JP2018158083A - Posture Distortion Evaluation Method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To propose a posture distortion evaluation method for inhibiting an elderly person from falling down, by presenting to a subject "a distortion age" converted from the magnitude of posture distortion using a variation between posture distortion at the time of linear walking and that at the time of standing still.SOLUTION: Respective predetermined parts of the head, left or right shoulder, and hip are measured using an evaluation system comprising motion capture means for measuring three-dimensional positions at the time of standing still and at the time of walking, a computer for calculating a distortion age, and a display device, storage device, or transmission means therefor. In the case that a shoulder position is located on a line connecting the left and right shoulders, y is defined as a distortion age, and xand xare defined as a distance between the head and the shoulder position at the time of standing still and a distance between the head and the hip at the time of walking, respectively, in a plan view, a posture distortion evaluation method includes: (1) measuring xat the time of standing still; (2) measuring x2 at the time of walking; (3) calculating y as a linear sum of a×x, b×x, and a constant c; and (4) transmitting the distortion age y to a display unit; where the coefficients a, b, c are calculated using a multiple regression analysis taking an actual age as an objective variable and taking measurement values of the xand xas explanatory variables.SELECTED DRAWING: Figure 4

Description

  The present invention measures the posture distortion of a subject, converts the magnitude of the posture distortion into a “distortion age” expressed in age, informs the subject of the “distortion age”, The present invention relates to a posture distortion evaluation method for suppressing falls in daily activities by utilizing the selection of clothes and footwear and exercises of a person.

  It is generally known that body distortion caused by aging is a cause of headache, visceral disease, and fall accident in addition to poor posture. In particular, it has been pointed out that the fall of elderly people causes fractures, brain diseases and the like, leading to bedriddenness, and is one of the major social problems that have been sought to be solved.

  When comparing the postures of young and elderly, it is well known that elderly people tend to be in a so-called leaning back posture. In addition, it is known that elderly people lean forward because the head position is downward and the line of sight is lowered, so there is a possibility that the visual attention range may be narrowed, and the risk of falls is increased. . Therefore, it is considered that quantifying the degree of the forward leaning posture is effective in estimating the risk of falls based on the elderly leaning forward posture.

Conventionally, the evaluation of body strain has been generally conducted in an upright and stationary state. However, according to the study by the inventor of the present invention, as a result of comparing the postures when standing upright and straight walking, there was no significant difference between the two in young people, There was a big difference.
From this, it can be considered that the characteristic of the elderly's body distortion lies in the amount of change in posture distortion during linear walking and standing upright.

  As a method for quantifying the coordinates of the body in motion, digitizing the important parts of the body using a motion capture device is performed. As the motion capture device used at this time, an optical type, an inertial sensor type, a mechanical type, a magnetic type, a video type, a composite sensor type, and the like are known.

In the optical system, multiple cameras are installed around the measurement space, images are taken using a reflective marker attached to the body as a tracker, and the movement of the subject is digitized by analyzing the coordinates of the tracker of multiple camera images. Is.
The inertial sensor type is to attach an inertial sensor consisting of a gyro sensor (angular velocity meter) and an accelerometer to the measurement point of the body and calculate the position and posture by calculating backward from the information measured as acceleration, which requires image analysis. There is no.
The mechanical type measures each joint angle using a sensor such as a potentiometer or an encoder that mechanically measures the rotation angle and displacement.
The magnetic type detects a magnetic field generated by a magnetic generator with a magnetic sensor attached to the body. The magnetic field in three axes is detected by each magnetic sensor to obtain the position and orientation from the magnetic field generator. Compared to the optical type, there is no fear of hiding the marker, and the position and orientation information can be obtained with a set of magnetic sensors.
In the video type, images taken simultaneously by a plurality of cameras are analyzed, the three-dimensional position of each pixel is determined and captured as a 3D image.
The compound sensor type is a motion capture that combines a plurality of sensors, for example, Kinect (registered trademark). This includes a color camera, an infrared laser light source that generates speckle patterns, an infrared camera, and an array microphone. By obtaining depth information in addition to a two-dimensional image, three-dimensional image information can be obtained.

  Patent Document 1 discloses a diagnostic system that captures the movement of a specific part of a dairy cow, for example, using motion capture means, and objectively diagnoses the state of the hooves of an imaged dairy cow based on the movement of the specific part A diagnostic method is disclosed. In this diagnostic method, for example, a reflector for a marker is provided at the tip of the femur of a cow, the movement of the marker is extracted from the video image of the marker, and the health of the cow is grasped by analyzing the movement of the marker. It is.

JP 2005-253435 A

  Conventionally, it has been common to evaluate the strain of an elderly person in a state of standing upright. However, it has become clear from the study of the inventors of the present invention that the characteristics of the distortion can be captured by the amount of change in the distortion of posture during straight walking and standing still. Therefore, we propose a measurement method that measures the posture during straight walking with sufficient accuracy in addition to the posture when standing upright, and quantitatively evaluates the difference in posture distortion.

The present invention is a posture distortion evaluation method for suppressing a subject from falling while walking, and a motion capture means for measuring a three-dimensional position of a plurality of portions of the subject at rest and during walking A posture of a person using an evaluation system comprising: a computer that calculates an evaluation age evaluated from the measured three-dimensional position; and a transmission means that transmits the evaluation age to a display device, a storage device, or another computer. An evaluation method for evaluating distortion,
The plurality of portions include predetermined portions of the head, the right or left shoulder, and the waist,
The position of the shoulder is a position determined on the line connecting the right shoulder and the left shoulder of the subject,
The evaluation age and y, in a plan view of the evaluation system, the distance between the position and the shoulder position of a predetermined portion of the stationary state of the head of the subject and x _1, the head during walking of the subject when the distance between the positions of the predetermined portion of the waist of a predetermined portion and x _2,
(1) step of measuring the distance x ₁ in the stationary state of the subject,
(2) step of measuring the distance x ₂ during walking of the subject,
(3) y the obtaining a linear sum of a substantially x ₂ term and a predetermined constant c terms of x ₁ term and the predetermined coefficient b of predetermined coefficients a step, and (4) display the evaluation age y And measuring to include a step of transmitting to a storage device or another computer.

The coefficients a, the coefficient b and the constant c is determined by performing a multiple regression analysis to the actual age of the plurality of human purposes variable the distance x ₁ and distance as explanatory variables the measured value of x ₂ belonging to a given population Value,
When the evaluation age y of the subject is obtained, the coefficient a, the coefficient b, and the constant c obtained by performing multiple regression analysis on the group to which the subject belongs are used.

  Further, the direction of the subject at the time of measurement is a direction of moving forward toward the motion capture means when stationary and walking.

The position of the subject when measuring the distance x ₁ is characterized by the same position as the position of the subject when measuring the distance x _2.

In addition, the evaluation system has a function of identifying a subject,
The computer manages data including measurement data and evaluation age for each identified subject,
By notifying the subject of the progress information of the evaluation age y, the subject is prevented from falling while walking.

The motion capture means is Kinect (registered trademark),
The positions of the plurality of parts measured by the motion capture means are the positions of joint points in the 3D skeleton of the Kinect (registered trademark).

  In general, it is known that the walking posture of an elderly person is leaning forward, and it is easy to fall down during walking. By notifying to the subject, it becomes possible to warn the subject of the deterioration of the posture. As a result, it is possible to suppress a fall accident during walking of an elderly person or the like.

  It is a figure which shows the example of the setup of the motion capture used by this invention, (a) is an example using the sensor which can detect the three-dimensional coordinate of a specific place without a marker, (b) is a plurality of marker coordinates. It is an example using the sensor detected using this imaging device.   It is a site | part of a subject paying attention by this invention, (a) is a front view, (b) is a side view. (A) is a view showing the vertical axis the measurement result of the distance x ₁ in the plan view of the position and the shoulder position of a predetermined portion of the stationary state of the head with respect to the subject's actual age (horizontal axis), (b) is the real age ordinate the measurement result of x ₂ distance plan view of the position of the predetermined portion of the waist of a predetermined portion of the head during walking of the subject with respect to (horizontal axis) FIG. (C) is a correlation between the x ₁ and x ₂ of the measurement result is a diagram showing together with the measurement results for young people. Shows 3 (a) and the distance x ₁ and the distance indicated in (b) as explanatory variables the measured value of x _2, linear function of the actual age was determined by performing a multiple regression analysis as objective variable (solid line) It is.

  Embodiments of the present invention will be described below in detail with reference to the drawings. In this description, devices having the same function or similar functions are denoted by the same reference numerals unless there is a special reason.

  FIG. 1 shows an example of motion capture setup used in the present invention. FIG. 1A shows an example using a sensor that can detect a three-dimensional coordinate of a specific place without a marker. In this example, the sensor 6 performs motion capture when the subject 5 is stationary and walking, and the computer 7 analyzes the output data of the sensor 6.

  This sensor 6 is of a composite sensor type, and has a two-dimensional distance measurement and depth distance measurement function based on an image by a normal video camera. Kinect (registered trademark, Microsoft) is known as such a sensor. The distance measurement of the depth of this sensor is to analyze the laser reflection pattern, but the method of scanning laser beam irradiation and the method of irradiating ultrasonic waves are also known, including those of other methods, Any material that can realize sufficient measurement accuracy with sufficient reproducibility can be used.

As shown in the front view of FIG. 2A and the side view of FIG. 2B, the measurement points of the sensor include predetermined portions of the head (measurement point 1), the right or left shoulder, and the waist. The position of the shoulder is a position (measurement point 2-3) determined on the line connecting the right shoulder and the left shoulder of the subject, and may be the position of the right shoulder or the left shoulder or an intermediate position thereof. . In the plan view of the evaluation system, the distance x ₁ between the position of the predetermined portion of the head and the shoulder position when the subject is stationary, and the position (measurement point) of the predetermined portion of the head when the subject is walking. Let x ₂ be the distance between 1) and the position of a predetermined part of the waist (measurement point 4). The use of such a distance is a feature of the present invention.

  The subject 5 takes an upright stationary posture at the start point in FIG. 1A and starts walking from the start point toward the sensor 6 along the auxiliary walking line 9. At this time, with respect to the above measurement, the measurement at the start point can be performed at the time of stationary, and the measurement at the time of walking can be performed at any point on the walking auxiliary line 9. However, in order to improve the reproducibility of the measurement result, the measurement at the time of walking is performed at a position where the subject takes a predetermined posture at the time of walking, and the subject is standing still at the measurement position at the time of walking. It is desirable to perform posture measurements. This is a measure for avoiding nonlinearity related to the distance of the sensor.

  FIG. 1B is an example using a sensor that detects the coordinates of a marker attached to the subject 5 using a plurality of imaging devices. This configuration uses the optical or video motion capture described above. In these methods, a plurality of video cameras can be used. FIG. 1B shows an example in which two video cameras 8a and 8b are used. The camera output data is sent to the computer 7 by wired or wireless connection. At this time, the method of fixing the markers 11, 12, 13, and 14 to the body of the subject has an advantage that the influence on the measurement value can be suppressed when the subject's clothes are disturbed by the walking motion. In the case of video motion capture, it is possible to set measurement points without using markers.

  In the computer 7, the evaluation age y is calculated using coefficients a, b, and c obtained by the method described below, and is displayed on the display screen or transmitted to a storage device or another computer. The evaluation age y indicates the age corresponding to the distortion of the body shape that accompanies aging in the standard body shape distribution, which is a so-called “distortion age”.

The above measurement procedure is summarized as follows. That means
(1) step of measuring the distance x ₁ in the stationary state of the subject,
(2) step of measuring the distance x ₂ during walking of the subject,
(3) y the obtaining a linear sum of a substantially x ₂ term and a predetermined constant c terms of x ₁ term and the predetermined coefficient b of predetermined coefficients a step, and (4) display the evaluation age y And measuring to a storage device or other computer. However, when measuring the distance x1 at the point of measuring the distance x ₂ is the (1), the order is reversed (2).

The coefficients a, b, and c are coefficients obtained by using a sufficient number of samples belonging to a predetermined group. In other words, a value obtained by performing a multiple regression analysis to the actual age of the plurality of human purposes variable the distance x ₁ and distance as explanatory variables the measured value of x ₂ belonging to a given population. This specific example is shown below.

Fig. 3 (a) shows the predetermined head at rest relative to the actual age (horizontal axis) for 20 men and women (10 men and 10 women) aged 70 to 83 (average 75.55 years). the measurement results of the distance x ₁ in the plan view of the position of the portion of the position and the shoulder on the vertical axis. Further in FIG. 3 (b), the measurement result of the distance of the plan view x ₂ between the position of the location and the portion of the waist of a predetermined portion of the actual age for (horizontal axis) during walking of the subject's head Shown on the vertical axis.
FIG. 3 (a), from (b), the x ₁ and x _2, the results correlation but is recognized, that demonstrate this correlation clearly examined the correlation to expand the subject to young Is shown in FIG. The young are 16 university students (average 20.8 years old) from 19 to 21 years old (12 men and 4 women).

  As described above, conventionally, the evaluation of body strain has been generally performed on an upright stationary state. That is, although it was evaluated by FIG. 3A, it is the distance on the plan view between the position of the predetermined part of the head and the position of the predetermined part of the waist during walking from FIGS. 3A and 3C. It can be seen that it is necessary to pay attention to x2.

The coefficients obtained by the multiple regression analysis from the data shown in FIGS. 3A and 3B were A = 1.068, b = 0.583, and c = 71.69. In FIG. 4, a solid example shows a plot example of Equation 1 having this coefficient in the three-dimensional coordinates of x ₁ , x ₂ , and y axis. In this plot example, the values of the above coefficients are used, and (x ₁ , x ₂ ) is a region of (−4, −4), (−4, 10), (10, −4), (10, 10). Is shown. y is a value obtained from Equation 1 above, and the plane indicated by Equation 1 is a plane including the plot.

The broken line (two-dot chain line) in FIG. 4 is synthesized from the least square line in FIGS. 3A and 3B using age as a parameter. Comparison of these solid and broken lines, the first and the least squares straight line was determined from the distribution of actual age and x _1, obtained by combining the second least squares straight line was determined from the distribution of chronological age and x2 and an evaluation age with a linear, as described above, evaluation age y (i.e. strain age the distance x ₁ and the distance to the objective variable actual age was determined by performing a multiple regression analysis as explanatory variables the measured value of x ₂ ) Is quite different.

  In the setup shown in FIGS. 1A and 1B, measurement of a subject approaching toward a sensor equipped with a video camera is performed. For this reason, it is easy for the computer 7 to identify the face of the subject. By identifying the subject in this way, it is easy to classify the measurement results and the like for each subject. It is also easy to show the time series of the measurement results to the subject from the measurement results arranged for each subject in this way. Furthermore, from this time series, by using software to judge abnormalities, it is possible to present a warning to the subject that an abnormal value has occurred, thereby preventing accidents such as falls and maintaining the health of the subject. It is clear that it is very useful in doing so.

1, 2, 3, 4 Measuring point 5 Subject 6 Sensor 7 Computer 8a, 8b Video camera 9, 10 Walking auxiliary line 11, 12, 13, 14 Marker

Claims (6)

Motion capture means for measuring a three-dimensional position of a plurality of parts of a subject at rest and when walking;
A computer for calculating an evaluation age evaluated from the measured three-dimensional position;
An evaluation method for evaluating distortion of a person's posture using an evaluation system comprising: a transmission means for transmitting the evaluation age to a display device, a storage device or another computer,
The plurality of portions include predetermined portions of the head, the right or left shoulder, and the waist,
The position of the shoulder is a position determined on the line connecting the right shoulder and the left shoulder of the subject,
The evaluation age and y, in a plan view of the evaluation system, the distance between the position and the shoulder position of a predetermined portion of the stationary state of the head of the subject and x _1, the head during walking of the subject when the distance between the positions of the predetermined portion of the waist of a predetermined portion and x _2,
(1) step of measuring the distance x ₁ in the stationary state of the subject,
(2) step of measuring the distance x ₂ during walking of the subject,
(3) y the obtaining a linear sum of a substantially x ₂ term and a predetermined constant c terms of x ₁ term and the predetermined coefficient b of predetermined coefficients a step, and (4) display the evaluation age y A posture distortion evaluation method, comprising: measuring a step including transmitting to a storage device or another computer to suppress the subject from falling while walking. The coefficients a, the coefficient b and the constant c is determined by performing a multiple regression analysis to the actual age of the plurality of human purposes variable the distance x ₁ and distance as explanatory variables the measured value of x ₂ belonging to a given population Value,
2. When determining the evaluation age y of the subject, the coefficient a, the coefficient b, and the constant c obtained by performing multiple regression analysis in a group to which the subject belongs are used. The posture distortion evaluation method described in 1.   The posture distortion evaluation method according to claim 1 or 2, wherein the direction of the subject at the time of measurement is a direction of moving forward toward the motion capture means when stationary and walking. Position of the subject when measuring the distance x ₁ is any one of claims 1 to 3, characterized in that the same position as the position of the subject when measuring the distance x ₂ The posture distortion evaluation method described in 1. The evaluation system has a function of identifying a subject,
The computer manages data including measurement data and evaluation age for each identified subject,
The posture distortion evaluation according to any one of claims 1 to 4, wherein the subject is prevented from falling while walking by notifying the subject of progress information of the evaluation age y. Method. The motion capture means is Kinect (registered trademark),
The posture according to any one of claims 1 to 5, wherein the positions of the plurality of parts measured by the motion capture means are positions of joint points in the 3D skeleton of the Kinect (registered trademark). Strain evaluation method.

Images