2 ,…,m n ]Represents a plurality of the standard operation angles, T = [ T ] 1 ,t 2 ,…,t n ]Representing a plurality of the follow-up action angles;
and generating the follow-up rating according to the contact ratio and a preset mapping relation.
On the other hand, the invention also provides a live broadcast follow-up video detection device, which comprises:
the device comprises an acquisition unit, a tracking unit and a tracking unit, wherein the acquisition unit is used for acquiring a live motion video and acquiring a follow-up video based on the live motion video;
the calculation unit is used for calculating the playing frequency of the live motion video according to the first video time length of the live motion video and the second video time length of the follow-up video;
the extraction unit is used for extracting a standard video sequence from the live motion video based on the playing frequency and extracting a follow-up action sequence from the follow-up video;
the detection unit is used for detecting a standard key point of each standard video frame in the standard video sequence and detecting a user key point of each follow-up frame in the follow-up action sequence;
the identification unit is used for identifying the standard action angle of each standard video frame based on the standard key points and identifying the follow-up action angle of each follow-up frame based on the user key points;
and the generation unit is used for generating a follow-up rating of the follow-up video according to the standard action angle and the follow-up action angle.
In another aspect, the present invention further provides an electronic device, including:
a memory storing computer readable instructions; and
a processor that executes computer readable instructions stored in the memory to implement the live follow-through video detection method.
In another aspect, the present invention further provides a computer-readable storage medium, where computer-readable instructions are stored in the computer-readable storage medium and executed by a processor in an electronic device to implement the live follow-through video detection method.
According to the technical scheme, the time length of the first video reaches the time length of the second video can be accurate and quantified the play frequency, and then based on the play frequency can ensure the standard video sequence and the matching relation of the exercise action sequence, so that the matching problem caused by the fact that the exercise action of a user cannot follow the exercise action in the live motion video can be avoided, and the accuracy of the exercise rating is improved.
Drawings
Fig. 1 is a flowchart of a live follow-through video detection method according to a preferred embodiment of the present invention.
FIG. 2 is a visual representation of a standard action angle in the present invention.
Fig. 3 is a functional block diagram of a live follow-through video detection apparatus according to a preferred embodiment of the present invention.
Fig. 4 is a schematic structural diagram of an electronic device implementing a live-action follow-through video detection method according to a preferred embodiment 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 will be described in detail with reference to the accompanying drawings and specific embodiments.
Fig. 1 is a flow chart of a live follow-through video detection method according to a preferred embodiment of the present invention. The order of the steps in the flow chart may be changed and some steps may be omitted according to different needs.
The live follow-up video detection method can acquire and process related data based on an artificial intelligence technology. Among them, artificial Intelligence (AI) is a theory, method, technique and application system that simulates, extends and expands human Intelligence using a digital computer or a machine controlled by a digital computer, senses the environment, acquires knowledge and uses the knowledge to obtain the best result.
The artificial intelligence infrastructure generally includes technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a robot technology, a biological recognition technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like.
The live broadcast follow-up video detection method is applied to one or more electronic devices, wherein the electronic devices are devices capable of automatically performing numerical calculation and/or information processing according to computer readable instructions which are set or stored in advance, and hardware of the electronic devices includes but is not limited to a microprocessor, an Application Specific Integrated Circuit (ASIC), a Programmable Gate Array (FPGA), a Digital Signal Processor (DSP), an embedded device and the like.
The electronic device may be any electronic product capable of performing human-computer interaction with a user, for example, a Personal computer, a tablet computer, a smart phone, a Personal Digital Assistant (PDA), a game machine, an interactive Internet Protocol Television (IPTV), a smart wearable device, and the like.
The electronic device may include a network device and/or a user device. Wherein the network device includes, but is not limited to, a single network electronic device, an electronic device group consisting of a plurality of network electronic devices, or a Cloud Computing (Cloud Computing) based Cloud consisting of a large number of hosts or network electronic devices.
The network in which the electronic device is located includes, but is not limited to: the internet, a wide area Network, a metropolitan area Network, a local area Network, a Virtual Private Network (VPN), and the like.
101, acquiring a live motion video and acquiring a follow-up video based on the live motion video.
In at least one embodiment of the present invention, the live motion video is typically a tutorial video taken by a motion coach for a certain motion. The live motion video refers to a video played in real time.
The follow-up video refers to a training video generated after a training user performs action simulation based on the live motion video.
102, calculating the playing frequency of the live motion video according to the first video time length of the live motion video and the second video time length of the follow-up video.
In at least one embodiment of the present invention, the first video duration refers to a total playing duration of the live motion video, and the second video duration refers to a total playing duration of the follow-up video. The playing frequency refers to a ratio of the duration of the second video to the duration of the first video.
In at least one embodiment of the present invention, the calculating, by the electronic device, the playing frequency of the live motion video according to the first video duration of the live motion video and the second video duration of the follow-up video includes:
identifying a starting generation time point of a starting video frame in the live motion video, and identifying an ending generation time point of an ending video frame in the live motion video;
calculating a difference value between the generation ending time point and the generation starting time point to obtain the first video time length;
and calculating the ratio of the second video time length to the first video time length to obtain the playing frequency.
The starting video frame refers to a first video frame in the live motion video, and the last video frame in the ending video frame. The starting generation time point is a time point generated by shooting the starting video frame, and the ending generation time point is a time point generated by shooting the ending video frame.
The first video time length can be accurately counted through the ending generation time point and the starting generation time point, so that the accuracy of the playing frequency can be improved by combining the second video time length and the first video time length.
Specifically, the generation manner of the second video duration is similar to the generation manner of the first video duration, and details thereof are not repeated herein.
In at least one embodiment of the invention, the method further comprises:
acquiring a demand frequency;
converting the live broadcast motion video based on the demand frequency to obtain a target motion video;
and playing the target motion video.
Wherein the demand frequency is used to indicate a demand of the training user for a play speed of the live sports video.
Through the requirement frequency, the live sports video is converted, the problem that a user cannot learn action details from the live sports video can be avoided, and the problem that the user cannot keep up with coach actions can also be avoided.
103, extracting a standard video sequence from the live motion video based on the playing frequency, and extracting a follow-up action sequence from the follow-up video.
In at least one embodiment of the present invention, the standard video sequence may be a part of video frames or all video frames in the live motion video, the follow-up action sequence refers to video frames corresponding to the standard video sequence, and the number of video frames in the standard video sequence is equal to the number of video frames in the follow-up action sequence.
In at least one embodiment of the invention, the electronic device extracting a standard video sequence from the live motion video based on the play frequency comprises:
comparing the playing frequency with a preset frequency;
if the playing frequency is greater than the preset frequency, converting the live broadcast motion video based on the playing frequency to obtain a standard motion video, and generating the standard video sequence according to a standard video frame in the standard motion video and a frame position of the standard video frame in the standard motion video; or
And if the playing frequency is less than or equal to the preset frequency, extracting the standard video frame from the live motion video based on the playing frequency to serve as the standard video sequence.
Wherein the preset frequency is usually set to 1.
Through the relationship between the playing frequency and the preset frequency, the standard video sequence can be generated in different modes, and the generation accuracy of the standard video sequence is improved.
In at least one embodiment of the present invention, the electronic device generates the follow-up action sequence according to a follow-up video frame in the follow-up video and a frame position of the follow-up video frame in the follow-up video.
104, detecting the standard key points of each standard video frame in the standard video sequence, and detecting the user key points of each follow-up frame in the follow-up action sequence.
In at least one embodiment of the present invention, the standard key points refer to human pose key points of the sport coach, and the user key points refer to human pose key points of the training user. The standard key points reach the user key points respectively include human posture key points such as a nose, a left eye, a right eye, a left ear, a right ear, a left shoulder, a right shoulder, a left elbow, a right elbow, a left wrist, a right wrist, a left hip joint, a right hip joint, a left knee, a right knee, a left ankle, a right ankle and the like.
In at least one embodiment of the invention, the electronic device detecting the canonical keypoints for each canonical video frame in the canonical video sequence comprises:
for each standard video frame, acquiring the pixel value and the pixel position of each pixel point in the standard video frame;
detecting the pixel values and the pixel positions based on a pre-trained human body detection model to obtain a detection category corresponding to each pixel point and a detection probability of the pixel point belonging to the detection category;
screening target pixel points from the plurality of pixel points based on the detection probability;
and determining an area formed by adjacent pixel points corresponding to the same detection category in the target pixel points as the standard key point based on the pixel position and the detection category.
The human body detection model is generated by training and testing the crawled pictures related to the people. The detection categories may include, but are not limited to: nose, left eye, right eye, left ear, right ear, left shoulder, right shoulder, left elbow, right elbow, left wrist, right wrist, left hip joint, right hip joint, left knee, right knee, left ankle, right ankle, and the like.
The target pixel points are pixel points with the detection probability being larger than a preset probability, and the preset probability is set according to actual requirements.
The adjacent pixel points refer to target pixel points adjacent to the pixel positions.
By combining the pixel values and the pixel positions, the detection categories and the corresponding detection probabilities can be accurately identified, interference pixel points can be extracted based on the detection probabilities, so that the screening accuracy of the target pixel points is improved, and the identification accuracy of the standard key points can be improved by further analyzing adjacent pixel points of the target pixel points.
Specifically, the generation manner of the user key points is similar to the generation manner of the standard key points, and details are not repeated herein.
And 105, identifying the standard action angle of each standard video frame based on the standard key points, and identifying the follow-up action angle of each follow-up frame based on the user key points.
In at least one embodiment of the present invention, the standard motion angle refers to an angle formed by a first connection edge and a second connection edge, where the first connection edge refers to a connection edge formed by any standard key point and a first connection point, the second connection edge refers to a connection edge formed by any standard key point and a second connection point, and the first connection point and the second connection point refer to the remaining standard key points connected to the any standard key point respectively.
In at least one embodiment of the invention, the electronic device identifying the standard action angle of each standard video frame based on the standard key point comprises:
constructing a plane rectangular coordinate system by taking any standard key point as an origin and taking an image edge parallel to any standard video frame as a coordinate axis;
identifying a key point coordinate value corresponding to the standard key point according to the pixel position corresponding to the standard key point and the plane rectangular coordinate system;
acquiring a connection key point pair of any standard key point from a plurality of standard key points, wherein the connection key point pair comprises a first connection point and a second connection point;
determining a connecting edge formed by any standard key point and the first connecting point as a first connecting edge, and calculating an included angle between the first connecting edge and the coordinate axis based on the key point coordinate value of the first connecting point to obtain a first angle;
determining a connecting edge formed by any standard key point and the second connecting point as a second connecting edge, and calculating an included angle between the second connecting edge and the coordinate axis based on the key point coordinate value of the second connecting point to obtain a second angle;
and generating the standard action angle based on the first angle, the second angle and a preset integer value.
For example, if the horizontal pixel position of the standard key point a is the 8 th pixel point, and the vertical pixel position of the standard key point a is the 5 th pixel point, the key point coordinate value is (8, 5).
The preset integer value is used for controlling the value of the standard action angle, when the first angle is larger than the second angle, the preset integer value is larger than 0, and when the first angle is smaller than or equal to the second angle, the preset integer value is 0. For example, if α =90 °, β =315 °, θ = β - α =225 °. If α =90 °, β =45 °, θ = β - α +2k pi =315 °.
The first angle and the second angle can be accurately calculated through the coordinate values of the key points, and then the accuracy of the standard action angle can be improved through the control of the preset integer value on the standard action angle.
As shown in fig. 2, fig. 2 is a visual diagram of a standard action angle in the present invention. Where α in fig. 2 represents the first angle, β represents the second angle, and θ represents the standard action angle.
In this embodiment, the accuracy of the standard operation angle can be improved by controlling the first angle and the second angle to have the same rotation direction.
Specifically, the calculation formula of the first angle is as follows:
α＝arccos(c)；
α＝arcsin(s)；
where α represents the first angle, c represents a lateral coordinate value among the key point coordinate values of the first connection point, and s represents a longitudinal coordinate value among the key point coordinate values of the first connection point.
By combining the lateral coordinate value and the longitudinal coordinate value, the accuracy of determining the first angle can be improved.
Specifically, the calculation manner of the second angle is similar to that of the first angle, and details thereof are not repeated herein.
Specifically, the calculation formula of the standard action angle is as follows:
θ＝β-α±(2kπ)；
wherein θ represents the standard action angle, β represents the second angle, α represents the first angle, k is the preset integer value, and θ >0.
Through the adjustment of the preset integral value to the standard action angle, the accuracy of the standard action angle can be improved.
In at least one embodiment of the present invention, the recognition manner of the follow-up exercise angle is similar to the recognition manner of the standard exercise angle, and details thereof are not repeated herein.
106, generating a follow-up rating of the follow-up video according to the standard action angle and the follow-up action angle.
It is emphasized that the follow-up rating may also be stored in a node of a blockchain in order to further ensure privacy and security of the follow-up rating.
In at least one embodiment of the present invention, the follow-up rating is used to quantify the athletic training profile of the training user. The drill rating includes, but is not limited to: fail rating, pass rating, great rating, perfect rating, and the like.
In at least one embodiment of the present invention, the electronic device generating a follow-up rating of the follow-up video according to the standard action angle and the follow-up action angle includes:
acquiring a key point weight threshold corresponding to any standard key point;
generating the contact ratio of each follow-up frame according to the standard action angle, the follow-up action angle and the corresponding key point weight threshold, wherein the calculation formula of the contact ratio is as follows:
Figure BDA0003837788210000081
wherein S represents the degree of coincidence, W = [ W = 1 ,w 2 ,…,w n ]Representing keypoint weight thresholds corresponding to a plurality of said canonical keypoint, M = [ M = 1 ,m 2 ,…,m n ]Represents a plurality of the standard operation angles, T = [ T ] 1 ,t 2 ,…,t n ]Representing a plurality of the follow-up action angles;
and generating the follow-up rating according to the contact ratio and a preset mapping relation.
The key point weight threshold and the preset mapping relation can be set according to actual requirements. For example, a degree of overlap of 0% -40% is a Fail rating; the contact ratio is 41% -60%, and the Pass rating is obtained; the contact ratio is 61-80%, and the Great rating is obtained; the contact ratio is 81% -100%, and the Perfect rating is obtained.
By setting a key point weight threshold value for each standard key point, the quantification accuracy of the contact ratio can be improved, and therefore the accuracy of the follow-up rating is improved.
Specifically, the generating, by the electronic device, the follow exercise rating according to the contact ratio and a preset mapping relationship includes:
calculating an average value of a plurality of the contact ratios;
and identifying the follow-up rating based on the average value and the preset mapping relation.
According to the technical scheme, the time of the first video reaches the time of the second video can be accurate and quantified, the playing frequency is further based on the playing frequency can be ensured, the standard video sequence can be matched with the exercise action sequence, so that the matching problem caused when the exercise action of a user cannot follow the exercise action in the live sports video can be avoided, and the accuracy of the exercise rating is improved.
Fig. 3 is a functional block diagram of a live-following video detection device according to a preferred embodiment of the present invention. The live drill video detection apparatus 11 includes an acquisition unit 110, a calculation unit 111, an extraction unit 112, a detection unit 113, a recognition unit 114, a generation unit 115, a conversion unit 116, and a playback unit 117. A module/unit as referred to herein is a series of computer readable instruction segments capable of being retrieved by the processor 13 and performing a fixed function, and stored in the memory 12. In the present embodiment, the functions of the modules/units will be described in detail in the following embodiments.
The acquisition unit 110 acquires a live motion video, and acquires a follow-up video based on the live motion video.
In at least one embodiment of the present invention, the live motion video is typically a tutorial video taken by a motion coach for a certain motion. The live motion video refers to a video played in real time.
The follow-up video refers to a training video generated after a training user performs action simulation based on the live motion video.
The calculating unit 111 calculates the playing frequency of the live motion video according to the first video time length of the live motion video and the second video time length of the follow-up video.
In at least one embodiment of the present invention, the first video duration refers to a total playing duration of the live motion video, and the second video duration refers to a total playing duration of the follow-up video. The playing frequency refers to a ratio of the duration of the second video to the duration of the first video.
In at least one embodiment of the present invention, the calculating unit 111 calculates the playing frequency of the live motion video according to the first video duration of the live motion video and the second video duration of the follow-up video, where the calculating unit includes:
identifying a starting generation time point of a starting video frame in the live broadcast motion video, and identifying an ending generation time point of an ending video frame in the live broadcast motion video;
calculating a difference value between the generation ending time point and the generation starting time point to obtain the first video time length;
and calculating the ratio of the second video time length to the first video time length to obtain the playing frequency.
The starting video frame refers to a first video frame in the live motion video, and the last video frame in the ending video frame. The starting generation time point is a time point generated by shooting the starting video frame, and the ending generation time point is a time point generated by shooting the ending video frame.
The first video time length can be accurately counted through the generation ending time point and the generation starting time point, and therefore the accuracy of the playing frequency can be improved by combining the second video time length and the first video time length.
Specifically, the generation manner of the second video duration is similar to the generation manner of the first video duration, and details thereof are not repeated herein.
In at least one embodiment of the present invention, the obtaining unit 110 obtains a demand frequency;
the conversion unit 116 performs conversion processing on the live motion video based on the required frequency to obtain a target motion video;
the playing unit 117 plays the target moving video.
Wherein the demand frequency is used to indicate a demand of the training user for a play speed of the live sports video.
Through the conversion of the live sports video by the demand frequency, the problem that a user cannot learn action details from the live sports video can be avoided, and the problem that the user cannot keep up with coach actions can also be avoided.
The extraction unit 112 extracts a standard video sequence from the live motion video based on the playback frequency, and extracts a follow-up action sequence from the follow-up video.
In at least one embodiment of the present invention, the standard video sequence may be a part of video frames or all video frames in the live motion video, the follow-up action sequence refers to video frames corresponding to the standard video sequence, and the number of video frames in the standard video sequence is equal to the number of video frames in the follow-up action sequence.
In at least one embodiment of the present invention, the extracting unit 112 extracts a standard video sequence from the live motion video based on the play frequency includes:
comparing the playing frequency with a preset frequency;
if the playing frequency is greater than the preset frequency, converting the live broadcast motion video based on the playing frequency to obtain a standard motion video, and generating a standard video sequence according to a standard video frame in the standard motion video and a frame position of the standard video frame in the standard motion video; or alternatively
And if the playing frequency is less than or equal to the preset frequency, extracting the standard video frame from the live motion video based on the playing frequency to serve as the standard video sequence.
Wherein the preset frequency is usually set to 1.
According to the relation between the playing frequency and the preset frequency, the standard video sequence can be generated in different modes, and the generation accuracy of the standard video sequence is improved.
In at least one embodiment of the present invention, the extracting unit 112 generates the follow-up action sequence according to a follow-up video frame in the follow-up video and a frame position of the follow-up video frame in the follow-up video.
The detection unit 113 detects a standard key point of each standard video frame in the standard video sequence, and detects a user key point of each follow-through frame in the follow-through action sequence.
In at least one embodiment of the present invention, the standard key points refer to human pose key points of the sport coach, and the user key points refer to human pose key points of the training user. The standard key points and the user key points respectively comprise human posture key points such as a nose, a left eye, a right eye, a left ear, a right ear, a left shoulder, a right shoulder, a left elbow, a right elbow, a left wrist, a right wrist, a left hip joint, a right hip joint, a left knee, a right knee, a left ankle and a right ankle.
In at least one embodiment of the present invention, the detecting unit 113 detects the standard keypoints of each standard video frame in the standard video sequence includes:
for each standard video frame, acquiring the pixel value and the pixel position of each pixel point in the standard video frame;
detecting the pixel values and the pixel positions based on a pre-trained human body detection model to obtain a detection category corresponding to each pixel point and a detection probability of the pixel point belonging to the detection category;
screening a target pixel point from the plurality of pixel points based on the detection probability;
and determining the area formed by adjacent pixel points corresponding to the same detection type in the target pixel points as the standard key point based on the pixel position and the detection type.
The human body detection model is generated by training and testing the crawled pictures related to the people. The detection categories may include, but are not limited to: nose, left eye, right eye, left ear, right ear, left shoulder, right shoulder, left elbow, right elbow, left wrist, right wrist, left hip joint, right hip joint, left knee, right knee, left ankle, right ankle, and the like.
The target pixel points are pixel points with the detection probability larger than a preset probability, and the preset probability is set according to actual requirements.
The adjacent pixel points refer to target pixel points adjacent to the pixel positions.
By combining the pixel values and the pixel positions, the detection categories and the corresponding detection probabilities can be accurately identified, interference pixel points can be extracted based on the detection probabilities, so that the screening accuracy of the target pixel points is improved, and the identification accuracy of the standard key points can be improved by further analyzing adjacent pixel points of the target pixel points.
Specifically, the generation manner of the user key points is similar to the generation manner of the standard key points, which is not described herein again.
The recognition unit 114 recognizes a standard action angle of each standard video frame based on the standard key points, and recognizes a follow-up action angle of each follow-up frame based on the user key points.
In at least one embodiment of the present invention, the standard motion angle refers to an angle formed by a first connection edge and a second connection edge, where the first connection edge refers to a connection edge formed by any standard key point and a first connection point, the second connection edge refers to a connection edge formed by any standard key point and a second connection point, and the first connection point and the second connection point refer to the remaining standard key points connected to the any standard key point respectively.
In at least one embodiment of the present invention, the identifying unit 114 identifies the standard motion angle of each standard video frame based on the standard key points comprises:
constructing a plane rectangular coordinate system by taking any standard key point as an origin and taking an image edge parallel to any standard video frame as a coordinate axis;
identifying a key point coordinate value corresponding to the standard key point according to the pixel position corresponding to the standard key point and the plane rectangular coordinate system;
acquiring a connection key point pair of any standard key point from a plurality of standard key points, wherein the connection key point pair comprises a first connection point and a second connection point;
determining a connecting edge formed by any standard key point and the first connecting point as a first connecting edge, and calculating an included angle between the first connecting edge and the coordinate axis based on the key point coordinate value of the first connecting point to obtain a first angle;
determining a connecting edge formed by any standard key point and the second connecting point as a second connecting edge, and calculating an included angle between the second connecting edge and the coordinate axis based on the key point coordinate value of the second connecting point to obtain a second angle;
and generating the standard action angle based on the first angle, the second angle and a preset integer value.
The key point coordinate value refers to a coordinate value of the standard key point in the planar rectangular coordinate system, for example, if the horizontal pixel position of the standard key point a is the 8 th pixel point, and the vertical pixel position is the 5 th pixel point, the key point coordinate value is (8, 5).
The preset integer value is used for controlling a value of the standard action angle, when the first angle is larger than the second angle, the preset integer value is larger than 0, and when the first angle is smaller than or equal to the second angle, the preset integer value is 0. For example, if α =90 °, β =315 °, then θ = β - α =225 °. If α =90 °, β =45 °, θ = β - α +2k pi =315 °.
The first angle and the second angle can be accurately calculated through the coordinate values of the key points, and then the accuracy of the standard action angle can be improved through the control of the preset integer value on the standard action angle.
As shown in FIG. 2, FIG. 2 is a visual diagram of a standard action angle in the present invention. Where α in fig. 2 represents the first angle, β represents the second angle, and θ represents the standard action angle.
In this embodiment, the first angle and the second angle are controlled to have the same rotation direction, so that the accuracy of the standard operation angle can be improved.
Specifically, the calculation formula of the first angle is as follows:
α＝arccos(c)；
α＝arcsin(s)；
where α represents the first angle, c represents a lateral coordinate value of the first connection point's keypoint coordinate values, and s represents a longitudinal coordinate value of the first connection point's keypoint coordinate values.
By combining the lateral coordinate value and the longitudinal coordinate value, the accuracy of determining the first angle can be improved.
Specifically, the calculation manner of the second angle is similar to that of the first angle, and details thereof are not repeated herein.
Specifically, the calculation formula of the standard action angle is as follows:
θ＝β-α±(2kπ)；
wherein θ represents the standard action angle, β represents the second angle, α represents the first angle, k is the preset integer value, and θ >0.
The accuracy of the standard action angle can be improved by adjusting the standard action angle through the preset integral value.
In at least one embodiment of the present invention, the recognition manner of the follow-up exercise angle is similar to the recognition manner of the standard exercise angle, and details thereof are not repeated herein.
The generation unit 115 generates a follow-up rating of the follow-up video from the standard motion angle and the follow-up motion angle.
It is emphasized that the follow-up rating may also be stored in a node of a block chain in order to further ensure privacy and security of the follow-up rating.
In at least one embodiment of the present invention, the follow-up rating is used to quantify the athletic training profile of the training user. The drill rating includes, but is not limited to: fail rating, pass rating, great rating, perfect rating, and the like.
In at least one embodiment of the present invention, the generating unit 115 generates the follow-up rating of the follow-up video according to the standard action angle and the follow-up action angle includes:
acquiring a key point weight threshold corresponding to any standard key point;
generating the contact ratio of each follow-up frame according to the standard action angle, the follow-up action angle and the corresponding key point weight threshold, wherein the calculation formula of the contact ratio is as follows:
Figure BDA0003837788210000131
wherein S represents the degree of coincidence, W = [ W = 1 ,w 2 ,…,w n ]Representing keypoint weight thresholds corresponding to a plurality of said canonical keypoint, M = [ M = 1 ,m 2 ,…,m n ]Represents a plurality of said standard motion angles, T = [ T ] 1 ,t 2 ,…,t n ]Representing a plurality of the follow-up action angles;
and generating the follow-up rating according to the contact ratio and a preset mapping relation.
The key point weight threshold and the preset mapping relation can be set according to actual requirements. For example, a coincidence of 0% -40% is a Fail rating; the contact ratio is 41% -60%, and the Pass rating is obtained; the contact ratio is 61% -80%, and then the Great rating is obtained; the contact ratio is 81% -100%, and the Perfect rating is obtained.
By setting a key point weight threshold value for each standard key point, the quantification accuracy of the contact ratio can be improved, and thus the accuracy of the follow-up rating is improved.
Specifically, the generating unit 115 generates the follow-up rating according to the contact ratio and a preset mapping relationship, including:
calculating an average value of a plurality of the contact ratios;
and identifying the follow-up rating based on the average value and the preset mapping relation.
According to the technical scheme, the time of the first video reaches the time of the second video can be accurate and quantified, the playing frequency is further based on the playing frequency can be ensured, the standard video sequence can be matched with the exercise action sequence, so that the matching problem caused when the exercise action of a user cannot follow the exercise action in the live sports video can be avoided, and the accuracy of the exercise rating is improved.
Fig. 4 is a schematic structural diagram of an electronic device according to a preferred embodiment of the present invention for implementing a live follow-through video detection method.
In one embodiment of the present invention, the electronic device 1 includes, but is not limited to, a memory 12, a processor 13, and computer readable instructions, such as a live drill video detection program, stored in the memory 12 and executable on the processor 13.
It will be appreciated by a person skilled in the art that the schematic diagram is only an example of the electronic device 1 and does not constitute a limitation of the electronic device 1, and that it may comprise more or less components than shown, or some components may be combined, or different components, e.g. the electronic device 1 may further comprise an input output device, a network access device, a bus, etc.
The Processor 13 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. The processor 13 is an operation core and a control center of the electronic device 1, and is connected to each part of the whole electronic device 1 by various interfaces and lines, and executes an operating system of the electronic device 1 and various installed application programs, program codes, and the like.
Illustratively, the computer readable instructions may be partitioned into one or more modules/units that are stored in the memory 12 and executed by the processor 13 to implement the present invention. The one or more modules/units may be a series of computer readable instruction segments capable of performing specific functions, which are used for describing the execution process of the computer readable instructions in the electronic device 1. For example, the computer-readable instructions may be divided into an acquisition unit 110, a calculation unit 111, an extraction unit 112, a detection unit 113, a recognition unit 114, a generation unit 115, a conversion unit 116, and a playback unit 117.
The memory 12 may be used for storing the computer readable instructions and/or modules, and the processor 13 implements various functions of the electronic device 1 by executing or executing the computer readable instructions and/or modules stored in the memory 12 and invoking data stored in the memory 12. The memory 12 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data created according to use of the electronic device, and the like. The memory 12 may include non-volatile and volatile memories, such as: a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other storage device.
The memory 12 may be an external memory and/or an internal memory of the electronic device 1. Further, the memory 12 may be a memory having a physical form, such as a memory stick, a TF Card (Trans-flash Card), or the like.
The integrated modules/units of the electronic device 1 may be stored in a computer-readable storage medium if they are implemented in the form of software functional units and sold or used as separate products. Based on such understanding, all or part of the flow of the method according to the above embodiments may be implemented by hardware that is configured to be instructed by computer readable instructions, which may be stored in a computer readable storage medium, and when the computer readable instructions are executed by a processor, the steps of the method embodiments may be implemented.
Wherein the computer readable instructions comprise computer readable instruction code which may be in source code form, object code form, an executable file or some intermediate form, and the like. The computer-readable medium may include: any entity or device capable of carrying said computer readable instruction code, a recording medium, a U disk, a removable hard disk, a magnetic diskette, an optical disk, a computer Memory, a Read-Only Memory (ROM), a Random Access Memory (RAM).
The block chain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a string of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, which is used for verifying the validity (anti-counterfeiting) of the information and generating a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.
With reference to fig. 1, the memory 12 in the electronic device 1 stores computer-readable instructions to implement a live-follow video detection method, and the processor 13 can execute the computer-readable instructions to implement:
acquiring a live motion video, and acquiring a follow-up video based on the live motion video;
calculating the playing frequency of the live motion video according to the first video time length of the live motion video and the second video time length of the follow-up video;
extracting a standard video sequence from the live motion video based on the playing frequency, and extracting a follow-up action sequence from the follow-up video;
detecting a standard key point of each standard video frame in the standard video sequence, and detecting a user key point of each follow-up frame in the follow-up action sequence;
identifying a standard action angle of each standard video frame based on the standard key points, and identifying a follow-up action angle of each follow-up frame based on the user key points;
and generating a follow-up training rating of the follow-up training video according to the standard action angle and the follow-up training action angle.
Specifically, the processor 13 may refer to the description of the relevant steps in the embodiment corresponding to fig. 1 for a specific implementation method of the computer readable instructions, which is not described herein again.
In the embodiments provided in the present invention, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is only one logical functional division, and other divisions may be realized in practice.
The computer readable storage medium has computer readable instructions stored thereon, wherein the computer readable instructions when executed by the processor 13 are configured to implement the steps of:
acquiring a live motion video, and acquiring a follow-up video based on the live motion video;
calculating the playing frequency of the live motion video according to the first video time length of the live motion video and the second video time length of the follow-up video;
extracting a standard video sequence from the live motion video based on the playing frequency, and extracting a follow-up action sequence from the follow-up video;
detecting a standard key point of each standard video frame in the standard video sequence, and detecting a user key point of each follow-up frame in the follow-up action sequence;
identifying a standard action angle of each standard video frame based on the standard key points, and identifying a follow-up action angle of each follow-up frame based on the user key points;
and generating a follow-up training rating of the follow-up training video according to the standard action angle and the follow-up training action angle.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one position, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
In addition, functional modules in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional module.
The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims shall not be construed as limiting the claim concerned.
Furthermore, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. The plurality of units or devices may also be implemented by one unit or device through software or hardware. The terms first, second, etc. are used to denote names, but not any particular order.
Finally, it should be noted that the above embodiments are only intended to illustrate the technical solutions of the present invention and not to limit the same, and although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims (10)

1. A live broadcast follow-through video detection method is characterized by comprising the following steps:
acquiring a live motion video, and acquiring a follow-up video based on the live motion video;
calculating the playing frequency of the live motion video according to the first video time length of the live motion video and the second video time length of the follow-up video;
extracting a standard video sequence from the live motion video based on the playing frequency, and extracting a follow-up action sequence from the follow-up video;
detecting a standard key point of each standard video frame in the standard video sequence, and detecting a user key point of each follow-up frame in the follow-up action sequence;
identifying a standard action angle of each standard video frame based on the standard key points, and identifying a follow-up action angle of each follow-up frame based on the user key points;
and generating a follow-up training rating of the follow-up training video according to the standard action angle and the follow-up training action angle.
2. The method for detecting the live follow-through video as claimed in claim 1, wherein the calculating the playing frequency of the live motion video according to the first video duration of the live motion video and the second video duration of the follow-through video comprises:
identifying a starting generation time point of a starting video frame in the live broadcast motion video, and identifying an ending generation time point of an ending video frame in the live broadcast motion video;
calculating a difference value between the generation ending time point and the generation starting time point to obtain the first video time length;
and calculating the ratio of the second video time length to the first video time length to obtain the playing frequency.
3. The live drill-following video detection method of claim 1, wherein the extracting a standard video sequence from the live motion video based on the play frequency comprises:
comparing the playing frequency with a preset frequency;
if the playing frequency is greater than the preset frequency, converting the live broadcast motion video based on the playing frequency to obtain a standard motion video, and generating the standard video sequence according to a standard video frame in the standard motion video and a frame position of the standard video frame in the standard motion video; or
And if the playing frequency is less than or equal to the preset frequency, extracting the standard video frame from the live motion video based on the playing frequency to serve as the standard video sequence.
4. The method of claim 1, wherein the detecting the standard keypoints for each standard video frame in the standard video sequence comprises:
for each standard video frame, acquiring the pixel value and the pixel position of each pixel point in the standard video frame;
detecting the pixel values and the pixel positions based on a pre-trained human body detection model to obtain a detection category corresponding to each pixel point and a detection probability of the pixel point belonging to the detection category;
screening target pixel points from the plurality of pixel points based on the detection probability;
and determining an area formed by adjacent pixel points corresponding to the same detection category in the target pixel points as the standard key point based on the pixel position and the detection category.
5. The live drill-following video detection method of claim 4, wherein the identifying of the standard action angle for each standard video frame based on the standard keypoints comprises:
constructing a plane rectangular coordinate system by taking any standard key point as an origin and taking an image edge parallel to any standard video frame as a coordinate axis;
identifying a key point coordinate value corresponding to the standard key point according to the pixel position corresponding to the standard key point and the plane rectangular coordinate system;
acquiring a connection key point pair of any standard key point from a plurality of standard key points, wherein the connection key point pair comprises a first connection point and a second connection point;
determining a connecting edge formed by any standard key point and the first connecting point as a first connecting edge, and calculating an included angle between the first connecting edge and the coordinate axis based on the key point coordinate value of the first connecting point to obtain a first angle;
determining a connecting edge formed by any standard key point and the second connecting point as a second connecting edge, and calculating an included angle between the second connecting edge and the coordinate axis based on the key point coordinate value of the second connecting point to obtain a second angle;
and generating the standard action angle based on the first angle, the second angle and a preset integer value.
6. The live follow-through video detection method of claim 5, wherein the standard action angle is calculated by the formula:
θ＝β-α±(2kπ)；
wherein θ represents the standard action angle, β represents the second angle, α represents the first angle, k is the preset integer value, and θ >0.
7. The live follow-up video detection method of claim 5, wherein the generating of the follow-up rating of the follow-up video according to the standard action angle and the follow-up action angle comprises:
acquiring a key point weight threshold corresponding to any standard key point;
generating the contact ratio of each follow-up frame according to the standard action angle, the follow-up action angle and the corresponding key point weight threshold, wherein the calculation formula of the contact ratio is as follows:
Figure FDA0003837788200000021
wherein S represents the degree of coincidence, W = [ W = 1 ,w 2 ,…,w n ]Representing keypoint weight thresholds corresponding to a plurality of said canonical keypoint, M = [ M = 1 ,m 2 ,…,m n ]Represents a plurality of said standard motion angles, T = [ T ] 1 ,t 2 ,…,t n ]Representing a plurality of the follow-up exercise angles;
and generating the follow-up rating according to the contact ratio and a preset mapping relation.
8. The utility model provides a live with practise video detection device which characterized in that, live with practise video detection device includes:
the device comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring a live motion video and acquiring a follow-up video based on the live motion video;
the calculation unit is used for calculating the playing frequency of the live motion video according to the first video time length of the live motion video and the second video time length of the follow-up video;
the extraction unit is used for extracting a standard video sequence from the live motion video based on the playing frequency and extracting a follow-up action sequence from the follow-up video;
the detection unit is used for detecting a standard key point of each standard video frame in the standard video sequence and detecting a user key point of each follow-up frame in the follow-up action sequence;
the identification unit is used for identifying the standard action angle of each standard video frame based on the standard key points and identifying the follow-up action angle of each follow-up frame based on the user key points;
and the generating unit is used for generating a follow-up training rating of the follow-up training video according to the standard action angle and the follow-up training action angle.
9. An electronic device, characterized in that the electronic device comprises:
a memory storing computer readable instructions; and
a processor executing computer readable instructions stored in the memory to implement the live follow-through video detection method of any of claims 1-7.
10. A computer-readable storage medium characterized by: the computer-readable storage medium has stored therein computer-readable instructions that are executed by a processor in an electronic device to implement the live follow-through video detection method as recited in any one of claims 1 to 7.
CN202211093077.9A 2022-09-08 2022-09-08 Live broadcast follow-up video detection method, device, equipment and storage medium Pending CN115588231A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211093077.9A CN115588231A (en) 2022-09-08 2022-09-08 Live broadcast follow-up video detection method, device, equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211093077.9A CN115588231A (en) 2022-09-08 2022-09-08 Live broadcast follow-up video detection method, device, equipment and storage medium

Publications (1)

Publication Number Publication Date
CN115588231A true CN115588231A (en) 2023-01-10

Family

ID=84771123

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211093077.9A Pending CN115588231A (en) 2022-09-08 2022-09-08 Live broadcast follow-up video detection method, device, equipment and storage medium

Country Status (1)

Country Link
CN (1) CN115588231A (en)

Similar Documents

Publication Publication Date Title
Patrona et al. Motion analysis: Action detection, recognition and evaluation based on motion capture data
WO2021129064A9 (en) Posture acquisition method and device, and key point coordinate positioning model training method and device
Chang et al. Feature interaction augmented sparse learning for fast kinect motion detection
CN112184705B (en) Human body acupuncture point identification, positioning and application system based on computer vision technology
WO2021217927A1 (en) Video-based exercise evaluation method and apparatus, and computer device and storage medium
CN110705390A (en) Body posture recognition method and device based on LSTM and storage medium
US11837028B2 (en) Dance segment recognition method, dance segment recognition apparatus, and storage medium
CN111222486B (en) Training method, device and equipment for hand gesture recognition model and storage medium
WO2023050650A1 (en) Animation video generation method and apparatus, and device and storage medium
CN110782482A (en) Motion evaluation method and device, computer equipment and storage medium
Muneesawang et al. A machine intelligence approach to virtual ballet training
Zou et al. Intelligent fitness trainer system based on human pose estimation
CN111783593A (en) Human face recognition method and device based on artificial intelligence, electronic equipment and medium
CN115131879A (en) Action evaluation method and device
WO2023108842A1 (en) Motion evaluation method and system based on fitness teaching training
Morel et al. Automatic evaluation of sports motion: A generic computation of spatial and temporal errors
CN116205723A (en) Artificial intelligence-based face tag risk detection method and related equipment
CN116844084A (en) Sports motion analysis and correction method and system integrating blockchain
CN115588231A (en) Live broadcast follow-up video detection method, device, equipment and storage medium
Li et al. Intelligent correction method of shooting action based on computer vision
CN116468043A (en) Nested entity identification method, device, equipment and storage medium
CN115530814A (en) Child motion rehabilitation training method based on visual posture detection and computer deep learning
CN115188074A (en) Interactive physical training evaluation method, device and system and computer equipment
CN114495169A (en) Training data processing method, device and equipment for human body posture recognition
Cao How Technology is Revolutionizing Sports Training and healthcare teaching.

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