CN107357419B - Action recognition system and method based on common shrinkage rate - Google Patents

Action recognition system and method based on common shrinkage rate Download PDF

Info

Publication number
CN107357419B
CN107357419B CN201710441909.4A CN201710441909A CN107357419B CN 107357419 B CN107357419 B CN 107357419B CN 201710441909 A CN201710441909 A CN 201710441909A CN 107357419 B CN107357419 B CN 107357419B
Authority
CN
China
Prior art keywords
muscle
muscles
antagonistic
myoelectric
active
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.)
Active
Application number
CN201710441909.4A
Other languages
Chinese (zh)
Other versions
CN107357419A (en
Inventor
杜文静
王磊
李慧慧
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.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
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 Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Priority to CN201710441909.4A priority Critical patent/CN107357419B/en
Publication of CN107357419A publication Critical patent/CN107357419A/en
Application granted granted Critical
Publication of CN107357419B publication Critical patent/CN107357419B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
    • G06F3/015—Input arrangements based on nervous system activity detection, e.g. brain waves [EEG] detection, electromyograms [EMG] detection, electrodermal response detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2203/00—Indexing scheme relating to G06F3/00 - G06F3/048
    • G06F2203/01—Indexing scheme relating to G06F3/01
    • G06F2203/011—Emotion or mood input determined on the basis of sensed human body parameters such as pulse, heart rate or beat, temperature of skin, facial expressions, iris, voice pitch, brain activity patterns
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching

Landscapes

  • Engineering & Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Neurosurgery (AREA)
  • General Health & Medical Sciences (AREA)
  • Neurology (AREA)
  • Health & Medical Sciences (AREA)
  • Dermatology (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Biomedical Technology (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The application relates to a common shrinkage rate-based action recognition system and a common shrinkage rate-based action recognition method, wherein the action recognition system comprises: the acquisition device acquires physiological electromyographic signals released by the motor muscles in the movement process; the filter adopts a filtering algorithm to preprocess the physiological electromyographic signals; the processor extracts the common shrinkage rate of the motor muscle from the preprocessed physiological electromyographic signals through an average electromyographic algorithm; and the processor performs characteristic identification according to the common shrinkage rate so as to identify the corresponding action of the motion muscle. The physiological electromyographic signals of the muscles are collected in a non-invasive, objective and scientific mode, and the common muscle contraction rate of the exercising muscles under different exercises is obtained, so that the different exercises are identified; according to the method, the significance difference characteristic of the common shrinkage rate in different motion modes is fully utilized, motion recognition is achieved in an objective and simple mode, and practical basis is provided for a human motion posture mode recognition system.

Description

Action recognition system and method based on common shrinkage rate
Technical Field
The application relates to the technical field of intelligence, in particular to a common shrinkage rate-based action recognition system and method.
Background
Human motion gesture recognition is a very active research direction in the field of computer vision. The human motion recognition mainly comprises the steps of detecting and tracking a moving human body in an image and video sequence, and extracting motion characteristics to represent human body behaviors, so that the motion state of the human body can be understood by using characteristic data. In recent years, with the popularization of smart phones and wearable devices, human motion recognition is widely applied in the fields of child monitoring, learning ability assessment and the like. Many motion recognition systems mainly acquire human motion information through an acceleration sensor to realize human motion detection.
Researchers at home and abroad initially insert the needle electrode into muscle to detect electromyogram, the interference is small, the positioning performance is good, and the myogram is easy to identify. After long-term research, people find that the surface electromyographic signals can also acquire the movement information of muscles, and because the surface electromyographic signals are a non-invasive detection method, the bioelectric signals issued during the activity of the neuromuscular are recorded by electrode plates adhered to the surfaces of the muscles, the operation is simple, the bioelectric signals are easy to accept by people, and therefore, the surface electromyographic signals are widely applied to the fields of clinical medicine, biomedical engineering, artificial limb bionics, pattern recognition, sports and the like. The application of surface electromyographic signals is still in a stage of development compared to needle electrode detection. The method is closely related to the basic research of the neuromuscular system, the physiological research of the neuromuscular system provides a solid foundation for the application of the electromyographic signals, and the detection of the surface electromyographic signals provides a better detection method for the research of the neuromuscular system. During the exercise of the human body, the muscles contract to release electric signals. The surface electromyogram signal is used as one of the human body movement muscle discharge physiological signals, and the operation is simple, and the real-time performance and the bionic performance are good. In recent years, many people recognize a movement pattern using electromyographic signal power, but can be applied to only a finger with a single degree of freedom.
The prior inventions CN200880012887.9 and CN201310502947.8 respectively adopt an acceleration sensor and a video sequence, and a training model of a motion mode is added, so that continuous training is required to identify different actions. For the young, the old and the sick, the long-term exercise training and the multi-node exercise sensor bring certain burden to the mind and body of the young, the old and the sick, and the technology has certain limitation. The human motion mode is recognized by adopting the portable and non-load motion mode recognition device which is simple to operate and based on the common shrinkage rate.
The movement of the human body is driven by the bones as levers, the joints as pivots and the muscle contraction as the power and is coordinated and completed under the control of the nervous system. Muscle contraction is the role of motor in human movement to provide power for human movement, and plays a very important role in movement, and the mechanical properties of muscles are very complex, so muscle research is an attractive and challenging field. The motion posture recognition is based on the dynamic state, and the motion necessarily causes the change of various organs, joints, muscles and the like of the human body. The muscle occupies the most components in human tissues, and various movements can drive the movement contraction of different muscles, so that the movement mode identification is a priority condition undoubtedly.
Therefore, in the technical field, how to simply, non-invasively, painlessly and conveniently obtain useful information from various moving tissues of a human body to perform motion recognition is a technical problem to be solved.
Disclosure of Invention
Therefore, it is necessary to provide a motion recognition system and method based on common shrinkage rate, which can perform motion recognition according to the physiological electromyographic signals released by motion in a non-invasive, painless, simple and effective manner, and has high recognition accuracy, thereby being beneficial to popularization and use.
The embodiment of the application provides a motion recognition system based on common shrinkage rate, the motion recognition system includes:
the acquisition device is used for acquiring physiological electromyographic signals released by the motor muscles in the movement process;
the filter is used for preprocessing the physiological electromyographic signals by adopting a filtering algorithm;
the processor is used for extracting the common shrinkage rate of the motor muscle from the preprocessed physiological electromyographic signals through an average electromyographic algorithm;
the processor is further configured to perform feature recognition according to the common shrinkage rate to identify a corresponding action of the exercise muscle.
The embodiment of the application also provides a common shrinkage rate-based action identification method, which comprises the following steps:
acquiring physiological electromyographic signals released by the motor muscles in the movement process;
preprocessing the physiological electromyographic signals by adopting a filtering algorithm;
extracting the common shrinkage rate of the motor muscles from the preprocessed physiological electromyographic signals through an average electromyographic algorithm;
and performing feature identification according to the common shrinkage rate to identify the corresponding action of the motion muscle.
According to the motion recognition system and method based on the common shrinkage rate, the physiological electromyographic signals released by the motor muscles in the motion process are obtained through the obtaining device, then the physiological electromyographic signals are preprocessed through the filter by adopting the filtering algorithm, finally the common shrinkage rate of the motor muscles is extracted from the preprocessed physiological electromyographic signals through the average electromyographic algorithm, and the motion corresponding to the motor muscles is recognized through feature recognition according to the common shrinkage rate. The physiological electromyographic signals of the muscles are collected in a non-invasive, objective and scientific mode, and the common muscle contraction rate of the exercising muscles under different exercises is obtained, so that the different exercises are identified; according to the method, the significance difference characteristic of the common shrinkage rate in different motion modes is fully utilized, motion recognition is achieved in an objective and simple mode, and practical basis is provided for a human motion posture mode recognition system.
Drawings
FIG. 1 is a block diagram of a common shrinkage based motion recognition system in one embodiment;
FIG. 2 is a flow diagram of a method for common shrinkage based motion recognition in one embodiment;
fig. 3 is a schematic diagram illustrating an implementation process of a common shrinkage rate-based motion recognition system in a specific application example.
Detailed Description
In one embodiment, a common shrinkage factor based motion recognition system, as shown in fig. 1, includes, but is not limited to, an acquisition device 11, a filter 12, and a processor 13.
In this embodiment, the obtaining device 11 is configured to obtain a physiological electromyographic signal released by a motor muscle during exercise, the filter 12 is configured to pre-process the physiological electromyographic signal by using a filtering algorithm, and the processor 13 is configured to extract a common shrinkage rate of the motor muscle from the pre-processed physiological electromyographic signal by using a mean electromyographic algorithm, and perform feature recognition according to the common shrinkage rate to recognize a motion corresponding to the motor muscle.
In a preferred embodiment, the obtaining device 11 is specifically configured to obtain the myoelectric values of the antagonistic muscle and the active muscle released by the motor muscle during exercise.
Further, the obtaining device 11 is specifically configured to:
when performing anteversion movements, the harvested voluntary muscles include: left and right extraabdominal oblique muscles, and intra-abdominal oblique/transverse abdominal muscles, the antagonist muscles obtained comprising: left and right erector spinae/multifidus;
when performing a reclining exercise, the harvested voluntary muscles include: left and right erector spinae/multifidus muscles, and the acquired antagonistic muscles include: left and right extraabdominal oblique muscles, intra-abdominal oblique/transverse abdominal muscles;
when performing a left leaning exercise, the harvested voluntary muscles include: left erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse, and antagonistic muscles were obtained including: right erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse;
when performing right leaning exercises, the harvested voluntary muscles include: right erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse, and antagonistic muscles were obtained including: left erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse.
The acquiring device 11 is specifically configured to acquire the myoelectric values of the antagonistic muscle and the active muscle in the standard operation of the lumbar muscle group in accordance with the standard target of the age, height, and weight indexes.
In a specific embodiment, the acquiring device 11 acquires the myoelectric values of the antagonistic muscles and the active muscles of the lumbar muscle group of the standard subject according to the age, height and weight indexes, and specifically includes that disposable electrode pieces are pasted on the lumbar muscle group of the lumbar erector spinae/multifidus muscle, the abdominal external oblique muscle and the abdominal internal oblique muscle/abdominal transverse muscle of the standard subject along the muscle fiber direction. Wherein, in order to improve the accuracy, the pasting position of the standard object needs to be firstly wiped by using 60% -80% alcohol, and preferably, the pasting position of the standard object needs to be wiped by using 75% alcohol.
It should be noted that the processor 13 is specifically configured to extract an electromyography value of an antagonistic muscle of the motor muscle and an electromyography value of an active muscle from a physiological electromyography signal through a mean electromyography algorithm; calculating the common contraction rate according to the myoelectric value of the antagonistic muscle and the myoelectric value of the active muscle, wherein the calculation formula comprises but is not limited to the following formula:
Figure BDA0001320210570000041
wherein CCR is the common shrinkage, AEMGAntagonistic muscleTo antagonize the myoelectric value of the muscle, AEMGActive muscleThe myoelectric value of the active muscle.
It should be noted that, in this embodiment, in order to improve the stability and the accuracy of the identification of the system, a manner of obtaining and averaging multiple times may be adopted, specifically, the myoelectric value of the antagonistic muscle and the myoelectric value of the active muscle are the average myoelectric value of the antagonistic muscle and the average myoelectric value of the active muscle respectively, and the calculation formula adopted by the processor 13 includes:
Figure BDA0001320210570000051
Figure BDA0001320210570000052
wherein n is a natural number greater than 0, and | data [ i ] | is the myoelectric value of the ith antagonistic muscle, | data [ j ] | or the myoelectric value of the jth active muscle.
The processor 13 performs feature recognition according to the common shrinkage rate to identify the corresponding action of the exercise muscle, which includes, but is not limited to, the following processes:
identifying a corresponding motion of the motor muscle as a forward leaning motion when the common contraction rate is about equal to 0.2986;
when the common shrinkage rate is equal to 0.7219, identifying that the corresponding action of the motor muscle is a backward movement;
when the common contraction rate is equal to 0.5255, identifying that the corresponding action of the motor muscle is a left-leaning movement;
when the common contraction rate is equal to 0.4723, identifying that the corresponding action of the motor muscle is right-leaning movement;
wherein if a forward tilting motion is identified, the features are included: the level of voluntary muscle contraction is significantly greater than that of supine, left-leaning, and right-leaning movements, and greater than that of the antagonistic muscle itself; if a lean-back motion is identified, the features are included: the level of voluntary muscle contraction is significantly less than the level of antagonistic muscle contraction; if a left or right leaning motion is identified, the features are included: the levels of contraction of the active and antagonistic muscles are substantially the same.
The filter 12 is a 35-500Hz band-pass filter and is used for removing 50Hz power frequency signal interference noise corresponding to a power supply voltage of 220V and removing electrocardiosignal 0.25-35Hz frequency band interference noise.
The physiological electromyographic signals of the muscles are collected in a non-invasive, objective and scientific mode, and the common muscle contraction rate of the exercising muscles under different exercises is obtained, so that the different exercises are identified; according to the method, the significance difference characteristic of the common shrinkage rate in different motion modes is fully utilized, motion recognition is achieved in an objective and simple mode, and practical basis is provided for a human motion posture mode recognition system.
Referring next to fig. 2, in one embodiment, a method for motion recognition based on a common shrinkage rate is provided, which includes, but is not limited to, the following steps.
S201, acquiring physiological electromyographic signals released by the motor muscles in the movement process.
In a preferred embodiment, the S201 may specifically include: and acquiring the myoelectric values of antagonistic muscles and the myoelectric values of active muscles released by the motor muscles in the exercise process.
Further, the S201 may include the following acquiring process:
when performing anteversion movements, the harvested voluntary muscles include: left and right extraabdominal oblique muscles, and intra-abdominal oblique/transverse abdominal muscles, the antagonist muscles obtained comprising: left and right erector spinae/multifidus;
when performing a reclining exercise, the harvested voluntary muscles include: left and right erector spinae/multifidus muscles, and the acquired antagonistic muscles include: left and right extraabdominal oblique muscles, intra-abdominal oblique/transverse abdominal muscles;
when performing a left leaning exercise, the harvested voluntary muscles include: left erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse, and antagonistic muscles were obtained including: right erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse;
when performing right leaning exercises, the harvested voluntary muscles include: right erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse, and antagonistic muscles were obtained including: left erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse.
Specifically, the present embodiment can acquire the myoelectric values of the antagonistic muscle and the active muscle of the lumbar muscle group in the standard action according to the age, height, and weight indexes.
In a specific embodiment, a disposable electrode sheet may be attached to the lumbar muscle groups of the erector spinae/multifidus muscle, the extraabdominal oblique muscle, and the intra-abdominal oblique/transverse muscle of the standard subject in the muscle fiber direction. Wherein, in order to improve the accuracy, the pasting position of the standard object needs to be firstly wiped by using 60% -80% alcohol, and preferably, the pasting position of the standard object needs to be wiped by using 75% alcohol.
S202, preprocessing the physiological electromyographic signals by adopting a filtering algorithm.
In S202 of this embodiment, a 35-500Hz band-pass filter may be used to remove the 50Hz power frequency signal interference noise corresponding to the 220v power supply voltage, and to remove the frequency band interference noise of the 0.25-35Hz electrocardiosignal.
S203, extracting the common shrinkage rate of the motor muscle from the preprocessed physiological electromyographic signals through an average electromyographic algorithm.
In S203, the extracting, by using an average electromyography algorithm, the common contraction rate of the motor muscle from the preprocessed physiological electromyography signals in this embodiment specifically includes:
extracting myoelectric values of antagonistic muscles and active muscles of the motor muscles from physiological myoelectric signals through an average myoelectric algorithm;
calculating the common shrinkage rate according to the electromyographic value of the antagonistic muscle and the electromyographic value of the active muscle, wherein the calculation formula comprises the following steps:
Figure BDA0001320210570000071
wherein CCR is the common shrinkage, AEMGAntagonistic muscleTo antagonize the myoelectric value of the muscle, AEMGActive muscleThe myoelectric value of the active muscle.
The myoelectric value of the antagonistic muscle and the myoelectric value of the active muscle are respectively an average myoelectric value of the antagonistic muscle and an average myoelectric value of the active muscle, and the calculation formula comprises:
Figure BDA0001320210570000072
Figure BDA0001320210570000073
wherein n is a natural number greater than 0, and | data [ i ] | is the myoelectric value of the ith antagonistic muscle, | data [ j ] | or the myoelectric value of the jth active muscle.
And S204, performing feature identification according to the common shrinkage rate to identify the corresponding action of the motion muscle.
In S204, this embodiment specifically includes, but is not limited to, the following processes:
identifying a corresponding motion of the motor muscle as a forward leaning motion when the common contraction rate is about equal to 0.2986;
when the common shrinkage rate is equal to 0.7219, identifying that the corresponding action of the motor muscle is a backward movement;
when the common contraction rate is equal to 0.5255, identifying that the corresponding action of the motor muscle is a left-leaning movement;
when the common contraction rate is equal to 0.4723, identifying that the corresponding action of the motor muscle is right-leaning movement;
wherein if a forward tilting motion is identified, the features are included: the level of voluntary muscle contraction is significantly greater than that of supine, left-leaning, and right-leaning movements, and greater than that of the antagonistic muscle itself; if a lean-back motion is identified, the features are included: the level of voluntary muscle contraction is significantly less than the level of antagonistic muscle contraction; if a left or right leaning motion is identified, the features are included: the levels of contraction of the active and antagonistic muscles are substantially the same.
Referring to fig. 3, in a specific application example, the common shrinkage rate-based motion recognition system may include, but is not limited to, a lumbar muscle group electromyographic signal acquisition module, an electromyographic signal preprocessing module, a common shrinkage rate feature extraction module, a motion posture recognition module, and the like.
It is understood that the lumbar muscle group electromyographic signal acquisition module specifically needs to carry out volunteer recruitment, lumbar muscle group selection and exercise posture scheme design.
For example, since the electromyographic signals are influenced by age, fat and other factors, the healthy population recruiting part mainly screens volunteers with indexes of age, height, weight and the like matched with each other. The waist has the functions of holding the upper part and the lower part of the body like the back of an arrow, contracting when the arrow is pulled, expanding when the arrow is put on, supporting the balance of the upper body, maintaining the body posture, and matching with the lower body to maintain the stable gravity center. Meanwhile, the waist is the position with the largest strength of the human body, and the waist can initiate the release action of the force in the actions of throwing, whiplash and legs and the like. Therefore, the lumbar muscle group plays a decisive role in the overall movement of the body.
As described above, in this embodiment, disposable electrode sheets may be used, and the disposable electrode sheets may be attached to a plurality of muscle groups, such as the lumbar erector spinae/multifidus muscle, the extraabdominal oblique muscle, and the intra-abdominal oblique muscle/transverse abdominal muscle, which have been wiped with 75% alcohol, and attached along the direction of the muscle fibers. When a tester is in four movement postures of forward leaning, backward leaning, left leaning and right leaning, analog signals generated by muscles can be transmitted to a receiving device of an MP150 (multi-conductive physiological recorder) through a transmitting module of a BIOPAC (multi-conductive physiological recorder), the receiving device carries out analog-to-digital conversion, and the muscle analog signals are converted into one-dimensional random voltage signals, so that the electromyographic signal acquisition of the lumbar muscle group electromyographic signal acquisition module is completed.
The electromyographic signal preprocessing module mainly comprises filtering processing and standardization processing for band-pass filtering and power frequency denoising. It is understood that, because the effective frequency band of the electromyographic signals is between 10 and 500Hz, 50Hz of power frequency interference generated by 220v of voltage in China has a great influence on the electromyographic signals, and therefore, the power frequency signals need to be filtered. As the upper half of the human body is influenced by the heart beating, the main frequency band of the electrocardiosignal is 0.25-35Hz, which has certain interference to the electric signal generated by the muscle movement, and the electromyographic signal acquired for this purpose is subjected to 35-500Hz band-pass filtering processing.
It should be noted that the common shrinkage rate refers to a ratio of the average myoelectric value of the antagonistic muscle to the total average myoelectric value of the active muscle and the antagonistic muscle when the human body does exercise, which reflects the muscle coordination, wherein the common shrinkage rate feature extraction module of the embodiment adopts the calculation method mentioned in the above embodiment:
Figure BDA0001320210570000091
similarly, the average muscle discharge value is the average of the discharge amount of muscle over a period of time, calculated in the manner described in the above example:
Figure BDA0001320210570000092
Figure BDA0001320210570000093
wherein n is a natural number greater than 0, and | data [ i ] | is the myoelectric value of the ith antagonistic muscle, | data [ j ] | or the myoelectric value of the jth active muscle.
The specific operation principle of the motion gesture recognition module of the present embodiment is as follows. When the human body performs anteversion movement, the active muscles are: left and right extraabdominal oblique and intraabdominal oblique/transverse, antagonistic muscles are: left and right erector spinae/multifidus; when doing supine exercise, the active muscles are left and right erector spinae/multifidus muscles, and the antagonistic muscles are: left and right extraabdominal oblique and intra-abdominal oblique/transverse muscles; in a left-leaning exercise, the voluntary muscles are: left erector spinae/multifidus, extraabdominal oblique and intraabdominal oblique/transverse, antagonistic muscles are: right erector spinae/multifidus, extraabdominal oblique and intraabdominal oblique/transverse; during right-leaning exercise, the voluntary muscles are: right erector spinae/multifidus, extraabdominal oblique and intraabdominal oblique/transverse, antagonistic muscles are: left erector spinae/multifidus, extraabdominal oblique and intraabdominal oblique/transverse.
Under different movement modes, the average myoelectric value difference comparative value of the active muscle and the antagonistic muscle is as follows:
during anteversion movement: the average myoelectric value of the active muscle is about 0.45, and the average myoelectric value of the antagonistic muscle is about 0.19; when doing backward movement: the average myoelectric value of the active muscle is about 0.14, and the average myoelectric value of the antagonistic muscle is about 0.39; when the user leans to the left: the average myoelectric value of the active muscle is about 0.27, and the average myoelectric value of the antagonistic muscle is about 0.3; when the user leans rightly: the average myoelectric value of the active muscles is about 0.29 and the average myoelectric value of the antagonistic muscles is about 0.26. Therefore, the difference between the active muscles and the difference between the antagonistic muscles under each movement can be calculated and deduced.
It can be seen that in the anteversion, supination, left incline and right incline motor patterns, the level of voluntary muscle contraction for anteversion movement is significantly greater than that for supination, left incline and right incline motor patterns, and greater than the level of voluntary antagonistic muscle contraction; when the forward leaning exercise is performed, the main role of the muscle group of the active muscles is to coordinate the forward leaning exercise of the human body. In the backward movement, the contraction level of the active muscle is obviously lower than that of the antagonistic muscle, which indicates that the antagonistic muscle mainly coordinates the human body to perform the backward movement. In the leftward and rightward inclination movement, the contraction levels of the active muscles and the antagonistic muscles are basically consistent, and the active muscles and the antagonistic muscles coordinate the human body to perform the leftward and rightward inclination movement. The active and antagonistic muscles have significant variability between the four motor patterns of anteversion, supination, left inclination and right inclination, each p <0.05 in this example.
The differences in the common shrinkage rates among the four motion patterns of forward lean, backward lean, left lean and right lean are as follows:
the common shrinkage rate was about 0.2986 for forward leaning, about 0.7219 for backward leaning, about 0.5255 for left leaning and about 0.4723 for right leaning. The difference between each motion was calculated to be about 0.000, i.e., P-0.000.
Therefore, it is easy to see whether the four motion modes can be distinguished, mainly solving the difference of the characteristic parameters in the modes, and if the characteristic parameters have no significant difference in the four modes, it indicates that the four modes cannot be identified. By comparing the common shrinkage rate of each mode, the common shrinkage rate of the anteversion movement is found to have statistical significance difference (p is less than 0.05) with the retroversion movement, the left-leaning movement and the right-leaning movement; the common shrinkage rate of the backward tilting motion and the forward tilting motion, the left tilting motion and the right tilting motion have obvious difference respectively (p is less than 0.05); the common shrinkage rate of the left-leaning motion and the forward-leaning motion mode, the backward-leaning motion mode and the right-leaning motion mode have obvious difference (p is less than 0.05); the common shrinkage rate of the right-leaning movement is different from the forward-leaning movement mode, the backward-leaning movement mode and the left-leaning movement mode (p is less than 0.05), so that the common shrinkage rates of the forward-leaning movement mode, the backward-leaning movement mode, the left-leaning movement mode and the right-leaning movement mode are obviously different, and the movement modes can be identified by the common shrinkage rates.
The physiological electromyographic signals of the muscles are collected in a non-invasive, objective and scientific mode, and the common muscle contraction rate of the exercising muscles under different exercises is obtained, so that the different exercises are identified; according to the method, the significance difference characteristic of the common shrinkage rate in different motion modes is fully utilized, motion recognition is achieved in an objective and simple mode, and practical basis is provided for a human motion posture mode recognition system.
The above description is only an example of the present application and is not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings, or which are directly or indirectly applied to other related technical fields, are intended to be included within the scope of the present application.

Claims (8)

1. A common shrinkage rate-based motion recognition system, the motion recognition system comprising:
the acquisition device is used for acquiring physiological electromyographic signals released by the motor muscles in the movement process;
the filter is used for preprocessing the physiological electromyographic signals by adopting a filtering algorithm;
the processor is used for extracting the common shrinkage rate of the motor muscle from the preprocessed physiological electromyographic signals through an average electromyographic algorithm;
the processor is further used for performing feature recognition according to the common shrinkage rate to recognize corresponding actions of the motor muscles;
the processor is specifically configured to:
extracting myoelectric values of antagonistic muscles and active muscles of the motor muscles from physiological myoelectric signals through an average myoelectric algorithm;
calculating the common shrinkage rate according to the electromyographic value of the antagonistic muscle and the electromyographic value of the active muscle, wherein the calculation formula comprises the following steps:
Figure FDA0002353333460000011
wherein CCR is the common shrinkage, AEMGAntagonistic muscleTo antagonize the myoelectric value of the muscle, AEMGActive muscleIs the electromyographic value of the active muscle;
the myoelectric value of the antagonistic muscle and the myoelectric value of the active muscle are respectively an average myoelectric value of the antagonistic muscle and an average myoelectric value of the active muscle, and a calculation formula adopted by the processor comprises:
Figure FDA0002353333460000012
Figure FDA0002353333460000013
wherein n is a natural number greater than 0, and | data [ i ] | is the myoelectric value of the ith antagonistic muscle, | data [ j ] | or the myoelectric value of the jth active muscle.
2. The motion recognition system of claim 1, wherein the obtaining means is specifically configured to:
and acquiring the myoelectric values of antagonistic muscles and the myoelectric values of active muscles released by the motor muscles in the exercise process.
3. The motion recognition system of claim 2, wherein the obtaining means is specifically configured to:
when performing anteversion movements, the harvested voluntary muscles include: left and right extraabdominal oblique muscles, and intra-abdominal oblique/transverse abdominal muscles, the antagonist muscles obtained comprising: left and right erector spinae/multifidus;
when performing a reclining exercise, the harvested voluntary muscles include: left and right erector spinae/multifidus muscles, and the acquired antagonistic muscles include: left and right extraabdominal oblique muscles, intra-abdominal oblique/transverse abdominal muscles;
when performing a left leaning exercise, the harvested voluntary muscles include: left erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse, and antagonistic muscles were obtained including: right erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse;
when performing right leaning exercises, the harvested voluntary muscles include: right erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse, and antagonistic muscles were obtained including: left erector spinae/multifidus, extraabdominal oblique, intraabdominal oblique/transverse.
4. The motion recognition system of claim 3, wherein the processor performs feature recognition based on the common contraction rate to recognize the corresponding motion of the exercise muscle, and specifically comprises:
identifying a corresponding motion of the motor muscle as a forward leaning motion when the common contraction rate is about equal to 0.2986;
when the common shrinkage rate is equal to 0.7219, identifying that the corresponding action of the motor muscle is a backward movement;
when the common contraction rate is equal to 0.5255, identifying that the corresponding action of the motor muscle is a left-leaning movement;
when the common contraction rate is equal to 0.4723, identifying that the corresponding action of the motor muscle is right-leaning movement;
wherein if a forward tilting motion is identified, the features are included: the level of voluntary muscle contraction is significantly greater than that of supine, left-leaning, and right-leaning movements, and greater than that of the antagonistic muscle itself; if a lean-back motion is identified, the features are included: the level of voluntary muscle contraction is significantly less than the level of antagonistic muscle contraction; if a left or right leaning motion is identified, the features are included: the levels of contraction of the active and antagonistic muscles are substantially the same.
5. The motion recognition system of claim 2, wherein the obtaining means is specifically configured to:
and acquiring the myoelectric values of antagonistic muscles and the myoelectric values of active muscles of the lumbar muscle group of the standard object meeting the age, height and weight indexes when the lumbar muscle group performs standard actions.
6. The motion recognition system according to claim 5, wherein the obtaining device obtains the myoelectric values of the antagonistic muscles and the active muscles of the lumbar muscle group of the standard subject according to the age, height and weight indexes, and specifically comprises:
and a disposable electrode plate is stuck to the lumbar muscle groups of the lumbar erector spinae/multifidus muscle, the extraabdominal oblique muscle and the intra-abdominal oblique muscle/transverse muscle of the standard object along the muscle fiber direction.
7. The motion recognition system of claim 1, wherein the filter is a 35-500Hz band-pass filter, and is used for removing the interference noise of 50Hz power frequency signal corresponding to 220v power supply voltage and removing the interference noise of 0.25-35Hz frequency band of the electrocardiosignal.
8. A motion recognition method based on a common shrinkage rate is characterized by comprising the following steps:
acquiring physiological electromyographic signals released by the motor muscles in the movement process;
preprocessing the physiological electromyographic signals by adopting a filtering algorithm;
extracting the common shrinkage rate of the motor muscles from the preprocessed physiological electromyographic signals through an average electromyographic algorithm;
performing feature recognition according to the common shrinkage rate to recognize corresponding actions of the motor muscles;
the method for extracting the common shrinkage rate of the motor muscles from the preprocessed physiological electromyographic signals through the average electromyographic algorithm specifically comprises the following steps:
extracting myoelectric values of antagonistic muscles and active muscles of the motor muscles from physiological myoelectric signals through an average myoelectric algorithm;
calculating the common shrinkage rate according to the electromyographic value of the antagonistic muscle and the electromyographic value of the active muscle, wherein the calculation formula comprises the following steps:
Figure FDA0002353333460000041
wherein CCR is the common shrinkage, AEMGAntagonistic muscleTo antagonize the myoelectric value of the muscle, AEMGActive muscleIs the electromyographic value of the active muscle;
the myoelectric value of the antagonistic muscle and the myoelectric value of the active muscle are respectively an average myoelectric value of the antagonistic muscle and an average myoelectric value of the active muscle, and the calculation formula comprises:
Figure FDA0002353333460000042
Figure FDA0002353333460000043
wherein n is a natural number greater than 0, and | data [ i ] | is the myoelectric value of the ith antagonistic muscle, | data [ j ] | or the myoelectric value of the jth active muscle.
CN201710441909.4A 2017-06-13 2017-06-13 Action recognition system and method based on common shrinkage rate Active CN107357419B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710441909.4A CN107357419B (en) 2017-06-13 2017-06-13 Action recognition system and method based on common shrinkage rate

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710441909.4A CN107357419B (en) 2017-06-13 2017-06-13 Action recognition system and method based on common shrinkage rate

Publications (2)

Publication Number Publication Date
CN107357419A CN107357419A (en) 2017-11-17
CN107357419B true CN107357419B (en) 2020-05-19

Family

ID=60273885

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710441909.4A Active CN107357419B (en) 2017-06-13 2017-06-13 Action recognition system and method based on common shrinkage rate

Country Status (1)

Country Link
CN (1) CN107357419B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108198623A (en) * 2017-12-15 2018-06-22 东软集团股份有限公司 Human body condition detection method, device, storage medium and electronic equipment
CN110859619B (en) * 2018-08-28 2023-02-14 易适康连（上海）科技有限公司 Device, system and method for acquiring abdominal electromyogram data
CN110232976B (en) * 2019-07-01 2023-05-02 上海电机学院 Behavior identification method based on waist and shoulder surface myoelectricity measurement
CN113288532B (en) * 2021-05-31 2023-04-07 北京京东乾石科技有限公司 Myoelectric control method and device

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104107134A (en) * 2013-12-10 2014-10-22 中山大学 Myoelectricity feedback based upper limb training method and system
CN105148404A (en) * 2015-09-16 2015-12-16 燕铁斌 Electrical stimulation training system and training method based on health-side limb electromyographic signal trigger

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2613276A1 (en) * 2012-01-04 2013-07-10 Gabriele Ceruti Method and apparatus for neuromotor rehabilitation using interactive setting systems

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104107134A (en) * 2013-12-10 2014-10-22 中山大学 Myoelectricity feedback based upper limb training method and system
CN105148404A (en) * 2015-09-16 2015-12-16 燕铁斌 Electrical stimulation training system and training method based on health-side limb electromyographic signal trigger

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
肌肉共同收缩研究进展;魏勇，刘宇;《中国体育科技》;20091231(第05期);第54-59页 *

Also Published As

Publication number Publication date
CN107357419A (en) 2017-11-17

Similar Documents

Publication Publication Date Title
CN107357419B (en) Action recognition system and method based on common shrinkage rate
EP3273852B1 (en) Muscle activity monitoring
EP3065628B1 (en) Biomechanical activity monitoring
CN103212156B (en) The autonomous instrument of hemiplegic patient based on body image theory
Farrell A comparison of the effects of electrode implantation and targeting on pattern classification accuracy for prosthesis control
CN107928980A (en) A kind of autonomous rehabilitation training system of the hand of hemiplegic patient and training method
CN109288651A (en) Personalized upper-limbs rehabilitation training robot system and its recovery training method
CN108968952A (en) A kind of brain myoelectricity and Inertia information synchronous acquisition device
CN111091890B (en) Health cognitive system under self-excitation and feedback control
CN206228345U (en) Intelligent clothing for improving training effect
Shusong et al. EMG-driven computer game for post-stroke rehabilitation
US20190320944A1 (en) Biomechanical activity monitoring
CN114822761A (en) Wrist rehabilitation training system based on muscle cooperation and variable stiffness impedance control
US10912512B2 (en) System and method for muscle movements detection
CN106267557A (en) A kind of brain control based on wavelet transformation and support vector machine identification actively upper limb medical rehabilitation training system
Andrews et al. Optimal electrode configurations for finger movement classification using EMG
CN113995956B (en) Stroke electrical stimulation training intention recognition device based on myoelectricity expected posture adjustment
CN113713333B (en) Dynamic virtual induction method and system for lower limb rehabilitation full training process
KR101492478B1 (en) Locomotion Mode Pattern Recognition System using Fusion Sensor
Geng et al. Comparison of electromyography and mechanomyogram in control of prosthetic system in multiple limb positions
CN208302083U (en) A kind of pelvic diaphragm muscle instrument for training
Giuffrida et al. Upper-extremity stroke therapy task discrimination using motion sensors and electromyography
CN205994970U (en) A kind of EMG feedback respiratory training clothes
CN215017698U (en) Rehabilitation training motion simulation visualization system
CN115068286A (en) Wearable data acquisition device based on rehabilitation massage therapy and use method

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