CN110200601B - Pulse condition acquisition device and system - Google Patents
Pulse condition acquisition device and system Download PDFInfo
- Publication number
- CN110200601B CN110200601B CN201910522341.8A CN201910522341A CN110200601B CN 110200601 B CN110200601 B CN 110200601B CN 201910522341 A CN201910522341 A CN 201910522341A CN 110200601 B CN110200601 B CN 110200601B
- Authority
- CN
- China
- Prior art keywords
- pulse
- information
- position information
- feeling
- acquiring
- 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
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
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Physiology (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Engineering & Computer Science (AREA)
- Cardiology (AREA)
- Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
The invention discloses a pulse condition acquisition device, comprising: the stereoscopic vision acquisition module is used for acquiring an original image of the wrist of the human body and with depth information by using a stereoscopic camera; the pulse feeling position information acquisition module is used for inputting the original image into machine learning software to acquire pulse feeling position information; the pulse condition information acquisition module is used for outputting a motor control signal corresponding to pulse feeling position information to the three-axis motor and acquiring pulse condition information by using a pulse diagnosis probe connected with an output shaft of the three-axis motor; and the pulse condition information display module is used for displaying the pulse condition information on the display screen. In the whole working process of the device for acquiring the pulse condition information, all pulse conditions of cun, guan and chi of a patient can be accurately acquired without intervention of a doctor. The invention also discloses a pulse condition acquisition system which has corresponding technical effects.
Description
Technical Field
The invention relates to the technical field of medical instruments, in particular to a pulse condition acquisition device and system.
Background
The traditional Chinese medicine diagnoses the cause of disease by feeling pulse. The pulse feeling means that the doctor can feel the image of pulse corresponding to the finger by cutting and pressing the cun, guan and chi parts of the cun-kou radial artery of the patient with the fingers to understand the state of an illness and judge the diagnosis and treatment method of the disease. The cun-guan-chi region of the left and right hand indicates different organs, i.e., the right cun-lung, the right guan-spleen, the left cun-heart, the left guan-liver, and the right-chi region the kidney. Cun Hou vessel has the symptoms of exogenous fever, vasodilatation and superficial vessels, which makes the former superficial vessels more obvious. The right guan pertains to the spleen and the left guan pertains to the liver, as explained in Qing Yi Zong jin Jian: "diet impairment is excessive, so the right guan Mai is powerful; fatigue impairs qi, so the right guan pulse is weak. The ulnar pulse, which is the kidney-waiting and gate of life, is the primordial qi of a person, and when the person is seriously ill, the primordial qi is weak, the heart force is insufficient, the pulse is thready and weak, and the ulnar pulse is originally deep and weak, so the pulse is reflected weaker. Cun guan chi reflects the physiological changes of the corresponding organs and is an important basis for determining the pulse feeling.
Although the pulse diagnosis has a long history and a great deal of theory and practice, the pulse diagnosis still has problems in the present society, such as the standardization of pulse condition and the technical method of pulse diagnosis. Regarding the standardization of pulse conditions, the concept of pulse conditions in traditional Chinese medicine is general, and the specific judgment standard is fuzzy, so that the objective judgment standard is lacked, and the accuracy of pulse diagnosis is influenced. Regarding the pulse-taking technical method, the counting method of pulse-taking in traditional Chinese medicine is called rest, which means that doctors calculate the pulse count of patients by taking their own breath as the time unit in pulse-taking. The method for counting the rest is convenient to use, does not need any instrument, has certain scientific basis, but has obvious defects, such as inaccurate time calculation and great difficulty in medical diagnosis and teaching.
In recent years, more pulse diagnosis instruments/devices are on the market. For example, a device for distinguishing pulse conditions by acquiring blood vessel pressure distribution signals and pulse wave conduction velocity of the cun-guan-chi part of a human body is adopted, but the device has high requirements on signal processing, is difficult to realize and has high economic cost. In addition, all the current pulse diagnosis instruments use manual positioning, a pressure sensor for measuring blood pressure is manually placed at the cun-guan-chi position of a patient observed by human eyes, and pulse conditions are collected. This requires that the person collecting the pulse condition can accurately judge the cun-guan-chi position, and has the disadvantages of slow speed and labor consumption.
In summary, how to effectively solve the problems of automatically and accurately acquiring pulse condition information by a machine is a technical problem that needs to be solved urgently by those skilled in the art at present.
Disclosure of Invention
The invention aims to provide a pulse condition acquisition device and a pulse condition acquisition system, which are used for realizing automatic and accurate acquisition of pulse condition information by a machine so as to quickly acquire accurate pulse conditions and are beneficial to accurate diagnosis of doctors.
In order to solve the technical problems, the invention provides the following technical scheme:
a pulse manifestation obtaining apparatus comprising:
the stereoscopic vision acquisition module is used for acquiring an original image of the wrist of the human body and with depth information by using a stereoscopic camera;
the pulse feeling position information acquisition module is used for inputting the original image into machine learning software to acquire pulse feeling position information;
the pulse condition information acquisition module is used for outputting a motor control signal corresponding to the pulse feeling position information to the three-axis motor and acquiring pulse condition information by using a pulse diagnosis probe connected with an output shaft of the three-axis motor;
and the pulse condition information display module is used for displaying the pulse condition information on a display screen.
Preferably, the pulse-taking position information obtaining module includes:
the image preprocessing unit is used for preprocessing the original image by utilizing image processing software to obtain the wrist width, the origin of pulse feeling and a target image after background removal; wherein, the origin of pulse feeling is the intersection point of the connecting lines of the wrist and the ulnar, guan and cun pulses; inputting the target image, the pulse feeling origin and the wrist width into the machine learning software to obtain the pulse feeling position information.
Preferably, the pulse condition information acquisition module includes:
the motor control unit is used for sequentially and respectively outputting motor control signals corresponding to the ulnar position information, the guan-mai position information and the cun-mai position information to the three-axis motor when the pulse feeling position information comprises the ulnar position information, the guan-mai position information and the cun-mai position information;
and the pulse condition acquisition unit is used for sequentially acquiring ulnar pulse condition information, guan pulse condition information and cun pulse condition information by using the pulse diagnosis probe.
Preferably, the pulse condition information acquisition module is specifically configured to output the motor control signals with different pressure levels and corresponding to the pulse feeling position information to a vertical motor in the three-axis motor, and acquire pulse condition information corresponding to the different pressure levels by using the pulse diagnosis probe.
Preferably, the output shaft of the three-shaft motor is connected with the pulse-taking probe through a universal joint.
Preferably, the method further comprises the following steps: a machine learning module;
wherein the machine learning module comprises:
the training data acquisition unit is used for acquiring an image training sample with a label and depth information; the labels comprise ulnar position information, guan mai position information and cun mai position information;
and the training module is used for training initial machine learning software by using the image training sample to obtain the machine learning software.
Preferably, the method further comprises the following steps:
and the pulse condition result acquisition module is used for analyzing the pulse condition information to obtain a pulse diagnosis result.
A pulse manifestation acquisition system comprising:
the pulse diagnosis device comprises a computer, a stereo camera, a three-axis motor and a pulse diagnosis probe, wherein the computer is provided with visual positioning software and a display screen, the stereo camera is in communication connection with the computer, the three-axis motor is in communication connection with the computer, and the pulse diagnosis probe is connected to an output shaft of the three-axis motor;
the three-dimensional camera is used for acquiring an original image of the wrist of a human body and with depth information and sending the original image to the computer;
the computer is used for acquiring pulse feeling position information by using visual positioning software, outputting a motor control signal corresponding to the pulse feeling position information to the three-axis motor, and displaying the pulse feeling information on the display screen after receiving the pulse feeling information; the vision positioning software inputs the original image into a trained machine learning module to carry out pulse feeling position learning so as to obtain pulse feeling position information;
the three-axis motor is used for driving the pulse-taking probe to move under the control of the motor control signal;
the pulse diagnosis probe is used for acquiring the pulse condition information and sending the pulse condition information to the computer.
The device provided by the embodiment of the invention comprises: the stereoscopic vision acquisition module is used for acquiring an original image of the wrist of the human body and with depth information by using a stereoscopic camera; the pulse feeling position information acquisition module is used for inputting the original image into the trained machine learning software to acquire pulse feeling position information; the pulse condition information acquisition module is used for outputting a motor control signal corresponding to pulse feeling position information to the three-axis motor and acquiring pulse condition information by using a pulse diagnosis probe connected with an output shaft of the three-axis motor; and the pulse condition information display module is used for displaying the pulse condition information on the display screen.
In the device, trained machine learning software is used for carrying out vein-cutting position information on the original image acquired by the stereo camera to obtain the vein-cutting position information. Then, based on the pulse feeling position information, the three-axis motor is controlled to move, so that the pulse diagnosis probe connected with the output shaft of the three-axis motor collects pulse condition information, and then the pulse condition information is displayed on a display screen. Compared with the existing pulse diagnosis device, the pulse diagnosis device has the advantages that the pulse diagnosis probe does not need to be manually moved to the pulse diagnosis position, and the pulse diagnosis position can be determined. In the whole working process of the device for acquiring the pulse condition information, all pulse conditions of cun, guan and chi of a patient can be accurately acquired without intervention of a doctor. Namely, the device can realize that the machine can automatically and accurately acquire the pulse condition information, and is beneficial to the accurate diagnosis of doctors.
Accordingly, embodiments of the present invention further provide a pulse condition obtaining system corresponding to the pulse condition obtaining apparatus, which has the above technical effects, and are not described herein again.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic structural diagram of a pulse condition acquiring apparatus according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of the determination of the position of the elbow in accordance with an embodiment of the present invention;
FIG. 3 is a schematic view of a gimbal connection according to an embodiment of the present invention;
FIG. 4 is a schematic structural diagram of a pulse acquisition system according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a pulse acquisition system according to an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the disclosure, the invention will be described in further detail with reference to the accompanying drawings and specific embodiments. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The first embodiment is as follows:
referring to fig. 1, fig. 1 is a schematic structural diagram of a pulse condition obtaining apparatus according to an embodiment of the present invention, the apparatus including: the system comprises a stereoscopic vision acquisition module 101, a pulse feeling position information acquisition module 102, a pulse condition information acquisition module 103 and a pulse condition information display module 104;
the stereoscopic vision acquisition module is used for acquiring an original image of the wrist of the human body and with depth information by using the stereoscopic camera.
In the embodiment of the invention, the type of the stereo camera is not limited, as long as the stereo camera can acquire the original image of the wrist of the human body and with the depth information, and the acquired original image can meet the definition requirement of computer visual information. In other embodiments of the present invention, two flat cameras may also be used to acquire the raw image. The original image may be a three-dimensional image of the wrist covering the position required for pulse feeling.
And the pulse feeling position information acquisition module is used for inputting the original image into the trained machine learning software to carry out pulse feeling position learning and acquiring pulse feeling position information.
In the embodiment of the invention, in order to acquire the pulse feeling position information, the machine learning software can process the original image by utilizing the learning capability of the deep convolutional neural network. Because the deep convolutional neural network has the corresponding capabilities of identification and classification and the like after being trained, the device can also comprise a machine learning module in order to determine pulse feeling position information conveniently; wherein, the machine learning module includes:
the training data acquisition unit is used for acquiring an image training sample with a label and depth information; the labels comprise ulnar position information, guan mai position information and cun mai position information;
and the training module is used for training the initial deep convolution neural network by using the image training sample to obtain the deep convolution neural network.
That is, before pulse taking, machine learning software for learning pulse taking position information of an original image may be trained in advance. Specifically, a large number of image training samples can be collected, and then the positions of the pulses are accurately calibrated by using traditional Chinese medicine with rich experience, so that the image training samples with labels are obtained. And then, training the deep convolutional neural network by using the labeled image training samples to obtain the deep convolutional neural network meeting the requirements, namely obtaining the machine learning software. Wherein, the meeting requirement means that the loss value calculated based on the loss function meets less than the preset threshold or the loss value is not changed any more. After the machine learning software is trained, when an original image is obtained, the original image can be input into a deep convolution neural network in the trained machine learning software for information, and then pulse feeling position information is obtained. Of course, in order to further determine whether the pulse-taking position information is wrong or not, the pulse-taking position information may be displayed on the display screen, for example, the pulse-taking position is marked on the original image based on the pulse-taking position information.
Preferably, to further improve the accuracy of the pulse-taking location information, the original image may be preprocessed before the learning by the machine learning software. That is, the pulse feeling position information obtaining module may specifically include:
the image preprocessing unit is used for preprocessing the original image by utilizing image processing software to obtain the wrist width, the origin of pulse feeling and a target image after background removal; wherein, the origin of pulse feeling is the intersection point of the connecting lines of the wrist and the ulnar, guan and cun pulse; inputting the target image, the origin of pulse feeling and the width of the wrist into machine learning software to learn the pulse feeling position, and obtaining pulse feeling position information.
Referring to fig. 2, fig. 2 is a schematic diagram illustrating the determination of the position below the elbow according to an embodiment of the present invention. The original image can be processed by image processing software such as OpenCV to obtain the width of the wrist, i.e. an illustration line segment | AB |; the origin of the pulse feeling, i.e. the position of the illustrated O point, the connecting line of the O point and the pulse of the inch size forms the x axis, and the target image without background. And inputting the wrist width, the pulse feeling origin and the target image without the background into a trained deep convolution neural network for learning to obtain pulse feeling position information. In the embodiment of the invention, the machine learning software with the deep convolutional neural network can also be arranged in the server so as to improve the acquisition speed of the pulse feeling position information by utilizing the strong computing power of the server. It should be noted that, because the pulse conditions corresponding to the cun pulse, the chi pulse and the guan pulse respectively need to be palpated in the traditional Chinese medical science pulse diagnosis, in the embodiment of the present invention, the pulse feeling position information may specifically include the position of the cun pulse, the position of the chi pulse and the position of the guan pulse. For example, according to the proportion of | OA | in | AB | and the point O on the extension of the straight line of the inch-off scale, the point O can be marked by OpenCV from this proportion. And then inputting the processed target image into the trained deep convolutional neural network, wherein the deep convolutional neural network outputs the detected image, and the image of the part below the elbow, which is marked with cun pulse, guan pulse and chi pulse, can be obtained. Of course, the output result may also be the position information of cun pulse, guan pulse and chi pulse.
And the pulse condition information acquisition module is used for outputting a motor control signal corresponding to the pulse feeling position information to the three-axis motor and acquiring pulse condition information by using a pulse diagnosis probe connected with an output shaft of the three-axis motor.
After pulse feeling position information is obtained, a motor control signal corresponding to the pulse feeling position information can be output to the three-axis motor, and then pulse condition information is acquired by using the pulse diagnosis probe. That is, the three-axis motor is driven to make the pulse feeling probe perform pulse feeling (i.e., to acquire pulse conditions). Specifically, the position of the three-axis motor can be controlled by using the PID according to the position information, namely the three motors of the x axis, the y axis and the z axis are controlled to quickly and accurately move the pulse-taking probe to the correct pulse-taking position so as to obtain the pulse pressure signal.
Specifically, because the pulse feeling in traditional Chinese medicine needs to diagnose the pulse conditions at different positions, in the embodiment of the present invention, the pulse condition information acquisition module may specifically include:
the motor control unit is used for sequentially and respectively outputting motor control signals corresponding to the pulse size position information, the pulse guan position information and the pulse cun position information to the three-axis motor when the pulse feeling position information comprises the pulse size position information, the pulse guan position information and the pulse cun position information;
the pulse condition acquisition unit is used for sequentially acquiring ulnar pulse condition information, guan pulse condition information and cun pulse condition information by using the pulse diagnosis probe.
That is, after the pulse feeling position information is determined, the three-axis motor is moved in sequence, so that the pulse diagnosis probe acquires the ulnar pulse condition information, the guan pulse condition information and the cun pulse condition information in sequence. In the embodiment of the invention, position control can be carried out by adopting PID (proportion integration differentiation), so that the three motors of the x axis, the y axis and the z axis can quickly and accurately move the pulse diagnosis probe to the previous position for pulse condition acquisition.
Preferably, in consideration of the fact that the traditional Chinese medicine diagnosis needs to determine the depth of the pulse condition, the pulse condition information acquisition module can be specifically used for outputting motor control signals with different pressure levels and corresponding to pulse feeling position information to a motor in the vertical direction in a three-axis motor, and acquiring pulse condition information corresponding to different pressure levels by using a pulse diagnosis probe. Wherein, the vertical direction is perpendicular to the surface of the skin on the inner side of the wrist. That is, the depth of the pulse condition is detected by the magnitude of the applied pressure (e.g. setting three pressure levels of light, medium and heavy), while the intensity of the pulse condition can be known by the measured pressure change in the blood vessel, and the speed of the pulse condition can be determined by the acquired pressure waveform and period. Therefore, the device can obtain the pulse condition information of the pulse speed, the pulse intensity and the pulse depth in the traditional Chinese medicine. And when the speed of the pulse condition is collected, the respiration of a doctor does not need to be referred as a counter, so that the accuracy rate of the pulse condition information can be improved.
Preferably, in order to enable the pulse-taking probe to be capable of fitting the bending amplitude of the wrist of the human body and better collect pulse condition information, the output shaft of the three-shaft motor is connected with the pulse-taking probe through a universal joint. Specifically, reference may be made to fig. 3 for a connection relationship between an output shaft of the three-axis motor and the pulse-taking probe through a universal joint, and fig. 3 is a schematic diagram of a universal joint connection in an embodiment of the present invention. Wherein the pulse-taking probe is positioned below a flat plate, namely a pressure sensor array. Wherein D1 is a connecting part connected with the output end of the motor, D2 is a spherical part, D3 is a sensor array, D2 is rigidly connected with D1, and D1 is connected into a row of flexible wires and D3. Since D2 is a spherical member, the sensor array is allowed to rotate in all directions. Therefore, when the gimbal is operated in conjunction with the motor and probe, D1 is connected to the motor control section and the sensor array is connected to the pulse probe (not shown). Due to the function of the spherical D2, when the direction of the D1 connected with the motor control part is changed, the direction of the probe connected with the sensor array can be changed in a sliding manner and is always attached to the skin, and the effect of accurately measuring the pressure waveform is achieved. The specific size, specific connection details, component details, and the like of the universal joint can be referred to common universal joints and methods for using the universal joints, and are not described in detail herein.
And the pulse condition information display module is used for displaying the pulse condition information on the display screen.
Specifically, the pressure waveform of the pulse condition can be displayed on the display screen.
Preferably, in order to facilitate a doctor to quickly check for accurate diagnosis, the pulse condition obtaining apparatus provided in the embodiment of the present invention may further include: and the pulse condition result acquisition module is used for analyzing the pulse condition information to obtain a pulse diagnosis result. Specifically, the pulse condition information may be obtained by using a deep convolutional network, with reference to the above-mentioned obtaining manner of obtaining the pulse feeling location information, to obtain the pulse diagnosis result. Namely, the samples required by the training of the deep convolutional network for analyzing and processing the pulse condition information to obtain the pulse diagnosis result are the pulse condition information with the pulse diagnosis result. In the training process, the predicted application process can refer to the above-described pulse-feeling location information obtaining manner, and details are not repeated here.
The device provided by the embodiment of the invention comprises: the stereoscopic vision acquisition module is used for acquiring an original image of the wrist of the human body and with depth information by using a stereoscopic camera; the pulse feeling position information acquisition module is used for inputting the original image into the trained machine learning software to acquire pulse feeling position information; the pulse condition information acquisition module is used for outputting a motor control signal corresponding to pulse feeling position information to the three-axis motor and acquiring pulse condition information by using a pulse diagnosis probe connected with an output shaft of the three-axis motor; and the pulse condition information display module is used for displaying the pulse condition information on the display screen.
In the device, trained machine learning software is used for carrying out vein-cutting position information on the original image acquired by the stereo camera to obtain the vein-cutting position information. Then, based on the pulse feeling position information, the three-axis motor is controlled to move, so that the pulse diagnosis probe connected with the output shaft of the three-axis motor collects pulse condition information, and then the pulse condition information is displayed on a display screen. Compared with the existing pulse diagnosis device, the pulse diagnosis device has the advantages that the pulse diagnosis probe does not need to be manually moved to the pulse diagnosis position, and the pulse diagnosis position can be determined. In the whole working process of the device for acquiring the pulse condition information, all pulse conditions of cun, guan and chi of a patient can be accurately acquired without intervention of a doctor. Namely, the device can realize that the machine can automatically and accurately acquire the pulse condition information, and is beneficial to the accurate diagnosis of doctors.
Example two:
corresponding to the above method embodiments, the embodiments of the present invention further provide a pulse condition acquiring system, and the pulse condition acquiring system described below and the pulse condition acquiring apparatus described above may be referred to with each other.
Referring to fig. 4, the system includes:
the pulse diagnosis instrument comprises a computer 100 provided with visual positioning software 101 and a display screen 102, a stereo camera 200 in communication connection with the computer, a three-axis motor 300 in communication connection with the computer, and a pulse diagnosis probe 400 connected to an output shaft 301 of the three-axis motor;
the three-dimensional camera is used for acquiring an original image of the wrist of the human body and with depth information and sending the original image to the computer;
the computer is used for acquiring pulse feeling position information by using the visual positioning software, outputting a motor control signal corresponding to the pulse feeling position information to the three-axis motor, and displaying the pulse feeling information on the display screen after receiving the pulse feeling information; the visual positioning software inputs the original image into a trained machine learning module to carry out pulse feeling position learning so as to obtain pulse feeling position information;
the three-axis motor is used for driving the pulse-taking probe to move under the control of a motor control signal;
the pulse diagnosis probe is used for acquiring pulse condition information and sending the pulse condition information to the computer.
Wherein, the triaxial motor is connected with the pulse-taking probe through a universal joint.
Wherein, the triaxial motor is a miniature stepping motor or a direct current brushless motor.
The machine learning module may be referred to herein as machine learning software and its training process.
The system provided by the embodiment of the invention comprises the following components: the pulse diagnosis device comprises a computer, a stereo camera, a three-axis motor and a pulse diagnosis probe, wherein the computer is provided with visual positioning software and a display screen; the three-dimensional camera is used for acquiring an original image of the wrist of the human body and with depth information and sending the original image to the computer; the computer is used for acquiring pulse feeling position information by using the visual positioning software, outputting a motor control signal corresponding to the pulse feeling position information to the three-axis motor, and displaying the pulse feeling information on the display screen after receiving the pulse feeling information; the visual positioning software inputs the original image into a trained machine learning module to carry out pulse feeling position learning so as to obtain pulse feeling position information; the three-axis motor is used for driving the pulse-taking probe to move under the control of a motor control signal; the pulse diagnosis probe is used for acquiring pulse condition information and sending the pulse condition information to the computer.
In the system, trained machine learning software is used for carrying out vein-cutting position information on the original image acquired by the stereo camera to obtain vein-cutting position information. Then, based on the pulse feeling position information, the three-axis motor is controlled to move, so that the pulse diagnosis probe connected with the output shaft of the three-axis motor collects pulse condition information, and then the pulse condition information is displayed on a display screen. Compared with the existing pulse diagnosis device, the pulse diagnosis device has the advantages that the pulse diagnosis probe does not need to be manually moved to the pulse diagnosis position, and the pulse diagnosis position can be determined. In the whole working process of the device for acquiring the pulse condition information, all pulse conditions of cun, guan and chi of a patient can be accurately acquired without intervention of a doctor. Namely, the device can realize that the machine can automatically and accurately acquire the pulse condition information, and is beneficial to the accurate diagnosis of doctors.
In order to facilitate better understanding of the pulse acquisition system provided by the embodiments of the present invention, the pulse acquisition system will be described with reference to the prior art and specific application examples.
At present, instruments for detecting the cunguan position on the market all adopt a complex and signal processing method to determine the cunguan position, the emphasis is on measuring the pulse frequency, and the visual pulse diagnosis process cannot be realized. For example, the invention patent with the application number of CN201410451790.5, namely a traditional Chinese medicine pulse diagnosis instrument and a method for positioning cun, guan and chi pulse points thereof, comprises a flexible body, a flexible pressure sensor array, an air bag, a signal processing system and a connecting structure. The device determines the position of the highest point of the radial styloid process according to the pressure signals collected by the flexible pressure sensor array, and finally determines the positions of cun, guan and ulnar pulse points according to the position of the highest point of the radial styloid process and the pressure signals collected by the flexible pressure sensor array. However, the pressure signal requires a strict noise cancellation process, so this method is greatly limited by the noise.
The invention patent with application number of CN201510467648.4 discloses an adjustable longitudinal pulse condition detection driver, which comprises a support rod, two sliding buckles, two locking caps and two pressure springs. The work of the invention depends on the pressure spring, and the requirement on the stability of the pressure spring is higher. The smaller or larger elastic coefficient of the pressure spring can influence the positioning of the inch-off size position.
It can be seen that in the prior art such as the above invention, neither the ulnar position of the wrist can be automatically detected nor the sensor probe can be automatically moved to the ulnar position to acquire the pulse condition.
The pulse condition acquisition system provided by the embodiment of the invention can comprise a stereo camera, a server, a computer (namely a computer), a pressure sensor probe (namely a pulse diagnosis probe) and a plurality of motors controlled by the computer. When the device is used, the three-dimensional camera shoots 3D (three-dimensional) information of the wrist of a human body, the shot image is transmitted back to a computer, and then the image is processed by OpenCV. The method comprises the steps of removing a pure color background of an image through OpenCV, and then determining the width of a wrist, so that an intersection point between a straight line formed by the customs size and a wrist line segment can be obtained according to an output result of a trained neural network, and a straight line where three points of the customs size are located is obtained. And then inputting the image into a trained deep convolutional neural network for learning, wherein the deep convolutional neural network outputs specific positions of the inch, the close and the ruler points. The resulting specific locations may be uploaded to a server in order to enrich the data set. The original image can be acquired by shooting through a stereoscopic vision method. The stereoscopic vision method can shoot two-dimensional image information, and can also calculate the depth information of an object through the difference of images shot by two cameras. Of course, the depth information may also be collected by a depth camera, such as a camera using infrared structured light.
The machine learning can establish a three-dimensional image of the wrist according to the image information and the depth information and determine the position of the inch and the scale. Machine learning is divided into deep learning and simpler regression algorithms.
Deep learning trains a deep neural network according to an image with a manually measured inch-off-scale position, which is called a training set with a label. Usually, more than thousands of cases of labeled training data are needed for deep learning to complete training, and higher accuracy is achieved. The trained neural network can correctly judge the position of the inch-pass ruler according to the image input.
In the system, a machine learning method which is simpler and has low training data requirement, such as a random sampling consistency method, can be used for judging the position of the dimension. When the random sampling consistency method is used, the wrist part of a person is taken as a geometric model, and the geometric model is described by the position information of the wrist, the width, the height, the intercepted wrist length, the height of a detected person, the position of cun pulse, the position of guan pulse, the position of chi pulse and other parameters of the model. The method can accurately and objectively describe the model by the parameters, can comprehensively consider the influence on the parameters caused by the height and the thinness of the human body, abstract the wrist part of the human body, reserve the key part of the problem, ignore the part irrelevant to the problem and simplify the analysis of the problem. These parameters are then adjusted using random sample consistency. The random sampling consistency estimates the parameters of the mathematical model in an iteration mode, but because the random sampling consistency is an uncertain algorithm, the probability of obtaining a reasonable result is related to the iteration times, when the random sampling consistency algorithm is used, the iteration times are adapted to the model as much as possible and are as much as possible, so that the wrist model is matched, and the inch, the close and the ruler position parameters are found out.
The system controls three motors of x, y and z axes to move the probe of the pulse diagnosis instrument to an accurate position according to the cun, guan and chi position information and the cun-guan (point C, D, E) inclined angle of the skin surface in fig. 2, controls the inclined angle of the pulse diagnosis probe by using the motor, applies three different pressures of light, medium and heavy to the pulse respectively, and acquires the cun-guan pulse conditions of the left hand and the right hand. A micro stepper motor or a dc brushless motor drive may be used. After the position information (e.g., x, y, z) of the inch-scale is determined, the computer-controlled motor rapidly moves the probe to the specified position using a common PID position control algorithm.
In the whole working process of the system for acquiring the pulse condition information, all pulse conditions of cun guan chi of the left hand and the right hand of a patient can be accurately acquired without intervention of a doctor. Meanwhile, the cun-guan-chi of the radial artery is positioned by using a computer vision and machine learning and deep learning method. Thereby realizing the scientization, the quantification, the objectification and the automation of the pulse diagnosis and assisting doctors to obtain accurate traditional Chinese medicine diagnosis.
Referring to fig. 5, fig. 5 is a schematic structural diagram of a pulse condition obtaining system according to an embodiment of the present invention.
In practical application, a detector such as a medical staff shoots a human body by using the stereo camera 200 to form an original image of the human body, and further, various data of the posture of the human body of the detector can be obtained based on the original image, so that height information of the cunguan pulse relative to the wrist of the human body can be obtained, a camera on a plane shoots a graph of a part below the elbow and then transmits the graph to the computer 100, and the computer processes the image by using the OpenCV to obtain the width | AB | of the wrist and the position of the O point. And then transmitted to the server 500 by the computer, and the trained deep convolutional neural network is input. And after the server finishes the calculation, the image processing result is transmitted back to the computer. The recognition result can be saved in the server so as to continuously enrich the database of the image, so that the next detection result is more accurate; meanwhile, the obtained position signal is transmitted to the motor 300 (the three-axis motor in the text) through the computer, the motor drives the pulse-taking probe 400 to automatically perform machine pulse taking, the system performs position control by using PID according to the position information, and the three motors of the x axis, the y axis and the z axis are controlled to quickly and accurately move the pulse-taking probe to the correct position. By simulating the situation that the degree of pressing pressure is changed during the pulse feeling of the traditional Chinese medicine, three different pressures of light, medium and heavy are respectively applied to the pulse, the pulse conditions of cun pulse, guan pulse and chi pulse obtained by pressing the left hand and the right hand under different pressures are acquired and displayed, so that the pulse condition data obtained by the pulse feeling of the machine can reach the level of pulse feeling of the traditional Chinese medicine.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative components and steps have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
Claims (9)
1. A pulse manifestation obtaining apparatus, comprising:
the stereoscopic vision acquisition module is used for acquiring an original image of the wrist of the human body and with depth information by using a stereoscopic camera;
the pulse feeling position information acquisition module is used for inputting the original image into machine learning software to acquire pulse feeling position information;
the pulse condition information acquisition module is used for outputting a motor control signal corresponding to the pulse feeling position information to the three-axis motor and acquiring pulse condition information by using a pulse diagnosis probe connected with an output shaft of the three-axis motor; the pulse condition information acquisition module is used for carrying out position control on the three-axis motor through PID;
the pulse condition information display module is used for displaying the pulse condition information on a display screen;
the pulse feeling position information acquisition module comprises:
the image preprocessing unit is used for preprocessing the original image by utilizing image processing software to obtain the wrist width, the origin of pulse feeling and a target image after background removal; wherein, the origin of pulse feeling is the intersection point of the connecting lines of the wrist and the ulnar, guan and cun pulses; inputting the target image, the pulse feeling origin and the wrist width into the machine learning software to obtain the pulse feeling position information.
2. The pulse acquiring device according to claim 1, wherein the pulse information acquiring module comprises:
the motor control unit is used for sequentially and respectively outputting motor control signals corresponding to the ulnar position information, the guan-mai position information and the cun-mai position information to the three-axis motor when the pulse feeling position information comprises the ulnar position information, the guan-mai position information and the cun-mai position information;
and the pulse condition acquisition unit is used for sequentially acquiring ulnar pulse condition information, guan pulse condition information and cun pulse condition information by using the pulse diagnosis probe.
3. The pulse acquiring device according to claim 1, wherein the pulse information acquiring module is specifically configured to output the motor control signals corresponding to the pulse feeling position information and with different pressure levels to a vertical motor in the three-axis motor, and acquire the pulse information corresponding to the different pressure levels by using the pulse probe.
4. The pulse condition acquiring device according to claim 1, wherein the output shaft of the three-shaft motor is connected to the pulse probe via a universal joint.
5. The pulse condition obtaining apparatus according to claim 1, further comprising: a machine learning module;
wherein the machine learning module comprises:
the training data acquisition unit is used for acquiring an image training sample with a label and depth information; the labels comprise ulnar position information, guan mai position information and cun mai position information;
and the training module is used for training initial machine learning software by using the image training sample to obtain the machine learning software.
6. The pulse manifestation acquisition device of any one of claims 1 to 5, further comprising:
and the pulse condition result acquisition module is used for analyzing the pulse condition information to obtain a pulse diagnosis result.
7. A pulse acquisition system for realizing the pulse acquisition apparatus according to claim 1, comprising:
the pulse diagnosis device comprises a computer, a stereo camera, a three-axis motor and a pulse diagnosis probe, wherein the computer is provided with visual positioning software and a display screen, the stereo camera is in communication connection with the computer, the three-axis motor is in communication connection with the computer, and the pulse diagnosis probe is connected to an output shaft of the three-axis motor;
the three-dimensional camera is used for acquiring an original image of the wrist of a human body and with depth information and sending the original image to the computer;
the computer is used for acquiring pulse feeling position information by using visual positioning software, outputting a motor control signal corresponding to the pulse feeling position information to the three-axis motor, and displaying the pulse feeling information on the display screen after receiving the pulse feeling information; the vision positioning software inputs the original image into a trained machine learning module to carry out pulse feeling position learning so as to obtain pulse feeling position information;
the three-axis motor is used for driving the pulse-taking probe to move under the control of the motor control signal;
the pulse diagnosis probe is used for acquiring the pulse condition information and sending the pulse condition information to the computer.
8. The pulse acquisition system of claim 7, wherein the tri-axial motor is gimbaled to the pulse probe.
9. The pulse acquisition system according to claim 7, wherein the three-axis motor is a micro stepper motor or a DC brushless motor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910522341.8A CN110200601B (en) | 2019-06-17 | 2019-06-17 | Pulse condition acquisition device and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910522341.8A CN110200601B (en) | 2019-06-17 | 2019-06-17 | Pulse condition acquisition device and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110200601A CN110200601A (en) | 2019-09-06 |
| CN110200601B true CN110200601B (en) | 2022-04-19 |
Family
ID=67793050
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910522341.8A Active CN110200601B (en) | 2019-06-17 | 2019-06-17 | Pulse condition acquisition device and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110200601B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111643060A (en) * | 2020-05-22 | 2020-09-11 | 深圳市物联升科技有限公司 | Doctor-patient contact-free traditional Chinese medicine auxiliary diagnosis system based on artificial intelligence |
| CN111700599A (en) * | 2020-05-28 | 2020-09-25 | 中北大学 | Finger method pulse feeling device and system simulating traditional Chinese medicine |
| TWI810495B (en) * | 2020-10-29 | 2023-08-01 | 國立東華大學 | Pulse-measuring equipment and pulse-measuring method for pulse diagnosis |
| CN114431827B (en) * | 2020-11-03 | 2024-03-08 | 东华大学 | Bracelet type pulse diagnosis instrument |
| CN112826458B (en) * | 2021-01-08 | 2022-11-01 | 广州中医药大学(广州中医药研究院) | Pulse diagnosis system and pulse diagnosis method |
| CN113229788B (en) * | 2021-03-26 | 2022-07-19 | 广东粤港澳大湾区黄埔材料研究院 | Pulse wave denoising method and device based on film pressure sensor |
| CN115736907A (en) * | 2021-09-04 | 2023-03-07 | 天津大学 | Wearable microwave nondestructive testing human blood sugar based on cascade convolution neural network |
| CN113892919B (en) * | 2021-12-09 | 2022-04-22 | 季华实验室 | Pulse feeling data acquisition method and device, electronic equipment and system |
| CN113974581B (en) * | 2021-12-15 | 2023-10-20 | 曾国伟 | Matching method and system for dynamic incision characteristics of pulse diagnosis |
| CN115381411A (en) * | 2022-09-19 | 2022-11-25 | 慧医谷中医药科技(天津)股份有限公司 | Pulse condition analysis information processing method and system |
| CN116236165B (en) * | 2023-03-22 | 2023-10-31 | 中国中医科学院中医基础理论研究所 | Pulse diagnosis data generation and display method and corresponding device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1360440A (en) * | 2002-01-31 | 2002-07-24 | 北京理工大学 | Miniaturized real-time stereoscopic visual display |
| CN107122776A (en) * | 2017-04-14 | 2017-09-01 | 重庆邮电大学 | A kind of road traffic sign detection and recognition methods based on convolutional neural networks |
| CN108542358A (en) * | 2018-05-14 | 2018-09-18 | 广州中康科技发展有限公司 | The device and method of pulse condition 7 degree of freedom automatic acquisition of scientific information based on photoacoustic imaging |
| CN108784664A (en) * | 2018-06-28 | 2018-11-13 | 上海掌门科技有限公司 | Pulse diagnosing device based on pressure sensor and image capture device |
| CN108814564A (en) * | 2018-05-30 | 2018-11-16 | 广东工业大学 | A kind of intelligence pulse diagnosing device |
| CN108961678A (en) * | 2018-04-26 | 2018-12-07 | 华慧视科技（天津）有限公司 | One kind being based on Face datection Study in Driver Fatigue State Surveillance System and its detection method |
-
2019
- 2019-06-17 CN CN201910522341.8A patent/CN110200601B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1360440A (en) * | 2002-01-31 | 2002-07-24 | 北京理工大学 | Miniaturized real-time stereoscopic visual display |
| CN107122776A (en) * | 2017-04-14 | 2017-09-01 | 重庆邮电大学 | A kind of road traffic sign detection and recognition methods based on convolutional neural networks |
| CN108961678A (en) * | 2018-04-26 | 2018-12-07 | 华慧视科技（天津）有限公司 | One kind being based on Face datection Study in Driver Fatigue State Surveillance System and its detection method |
| CN108542358A (en) * | 2018-05-14 | 2018-09-18 | 广州中康科技发展有限公司 | The device and method of pulse condition 7 degree of freedom automatic acquisition of scientific information based on photoacoustic imaging |
| CN108814564A (en) * | 2018-05-30 | 2018-11-16 | 广东工业大学 | A kind of intelligence pulse diagnosing device |
| CN108784664A (en) * | 2018-06-28 | 2018-11-13 | 上海掌门科技有限公司 | Pulse diagnosing device based on pressure sensor and image capture device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110200601A (en) | 2019-09-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110200601B (en) | Pulse condition acquisition device and system | |
| US7478009B2 (en) | Apparatus and method for evaluating a hypertonic condition | |
| Oubre et al. | Estimating upper-limb impairment level in stroke survivors using wearable inertial sensors and a minimally-burdensome motor task | |
| CN105832343B (en) | Multidimensional vision hand function rehabilitation quantitative evaluation system and evaluation method | |
| EP3376414B1 (en) | Joint movement detection system and method, and dynamic assessment method and system for knee joint | |
| JP2021520939A (en) | Adaptive ultrasonic scanning | |
| CN102247128A (en) | Human body pulse information collecting device and human body health status monitoring device | |
| EP2667783A2 (en) | System, method and device for automatic and autonomous determination of hemodynamic and cardiac parameters using ultrasound | |
| WO2022037274A1 (en) | Muscle training method and system for providing visual feedback by using ultrasonic imaging | |
| CN112826458B (en) | Pulse diagnosis system and pulse diagnosis method | |
| CN104146684A (en) | Blinder type dizziness detector | |
| Bumacod et al. | Image-processing-based digital goniometer using OpenCV | |
| Mykhalyk et al. | Modern hardware and software solution for identification of abnormal neurological movements of patients with essential tremor | |
| Zhu et al. | A protocol for digitalized collection of Traditional Chinese Medicine (TCM) pulse information using bionic pulse diagnosis equipment | |
| CN113995383B (en) | Internet of things-based traditional Chinese medicine remote inspection informatization processing system and method | |
| Lu et al. | Comparing limb-volume measurement techniques: 3D models from an infrared depth sensor versus water displacement | |
| CN108804246A (en) | The usability evaluation method of upper limb rehabilitation robot | |
| Rovini et al. | Vision optical-based evaluation of SensHand accuracy for Parkinson’s disease motor assessment | |
| CN115605128A (en) | Pulse wave intelligent distinguishing method and system based on human body sequence parameter model | |
| Li et al. | A virtual scanning framework for robotic spinal sonography with automatic real-time recognition of standard views | |
| Jobbagy et al. | PAM: passive marker-based analyzer to test patients with neural diseases | |
| CN116236165B (en) | Pulse diagnosis data generation and display method and corresponding device | |
| CN113907729B (en) | Scientific and technological data acquisition method based on medical robot | |
| JP2019048211A (en) | Ultrasonic image diagnostic system, and control method of the same | |
| CN116012938B (en) | Construction method and system of CPR automatic feedback detection model based on AlphaPose algorithm |
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 |