CN113627396A - Health monitoring-based skipping rope counting method - Google Patents

Health monitoring-based skipping rope counting method Download PDF

Info

Publication number
CN113627396A
CN113627396A CN202111107598.0A CN202111107598A CN113627396A CN 113627396 A CN113627396 A CN 113627396A CN 202111107598 A CN202111107598 A CN 202111107598A CN 113627396 A CN113627396 A CN 113627396A
Authority
CN
China
Prior art keywords
rope skipping
coordinates
round
coordinate
heart rate
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
Application number
CN202111107598.0A
Other languages
Chinese (zh)
Other versions
CN113627396B (en
Inventor
林峰
鲁昱舟
韩涛
厉位阳
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.)
Zhejiang University ZJU
Original Assignee
Zhejiang University ZJU
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 Zhejiang University ZJU filed Critical Zhejiang University ZJU
Priority to CN202111107598.0A priority Critical patent/CN113627396B/en
Publication of CN113627396A publication Critical patent/CN113627396A/en
Application granted granted Critical
Publication of CN113627396B publication Critical patent/CN113627396B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
    • A61B5/024—Detecting, measuring or recording pulse rate or heart rate
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Medical Informatics (AREA)
  • Surgery (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Physiology (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Cardiology (AREA)
  • Molecular Biology (AREA)
  • Physics & Mathematics (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Psychiatry (AREA)
  • Signal Processing (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The invention discloses a rope skipping counting method based on health monitoring, which adopts a non-contact video image method to count rope skipping of rope skipping personnel by identifying rope skipping people and rope skipping states in a video; and meanwhile, the condition that the heart rate changes according to the rope skipping times when the human body jumps by the rope is monitored according to the video, so that the motion quantity is controlled within an expected range, the sudden cardiac death risk is warned, and the like. The method adopts a non-contact technology, has high counting accuracy and high counting speed, can simultaneously monitor the conditions of heart rate change and the like during exercise, and has high application value.

Description

Health monitoring-based skipping rope counting method
Technical Field
The invention relates to the field of exercise monitoring methods, in particular to a rope skipping counting method based on health monitoring.
Background
With the improvement of living standard and the change of health concept, body-building exercises such as rope skipping, ball patting, shuttlecock kicking and the like become a part of daily life of people, and the exercises are greatly helpful for various visceral organs, coordination, posture, weight reduction and the like of heart-lung systems and the like. However, in the process of competition or exercise, the number of times of exercises needs to be accurately and effectively judged, and a plurality of rope skipping counting methods based on artificial intelligence technology exist at present.
As disclosed in chinese patent documents publication nos. CN109876416A and CN110210360A, a rope skipping counting method based on image information and a rope skipping counting method based on video image object recognition perform counting by methods of recognizing positions of ropes and faces. Chinese patent publication No. CN110102040A discloses an audio rope skipping counting method based on a cross-correlation coefficient method, which achieves the purpose of determining the number of rope skipping times by recognizing the sound of a rope contacting the ground. The method for counting skipping ropes based on deep learning disclosed in Chinese patent publication No. CN112044046A comprises preprocessing obtained image data, classifying by using trained models, judging current motion state according to classification results, and counting the number of skipping rope state changes.
These methods are only used for rope skipping counting, but do not monitor the health condition of the human body during rope skipping, such as the physiological indicators of heart rate and the like. In order to control the exercise amount within a reasonable range, warn sudden cardiac death risk, etc., it is also necessary to monitor the relationship between the exercise and the change of physiological information (such as heart rate, etc.), and in this respect, there are some methods for monitoring the heart rate by adopting a non-contact technology.
For example, the invention combines the CNN feature extraction and LSTM long-and-short-term memory neural network, and embeds the channel attention network, so as to realize the heart rate non-contact measurement with low error rate and high efficiency. Chinese patent publication No. CN 112381011 a discloses a heart rate non-contact measurement method based on a face video sequence, which detects and tracks a face region of a human body in real time to realize non-contact measurement of heart rate by acquiring a video sequence containing face information of the human body and combining local texture features and a skin color model of an image.
The method mainly aims at non-contact heart rate monitoring of a static body, and cannot effectively monitor the heart rate of a moving human body.
Disclosure of Invention
The invention aims to provide a rope skipping counting method based on health monitoring, and the method is used for solving the problem that the prior art cannot effectively monitor the heart rate of a human body in rope skipping sports.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
step 1, obtaining original video data of rope skipping movement, and extracting image data from the original video data;
step 2, counting the rope skipping movement;
step 3, monitoring the heart rate during rope skipping;
and 4, outputting and displaying the result.
Among the above-mentioned technical scheme, utilize the camera to shoot rope skipping sportsman's whole body, and the camera keeps static, under the stable condition, handles video information, then statistics rope skipping state change number of times counts, monitors rope skipping in-process sportsman's rhythm of the heart again according to this video. The method has the advantages of high accuracy, high counting speed, heart rate monitoring speed meeting the actual requirement and high application value.
The rope skipping movement counting method in the step 2 adopts a rope skipping counting scheme based on PoseNet. Preferably, a MobileNet model is adopted, and the method specifically comprises the following steps:
step 2-1, inputting a PoseNet attitude estimation model; the method specifically comprises the following steps: an image; an image scale factor; horizontally turning over variables; and outputting the step.
Wherein, the image scale factor is the scale of zooming the image before inputting the image to the model; the output step is the ratio of the input image to the output image, the larger the output step is, the smaller the sizes of the characteristic diagram and the output image in the network are, the detection accuracy is reduced, the speed is the fastest, and the optional value of the parameter is 8, 16 or 32.
Step 2-2, PoseNet processing the image; when Posenet processes images, the Heatmap (Heatmap) and Offset Vectors (Offset Vectors) for the 17 body pose keypoints are actually output. The 17 key point parts detected by PoseNet are respectively as follows according to the index sequence: nose, left eye, right eye, left ear, right ear, left shoulder, right shoulder, left elbow, right elbow, left wrist, right wrist, left hip, right hip, left knee, right knee, left ankle, right ankle. An example of detecting the key points of the PoseNet human body part is shown in FIG. 2. The method specifically comprises the following steps:
step 2-2-1, heat map coding;
the heat map is a 3D tensor with the number of channels being 17, and each channel corresponds to the information of one key point; the offset vector is a 3D tensor with 34 channels, the first 17 channels represent x coordinate information, and the last 17 channels represent y coordinate information, which can be understood as an encoding process.
Step 2-2-2, decoding; decoding the heatmap and offset vector can find the region coordinates of the corresponding key points with high confidence. The decoding scheme is that the heat map is applied first
Figure BDA0003273086600000031
Activating a function to obtain a keypoint confidence; for these confidence scores, arg max is taken, and the p-index and q-index with the highest score for each body part in the heat map are retrieved, where arg max is defined as follows:
Figure BDA0003273086600000032
wherein f (p, q) represents the confidence value of the heatmap calculated by sigmoid.
Obtaining an offset vector of each part according to the offset of the x index and the y index in the heat map; and restoring the heat map to the size of the original image according to the output stride and the corresponding offset vector to obtain the x coordinate, the y coordinate and the confidence coefficient of each key point on the heat map.
Step 2-3, rope skipping counting:
carrying out rope skipping counting, wherein 17 key point coordinates of the body comprise x and y coordinates, transverse offset is not concerned in rope skipping counting, error is increased by incorporating the x coordinate, and therefore the algorithm only relates to the y coordinate; formula (II)
Figure BDA0003273086600000033
Remove from0-4 key point coordinates, namely the coordinates of the nose, the left eye, the right eye, the left ear and the right ear, so as to reduce unnecessary errors caused by head shaking, wherein ypThe y coordinates of 5 th to 17 th key points in the original image without the x coordinates are represented, and in addition, the y coordinates of the left wrist and the right wrist are excluded in actual operation, because the movement trend of the left wrist and the right wrist is not consistent with that of the whole trunk in the rope skipping process; dyprevious_roundDy obtained for the previous round of calculationcurrent_roundThe parameter runs through the whole counting process, so that the fluctuation of the calculation coordinate between two frames is stable, and the stability of the algorithm is ensured.
Step 2-4, jitter suppression:
in order to reduce the error influence caused by shaking, such as unexpected shaking of a camera, the coordinate change of key points between certain frames is too violent or slight, and the coordinate change is dyWhen calculating, dyprevious_roundAs a jitter suppression parameter, the characteristics of square root amplification value and reduction value are used for modification as follows:
Figure BDA0003273086600000034
Figure BDA0003273086600000041
wherein sgn (x) is a function for judging positive and negative, ymean_i，ymean_i+1，ymean_i+2Respectively representing the average value of y coordinates, dy, of 5-17 key points of the ith, i +1 and i +2 framesprevious_roundDy obtained for the previous round of calculationcurrent_roundAnd dycurrent_roundDefined by the following equation:
dycurrent_round＝(dyprevious_round+|ymean_i+1-ymean_i|+|ymean_i+2-ymean_i+1|)/3
the parameter dycurrent_roundThrough the whole counting process, the fluctuation of the calculation coordinates between two frames is stable, and the stability of the algorithm is ensured.
The heart rate monitoring during rope skipping in step 3 adopts a self-adaptive face detection and extracted rPPG heart rate detection method, and the specific steps are as follows:
step 3-1, self-adaptive face detection and extraction, wherein a self-adaptive face detection and extraction scheme based on YOLOv5 is adopted, and the method specifically comprises the following steps:
step 3-1-1, taking out 10 frame indexes from the video frame images, and taking out the frame images according to the indexes;
3-1-2, detecting a face area by adopting a YOLOv5s network, and taking the minimum value of the upper left coordinate and the maximum value of the lower right coordinate from 10 coordinates;
3-1-3, expanding the coordinates, and cutting according to the expanded coordinates;
and 3-1-4, inputting the data to a heart rate detection network.
The self-adaptive face detection and extraction module greatly increases the recognition rate of the face, so that the whole heart rate detection system is more stable and reliable, and the robustness is enhanced.
Step 3-2, ROI extraction; and positioning and extracting a region of interest ROI required by heart rate detection. The temperature of the forehead and cheek areas is relatively constant, so that the cheek areas with more exposed parts and stronger stability are selected to extract the pulse wave signals.
Step 3-3, extracting a blood volume wave signal (BVP) and extracting a heart rate; after the face ROI extraction is completed, signal extraction needs to be carried out on the area, a green channel (G channel) signal in an RBG channel of the face ROI area is adopted for digital signal filtering to obtain a BVP signal, and the cheek is divided into two parts, so that the average value of the G channel signals of the two ROIs is used as an input pulse wave signal. The processing steps are as follows:
3-3-1, filtering by using a Butterworth band-pass Filter (Butterworth Filter); to reduce the effects of environmental changes and motion-induced noise.
Step 3-3-2, trend removing (Detrend) processing; the method is used for reducing errors caused by illumination change, and the influence of the offset of the sensor can be effectively reduced through trend removing processing.
3-3-3, collecting signals by a Hamming (Hamming) window; the frequency leakage problem can be improved relative to directly truncating the signal (e.g., rectangular window truncation).
3-3-4, performing discrete Fourier transform; and (4) converting the signal into a frequency domain to extract the periodic characteristic, wherein the frequency corresponding to the energy spectrum peak in the frequency domain can be regarded as the heart rate.
Compared with the prior art, the invention has the advantages that:
the invention provides a rope skipping counting method based on health monitoring, which adopts a non-contact video image method to count rope skipping of rope skipping personnel by identifying rope skipping people and rope skipping states in a video; and meanwhile, the condition that the heart rate changes according to the rope skipping times when the human body jumps by the rope is monitored according to the video, so that the motion quantity is controlled within an expected range, the sudden cardiac death risk is warned, and the like. The method adopts a non-contact technology, has high counting accuracy and high counting speed, can simultaneously monitor the conditions of heart rate change and the like during exercise, and has high application value.
Drawings
Fig. 1 is a flowchart of a rope skipping counting method based on deep learning in an embodiment of the present invention.
Fig. 2 is an example of detecting key points of a PoseNet human body part in the embodiment of the present invention.
Fig. 3 is a skipping rope counting flow chart in the embodiment of the invention.
Fig. 4 is a schematic diagram of an adaptive face detection and extraction algorithm in the embodiment of the present invention.
FIG. 5 is a diagram illustrating an example of ROI extraction positioning in an embodiment of the present invention.
FIG. 6 is a waveform of a processed blood volume wave signal (BVP) and a heart rate signal according to an 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 further described with reference to the following embodiments and accompanying drawings. It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the described embodiments without any inventive step, are within the scope of protection of the invention.
Unless defined otherwise, technical or scientific terms used herein shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention belongs. The use of the word "comprise" or "comprises", and the like, in the context of this application, is intended to mean that the elements or items listed before that word, in addition to those listed after that word, do not exclude other elements or items. The terms "connected" or "coupled" and the like are not restricted to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "upper", "lower", "left", "right", and the like are used merely to indicate relative positional relationships, and when the absolute position of the object being described is changed, the relative positional relationships may also be changed accordingly.
Examples
Referring to fig. 1 to 6, a rope skipping counting method based on health monitoring of the present embodiment includes the following steps:
and S100, acquiring original video data of rope skipping movement, and extracting image data from the original video data.
And step S200, counting the rope skipping movement. This example employs a PoseNet-based rope skipping counting scheme. Preferably, the method adopts a MobileNet model, and comprises the following specific steps:
step S210, inputting a PoseNet attitude estimation model; the method specifically comprises the following steps: an image; an image scale factor; horizontally turning over variables; and outputting the step.
Step S220, PoseNet processes the image; when Posenet processes images, the Heatmap (Heatmap) and Offset Vectors (Offset Vectors) for the 17 body pose keypoints are actually output. The 17 key point parts detected by PoseNet are respectively as follows according to the index sequence: nose, left eye, right eye, left ear, right ear, left shoulder, right shoulder, left elbow, right elbow, left wrist, right wrist, left hip, right hip, left knee, right knee, left ankle, right ankle. An example of detecting the key points of the PoseNet human body part is shown in FIG. 2. The method specifically comprises the following steps:
step S221, heat map coding; the heat map is a 3D tensor with the number of channels being 17, and each channel corresponds to the information of one key point; the offset vector is a 3D tensor with 34 channels, the first 17 channels represent x coordinate information, and the last 17 channels represent y coordinate information, which can be understood as an encoding process.
Step S222, decoding; decoding the heatmap and offset vector can find the region coordinates of the corresponding key points with high confidence. The decoding scheme is that the heat map is applied first
Figure BDA0003273086600000061
Activating a function to obtain a keypoint confidence; for these confidence scores, arg max is taken, and the p-index and q-index with the highest score for each body part in the heat map are retrieved, where arg max is defined as follows:
Figure BDA0003273086600000062
wherein f (p, q) represents the confidence value of the heatmap calculated by sigmoid.
Obtaining an offset vector of each part according to the offset of the p index and the q index in the heat map; and restoring the heat map to the size of the original image according to the output stride and the corresponding offset vector to obtain the x coordinate, the y coordinate and the confidence coefficient of each key point on the original image.
Step S230, counting skipping ropes; the flow is shown in fig. 3.
The 17 key point coordinates of the body comprise x and y coordinates, and transverse offset is not concerned in rope skipping counting, and the error is increased by the inclusion of the x coordinate, so that the algorithm only relates to the y coordinate. Formula (II)
Figure BDA0003273086600000071
Removing 0-4 key point coordinates, namely the coordinates of the nose, the left eye, the right eye, the left ear and the right ear to reduce unnecessary errors caused by head shaking, wherein ypAnd the y coordinates of 5 th to 17 th key points in the original image without the x coordinates are represented, and in addition, the y coordinates of the left wrist and the right wrist are excluded in actual operation, because the movement trend of the left wrist and the right wrist is not consistent with the whole trunk in the rope skipping process. dyprevious_roundFor the previous round of calculationObtained dycurrent_roundThe parameter runs through the whole counting process, so that the fluctuation of the calculation coordinate between two frames is stable, and the stability of the algorithm is ensured.
Step S240, jitter suppression; the description is as follows:
at dyWhen calculating, dyprevious_roundAs a jitter suppression parameter, the characteristics of square root amplification value and reduction value are used for modification as follows:
Figure BDA0003273086600000072
Figure BDA0003273086600000073
wherein sgn (x) is a function for judging the negativity and the negativity.
And step S300, monitoring the heart rate during rope skipping. The example adopts an adaptive face detection and extracted rPPG heart rate detection method, and the specific steps are as follows:
step S310, self-adaptive face detection and extraction; the embodiment adopts a scheme of adaptive face detection and extraction based on YOLOv5, and the flow is shown in fig. 4, and specifically includes:
in step S311, a 10-frame index is extracted from the video frame image, and the frame image is extracted according to the index.
In step S312, a face area is detected by using the YOLOv5S network, and the minimum value of the upper left coordinate and the maximum value of the lower right coordinate are taken from the 10 coordinates.
And step 313, expanding the coordinates and cutting according to the expanded coordinates.
And step S314, inputting the data to a heart rate detection network.
Step S320, ROI extraction; and positioning and extracting a region of interest ROI required by heart rate detection. The temperature of the forehead and cheek areas is relatively constant, and the cheek areas with more exposed parts and stronger stability are selected in the example to extract the pulse wave signals, as shown in fig. 5.
Step S330, extracting a blood volume wave signal (BVP) and a heart rate; after the face ROI extraction is completed, signal extraction needs to be performed on the region, in this example, a green channel (G channel) signal in an RBG channel of the face ROI region is used for digital signal filtering to obtain a BVP signal, and since the cheek is divided into two parts, the average value of the G channel signals of the two ROIs is used as an input pulse wave signal. The processing steps are as follows:
in step S331, a Butterworth band pass Filter (Butterworth Filter) performs filtering processing.
In step S332, detrending (Detrend) processing is performed.
In step S333, Hamming (Hamming) window collects signals.
Step S334, discrete fourier transform.
And step S400, outputting and displaying the result.
The embodiments of the present invention are described only for the preferred embodiments of the present invention, and not for the limitation of the concept and scope of the present invention, and various modifications and improvements made to the technical solution of the present invention by those skilled in the art without departing from the design concept of the present invention shall fall into the protection scope of the present invention, and the technical content of the present invention which is claimed is fully set forth in the claims.

Claims (3)

1. A rope skipping counting method based on health monitoring is characterized by comprising the following steps:
step 1, obtaining original video data of rope skipping movement, and extracting image data from the original video data;
step 2, obtaining rope skipping movement count from the image data extracted in the step 1 by adopting a rope skipping counting method based on PoseNet;
step 3, obtaining heart rate monitoring data during rope skipping movement from the image data extracted in the step 1 by adopting a self-adaptive face detection and extracted rPPG heart rate detection method;
and 4, outputting results obtained in the steps 2 and 3.
2. The rope skipping counting method based on health monitoring as claimed in claim 1, wherein the step 2 process is as follows:
(2.1) constructing a Posenet attitude estimation model based on a MobileNet architecture, wherein the input of the Posenet attitude estimation model comprises the following steps: images, image scale factors, horizontal turnover variables and output steps;
(2.2) processing the image data extracted in the step 1 by adopting a PoseNet attitude estimation model; after the PoseNet attitude estimation model processes image data, outputting heat maps and offset vectors of 17 body attitude key points; the 17 key point parts detected by the PoseNet attitude estimation model are respectively as follows according to the index sequence: nose, left eye, right eye, left ear, right ear, left shoulder, right shoulder, left elbow, right elbow, left wrist, right wrist, left hip, right hip, left knee, right knee, left ankle, right ankle, specific treatment process as follows:
(2.2.1) firstly, carrying out heat map coding, wherein the heat map is a 3D tensor with 17 channels, and each channel corresponds to the information of one key point; the offset vector is a 3D tensor with 34 channels, the first 17 channels represent x coordinate information, and the last 17 channels represent y coordinate information, which can be understood as an encoding process;
(2.2.2) then decoding, and decoding the heatmap and the offset vector to find out the corresponding region coordinates with high confidence coefficient of the key point; the decoding scheme is that the heat map is applied first
Figure FDA0003273086590000011
Activating a function to obtain a keypoint confidence; for these confidence scores, arg max is taken, and the p-index and q-index with the highest score for each body part in the heat map are retrieved, where arg max is defined as follows:
Figure FDA0003273086590000012
wherein f (p, q) represents the confidence value of the heat map calculated by sigmoid;
obtaining an offset vector of each part according to the offset of the x index and the y index in the heat map; restoring the heat map to the size of the original image according to the output stride and the corresponding offset vector to obtain the x coordinate, the y coordinate and the confidence coefficient of each key point on the heat map;
(2.3) carrying out rope skipping counting, wherein the 17 key point coordinates of the body comprise x and y coordinates, lateral deviation is not concerned in the rope skipping counting, and the error is increased by the inclusion of the x coordinate, so that the algorithm only relates to the y coordinate; formula (II)
Figure FDA0003273086590000021
Removing 0-4 key point coordinates, namely the coordinates of the nose, the left eye, the right eye, the left ear and the right ear to reduce unnecessary errors caused by head shaking, wherein ypThe y coordinates of 5 th to 17 th key points in the original image without the x coordinates are represented, and in addition, the y coordinates of the left wrist and the right wrist are excluded in actual operation, because the movement trend of the left wrist and the right wrist is not consistent with that of the whole trunk in the rope skipping process; dyprevious_roundDy obtained for the previous round of calculationcurrent_roundThe parameter runs through the whole counting process, so that the fluctuation of the calculation coordinates between two frames is stable, and the stability of the algorithm is ensured;
(2.4), jitter suppression:
to reduce the effect of errors due to jitter, at dyWhen calculating, dyprevious_roundAs a jitter suppression parameter, the characteristics of square root amplification value and reduction value are used for modification as follows:
Figure FDA0003273086590000022
Figure FDA0003273086590000023
wherein sgn (x) is a function for judging positive and negative, ymean_i，ymean_i+1，ymean_i+2Respectively representing the average value of y coordinates, dy, of 5-17 key points of the ith, i +1 and i +2 framesprevious_roundDy obtained for the previous round of calculationcurrent_roundAnd dycurrent_roundIs composed ofThe formula defines:
dycurrent_round＝(dyprevious_round+|ymean_i+1-ymean_i|+|ymean_i+2-ymean_i+1|)/3，
parameter dycurrent_roundThrough the whole counting process, the fluctuation of the calculation coordinates between two frames is stable, and the stability of the algorithm is ensured.
3. The rope skipping counting method based on health monitoring as claimed in claim 1, wherein the procedure of step 3 is as follows:
(3.1) self-adaptive face detection and extraction: the adaptive face detection and extraction scheme based on YOLOv5 specifically comprises the following steps:
(3.1.1) taking out 10 frame indexes from the video frame images, and taking out the frame images according to the indexes;
(3.1.2) detecting a face area by adopting a YOLOv5s network, and taking the minimum value of the upper left coordinate and the maximum value of the lower right coordinate from 10 coordinates;
(3.1.3) expanding the coordinates, and cutting according to the expanded coordinates;
and (3.1.4) inputting the data to a heart rate detection network.
(3.2) extracting human face ROI: the ROI required by heart rate detection is positioned and extracted, and because the temperatures of the forehead area and the cheek area are relatively constant, the cheek area with more exposed parts and stronger stability is selected to extract pulse wave signals;
(3.3) extracting the blood volume wave signal BVP and extracting the heart rate: after the face ROI extraction is completed, signal extraction needs to be carried out on the area, a green channel, namely a G channel signal in an RBG channel of the face ROI area is adopted for digital signal filtering to obtain a BVP signal, the cheek is divided into two parts, so that the average value of the G channel signals of the two ROIs is used as an input pulse wave signal, and the processing steps are as follows:
(3.3.1) filtering by a Butterworth band-pass filter;
(3.3.2) trend removing treatment;
(3.3.3), collecting signals by a Hamming window;
(3.3.4), discrete fourier transform.
CN202111107598.0A 2021-09-22 2021-09-22 Rope skipping counting method based on health monitoring Active CN113627396B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111107598.0A CN113627396B (en) 2021-09-22 2021-09-22 Rope skipping counting method based on health monitoring

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111107598.0A CN113627396B (en) 2021-09-22 2021-09-22 Rope skipping counting method based on health monitoring

Publications (2)

Publication Number Publication Date
CN113627396A true CN113627396A (en) 2021-11-09
CN113627396B CN113627396B (en) 2023-09-05

Family

ID=78390543

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111107598.0A Active CN113627396B (en) 2021-09-22 2021-09-22 Rope skipping counting method based on health monitoring

Country Status (1)

Country Link
CN (1) CN113627396B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114287938A (en) * 2021-12-13 2022-04-08 重庆大学 Method and device for obtaining safety interval of human body parameters in building environment

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106778695A (en) * 2017-01-19 2017-05-31 北京理工大学 A kind of many people's examing heartbeat fastly methods based on video
CN109830277A (en) * 2018-12-12 2019-05-31 平安科技（深圳）有限公司 A kind of rope skipping monitoring method, electronic device and storage medium
WO2020191989A1 (en) * 2019-03-26 2020-10-01 浙江大学 Rope skipping counting method
CN112163516A (en) * 2020-09-27 2021-01-01 深圳市悦动天下科技有限公司 Rope skipping counting method and device and computer storage medium
CN112633220A (en) * 2020-12-30 2021-04-09 浙江工商大学 Human body posture estimation method based on bidirectional serialization modeling
CN113297883A (en) * 2020-07-03 2021-08-24 阿里巴巴集团控股有限公司 Information processing method, analysis model obtaining device and electronic equipment

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106778695A (en) * 2017-01-19 2017-05-31 北京理工大学 A kind of many people's examing heartbeat fastly methods based on video
CN109830277A (en) * 2018-12-12 2019-05-31 平安科技（深圳）有限公司 A kind of rope skipping monitoring method, electronic device and storage medium
WO2020191989A1 (en) * 2019-03-26 2020-10-01 浙江大学 Rope skipping counting method
CN113297883A (en) * 2020-07-03 2021-08-24 阿里巴巴集团控股有限公司 Information processing method, analysis model obtaining device and electronic equipment
CN112163516A (en) * 2020-09-27 2021-01-01 深圳市悦动天下科技有限公司 Rope skipping counting method and device and computer storage medium
CN112633220A (en) * 2020-12-30 2021-04-09 浙江工商大学 Human body posture estimation method based on bidirectional serialization modeling

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
GEORGE PAPANDREOU ET AL.: "Towards Accurate Multi-person Pose Estimation in the Wild", 《ARXIV:1701.01779V2》, pages 1 - 9 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114287938A (en) * 2021-12-13 2022-04-08 重庆大学 Method and device for obtaining safety interval of human body parameters in building environment
CN114287938B (en) * 2021-12-13 2024-02-13 重庆大学 Method and equipment for obtaining safety interval of human body parameters in building environment

Also Published As

Publication number Publication date
CN113627396B (en) 2023-09-05

Similar Documents

Publication Publication Date Title
CN110464357B (en) Rehabilitation process quality monitoring method and system
KR20200068545A (en) System and method for training a convolutional neural network and classifying an action performed by a subject in a video using the trained convolutional neural network
CN110705390A (en) Body posture recognition method and device based on LSTM and storage medium
CN111728602A (en) Non-contact blood pressure measuring device based on PPG
CN111839489B (en) Non-contact physiological and psychological health detection system
Chaudhari et al. Yog-guru: Real-time yoga pose correction system using deep learning methods
CN111753747B (en) Violent motion detection method based on monocular camera and three-dimensional attitude estimation
KR101738278B1 (en) Emotion recognition method based on image
CN110490109B (en) Monocular vision-based online human body rehabilitation action recognition method
US20220207921A1 (en) Motion recognition method, storage medium, and information processing device
US11127181B2 (en) Avatar facial expression generating system and method of avatar facial expression generation
CN113298047B (en) 3D form and posture estimation method and device based on space-time correlation image
CN115482485A (en) Video processing method and device, computer equipment and readable storage medium
CN113627396B (en) Rope skipping counting method based on health monitoring
CN112489129A (en) Pose recognition model training method and device, pose recognition method and terminal equipment
CN116128814A (en) Standardized acquisition method and related device for tongue diagnosis image
CN115346272A (en) Real-time tumble detection method based on depth image sequence
EP3872694A1 (en) Avatar facial expression generating system and method of avatar facial expression generation
Ilyas et al. Rehabilitation of traumatic brain injured patients: Patient mood analysis from multimodal video
Yuan et al. Pain intensity recognition from masked facial expressions using swin-transformer
Brieva et al. An intelligent human fall detection system using a vision-based strategy
CN116311497A (en) Tunnel worker abnormal behavior detection method and system based on machine vision
Tran et al. Intelligent non-invasive vital signs estimation from image analysis
CN107480604A (en) Gait recognition method based on the fusion of more contour features
CN112233769A (en) Recovery system after suffering from illness based on data acquisition

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