一种基于Microsoft Kinect的脑卒中患者上肢运动功能评定方法

Microsoft Kinect-based method for evaluating upper limb motor function of stroke patient

文档序号：9915 发布日期：2021-09-17 浏览：51次中文

1. A method for evaluating the motor function of the upper limb of a cerebral apoplexy patient based on Microsoft Kinect 2.0 is characterized by comprising the following steps:

1) initially defined: establishing a human body space coordinate system, calling joint point parameter definitions related to evaluation actions, initializing various parameter values, and setting a condition algorithm of a basic posture;

2) and (3) executing actions: the hemiplegic patient executes evaluation action under the guidance of characters, images and sounds in a screen;

3) data acquisition: acquiring three-dimensional coordinate information of key points of a human body of a patient by using a Kinect camera arranged in a treatment room;

4) and (3) data analysis: decomposing the evaluation action through an action algorithm, calculating and analyzing the obtained joint point coordinates according to a flow, and judging the degree of sufficiency of the action execution of the patient;

5) and (4) counting results: and (4) counting the completion conditions of all evaluation actions of the patient and determining the recovery stage of the upper limb movement function of the patient.

2. The method for assessing the motor function of the upper limbs of the cerebral apoplexy patient based on Microsoft Kinect 2.0 as claimed in claim 1, wherein the step 1) is implemented by establishing a human body space coordinate system: the HEAD, NECK, SPINE SHOULDER, SPINE MID, SPINE BASE are assumed to be on a straight line; taking the SPINE BASE skeleton coordinate point as (0, 0), the left-right direction as (X), the up-down direction as (Y), and the front-back direction as (Z); positive values to the right, up and forward, and negative values to the left, down and backward.

3. The method for assessing the motor function of the upper limbs of the stroke patient based on the Microsoft Kinect 2.0 as claimed in claim 1, wherein the step 1) assesses the joint point parameter definition related to the motion, and directly calls the self-contained parameter definition of the Microsoft Kinect 2.0 system.

4. The method for assessing the motor function of the upper limbs of the stroke patient based on the Microsoft Kinect 2.0 as claimed in claim 1, wherein the step 1) initializes each parameter, specifically, determines the values of the arm length and the trunk length of the upper limbs of the human body:

length of left arm: l _ AMLT = J _ SRLT _ Y-J _ HDLT _ Y

Length of right arm: l _ AMRT = J _ SRRT _ Y-J _ HDRT _ Y

Left palm to elbow length: l _ AMLT _1 = J _ EWLT _ Y-J _ HDLT _ Y

Right hand palm to elbow joint length: l _ AMRT _1 = J _ EWRT _ Y-J _ HDRT _ Y

Left elbow to shoulder length: l _ AMLT _2 = J _ SRLT _ Y-J _ EWLT _ Y

Right elbow to shoulder length: l _ AMRT _2 = J _ SRRT _ Y-J _ EWRT _ Y.

5. The method for assessing the motor function of the upper limbs of the stroke patient based on Microsoft Kinect 2.0 as claimed in claim 1, wherein the step 1) is a conditional algorithm for setting the basic posture: setting the tolerance to delt = 5 cm;

when the human body is in the end sitting position, the conditions met by all actions are as follows:

the left hand droops naturally without bending: i J _ SRLT _ X-J _ HDLT _ X < delt & | J _ SRLT _ Y-J _ HDLT _ Y-L _ AMLT | < delt

The right hand droops naturally without bending: i J _ SRRT _ X-J _ HDRT _ X | < delt & | J _ SRRT _ Y-J _ HDRT _ Y-L _ AMRT | < delt

The straight satisfying conditions of the left arm and the right arm are as follows:

|（（S_SRLT_X - S_HDLT_X）2 +（S_SRLT_Y - S_HDLT_Y）2 + S_SRLT_Z - S_HDLT_Z）2）^0.5 - L_AMLT| < delt

|（（S_SRRT_X - S_HDRT_X）2 +（S_SRRT_Y - S_HDRT_Y）2 + S_SRRT_Z - S_HDRT_Z）2）^0.5 - L_AMRT| < delt。

6. the method for assessing the motor function of the upper limbs of the stroke patient based on the Microsoft Kinect 2.0 as claimed in claim 1, wherein the step 2) of assessing the actions improves the actions in the brunstrom upper limb motor function assessment table, and specifically comprises the following steps:

(1) the affected hand is placed at the waist of the opposite side, the affected hand is lifted from the waist of the opposite side from the front of the body, and the affected hand tries to touch the ears of the same side;

(2) the affected hand is placed at the waist position on the same side, and the affected hand is extended to the knee on the opposite side from the front of the body;

(3) the affected hand is placed in the middle of the thigh on the same side, extended to the back and tried to touch the back of the waistband;

(4) the affected upper limb is placed on the side of the body with the affected upper limb hanging down, and the elbow is straightened to lift the affected upper limb forward to the same height as the shoulder;

(5) the elbow of the affected side is close to the tight side, the elbow joint is flexed by 90 degrees, the forearm is in a neutral position (the thumb is upward), and the forearm is rotated forward (namely the forearm is rotated to lead the palm center to be downward);

(6) on the basis of the action (5), the forearm is rotated backwards (namely the forearm is rotated to enable the palm center to face upwards);

(7) the affected upper limb is placed on the side of the body with the affected upper limb hanging down, the elbow is straightened, and the affected upper limb is extended horizontally to the side and is as high as the shoulder;

(8) the affected upper limb is vertically placed on the side of the body, the elbow is straightened, the affected upper limb is lifted forwards, and the shoulder joint is bent to 180 degrees as much as possible;

(9) straightening the elbow of the affected side, and raising the shoulder joint to 30-90 deg. in the middle position of the forearm (with the thumb facing upwards), and pronating the forearm (i.e. rotating the forearm to face the palm downwards);

(10) on the basis of the action (9), the forearm is rotated backwards (namely the forearm is rotated to enable the palm center to face upwards);

(11) the elbow joint is flexed, the upper arm is at the front side of the body, and the finger tip is placed on the shoulder at the same side; performing rapid lifting action by touching shoulder with fingertip, measuring time required for repeating 10 times, measuring healthy side, and comparing the measurement results of the affected side and healthy side (determining that the time required for the affected side is 1.5 times or less of that of the healthy side is sufficient);

(12) the affected hand is placed in the middle of the thigh on the same side, the chin is touched by the fingers with the fastest speed, and then the affected hand is placed back on the thigh; measuring the time required for repeating 10 times, measuring the healthy side, and comparing the measurement results of the affected side and the healthy side (determining that the time required for the affected side is 1.5 times or less of the healthy side is sufficient);

actions (1) to (12) are all performed in the sitting position.

7. The method for assessing the motor function of the upper limbs of the stroke patient based on the Microsoft Kinect 2.0 in the claim 1, wherein the step 4) decomposes the assessment action through an action algorithm, and the specific process is as follows: based on human body joint points captured by the Kinect, a motion algorithm capable of decomposing complex evaluation motions is designed by utilizing X, Y, Z coordinate reference relations among the joint points of all parts of the body; and substituting the three-dimensional coordinates of the relevant joint points in the action process of the patient captured by the Kinect into an algorithm, judging how many decomposition actions are completed by the patient, and further determining the execution sufficiency of the evaluation action.

8. The method for assessing the motor function of the upper limbs of the stroke patient based on Microsoft Kinect 2.0 as claimed in claim 1, wherein the specific process of the step 5) of counting the completion of each assessment action of the patient is as follows: judging that the patient can finish one of the actions (1) and (2) as Brunnstrom III, otherwise, outputting evaluation results from Brunnstrom I to Brunnstrom II; on the basis of judging as stage III, judging that the patient can finish one of (3), (4), (5) and (6) as Brunnstrom stage IV, otherwise, outputting an evaluation result as Brunnstrom stage III; on the basis of judging as the IV stage, judging that the patient can finish one of (7), (8), (9) and (10) as a Brunnstrom V stage, and otherwise, outputting an evaluation result as the Brunnstrom IV stage; and (3) on the basis of judging as the stage V, judging that the patient can complete the two action tasks (11) and (12), namely judging as the stage Brunnstrom VI, and otherwise, outputting an evaluation result as Brunnstrom V.

Background

The incidence and the prevalence of stroke in China rise year by year, and the incidence age is in a descending trend. Meanwhile, with the development and progress of medical science, the death rate of stroke patients is gradually reduced, and the rate of disability and the number of disabled people are increased continuously. Alleviate the dysfunction of the patient and help the patient to return to the family and the society again and need the intervention of rehabilitation therapy. Rehabilitation assessment is the basis of rehabilitation, and by evaluating the motor function of the upper limbs of a stroke patient, the motor disorder degree of the upper limbs of the patient can be determined, and the rehabilitation treatment items and the training plan can be guided to be made. In some areas where economic development is relatively laggard, the number of rehabilitation professionals is far from sufficient to serve local disabled people, and stroke patients often cannot obtain timely and accurate functional assessment, so that subsequent rehabilitation treatment is influenced.

Currently, the domestic upper limb motor function assessment for stroke hemiplegia patients mainly takes the manual assessment dominated by rehabilitation therapists as main assessment. During the assessment process, the contact between the therapist and the patient's body and the interference of the therapist to the patient's psychological state can cause the generation process of the brain movement plan to change, and finally the performance of the patient in the assessment process is influenced, so that the assessment result is inaccurate. Meanwhile, the subjective property of manual evaluation is strong, and the same therapist is required to evaluate the patient in the rehabilitation process.

The Microsoft Kinect is a somatosensory device, realizes tracking of a human body by utilizing a skeletal tracking technology and depth image data based on environment, and is mainly used for capturing the motion of a patient in a three-dimensional space and acquiring three-dimensional coordinate information of a human body joint point in the process of performing evaluation action by the patient.

Disclosure of Invention

In order to solve the above problems, an object of the present invention is to provide an evaluation method capable of determining the recovery stage of upper limb motor function of a paraplegia patient after stroke even if a rehabilitation doctor or a therapist is absent. The method comprises the steps of obtaining three-dimensional coordinate information of human body joint points when a patient executes evaluation actions by using Microsoft Kinect 2.0, decomposing the evaluation actions through an action algorithm, carrying out calculation analysis on the obtained joint point coordinates according to a flow, judging the degree of sufficiency of the execution of the actions of the patient, finally counting the completion conditions of all the evaluation actions of the patient, and determining the recovery stage of the upper limb movement function of the patient.

In order to achieve the purpose, the invention adopts the following technical scheme: a method for evaluating the motor function of the upper limbs of a cerebral apoplexy patient based on Microsoft Kinect 2.0 comprises the following steps:

2) and (3) executing actions: the hemiplegic patient executes evaluation action under the guidance of characters, images and sounds in a screen;

3) data acquisition: acquiring three-dimensional coordinate information of key points of a human body of a patient by using a Kinect camera arranged in a treatment room;

5) and (4) counting results: and (4) counting the completion conditions of all evaluation actions of the patient and determining the recovery stage of the upper limb movement function of the patient.

Further, establishing a human body space coordinate system in the step 1): the HEAD, NECK, SPINE SHOULDER, SPINE MID, SPINE BASE are assumed to be on a straight line. The SPINE BASE skeleton coordinate point is (0, 0), the left-right direction is (X), the up-down direction is (Y), and the front-back direction is (Z). Positive values to the right, up and forward, and negative values to the left, down and backward. Further, the step 1) evaluates the joint point parameter definition related to the motion, and directly calls the own parameter definition of the Microsoft Kinect 2.0 system, specifically:

further, the step 1) initializes each parameter, specifically, determines the length of the arm and the length of the trunk of the upper limb of the human body.

Length of left arm: l _ AMLT = J _ SRLT _ Y-J _ HDLT _ Y

Length of right arm: l _ AMRT = J _ SRRT _ Y-J _ HDRT _ Y

Left palm to elbow length: l _ AMLT _1 = J _ EWLT _ Y-J _ HDLT _ Y

Right hand palm to elbow joint length: l _ AMRT _1 = J _ EWRT _ Y-J _ HDRT _ Y

Left elbow to shoulder length: l _ AMLT _2 = J _ SRLT _ Y-J _ EWLT _ Y

Right elbow to shoulder length: l _ AMRT _2 = J _ SRRT _ Y-J _ EWRT _ Y.

Further, the step 1) sets a conditional algorithm of the basic gesture: the tolerance is set to delt = 5 cm.

When the human body is in the end sitting position, the conditions met by all actions are as follows:

the left hand droops naturally without bending: i J _ SRLT _ X-J _ HDLT _ X < delt & | J _ SRLT _ Y-J _ HDLT _ Y-L _ AMLT | < delt

The right hand droops naturally without bending: i J _ SRRT _ X-J _ HDRT _ X | < delt & | J _ SRRT _ Y-J _ HDRT _ Y-L _ AMRT | < delt.

The straight satisfying conditions of the left arm and the right arm are as follows:

|（（S_SRLT_X - S_HDLT_X）²+（S_SRLT_Y - S_HDLT_Y）² + S_SRLT_Z - S_HDLT_Z）²）^0.5 - L_AMLT| < delt

|（（S_SRRT_X - S_HDRT_X）² +（S_SRRT_Y - S_HDRT_Y）²+ S_SRRT_Z - S_HDRT_Z）²）^0.5 - L_AMRT| < delt。

further, the step 2) of evaluating the actions improves the actions in the Brunnstrom upper limb motor function evaluation table, and specifically comprises the following steps:

(2) the affected hand is placed at the waist position on the same side, and the affected hand is extended to the knee on the opposite side from the front of the body;

(3) the affected hand is placed in the middle of the thigh on the same side, extended to the back and tried to touch the back of the waistband;

(6) on the basis of the action (5), the forearm is rotated backwards (namely the forearm is rotated to enable the palm center to face upwards);

(10) on the basis of the action (9), the forearm is rotated backwards (namely the forearm is rotated to enable the palm center to face upwards);

(11) the elbow joint is flexed with the upper arm on the front side of the body and the finger tip on the ipsilateral shoulder. Performing rapid lifting action by touching shoulder with fingertip, measuring time required for repeating 10 times, measuring healthy side, and comparing the measurement results of the affected side and healthy side (determining that the time required for the affected side is 1.5 times or less of that of the healthy side is sufficient);

(12) the affected hand is placed in the middle of the thigh on the same side, the chin is touched with the fingers at the fastest speed, and then the affected hand is placed back on the thigh. Measuring the time required for repeating 10 times, measuring the healthy side, and comparing the measurement results of the affected side and the healthy side (determining that the time required for the affected side is 1.5 times or less of the healthy side is sufficient);

actions (1) to (12) are all performed in the sitting position.

Further, the step 4) decomposes the evaluation action through an action algorithm, and the specific process is as follows: based on the human body joint points captured by the Kinect, a motion algorithm capable of decomposing complex evaluation motions is designed by utilizing the coordinate reference relationship of X, Y, Z among the joint points of all parts of the human body. And substituting the three-dimensional coordinates of the relevant joint points in the action process of the patient captured by the Kinect into an algorithm, judging how many decomposition actions are completed by the patient, and further determining the execution sufficiency of the evaluation action.

Further, the specific process of counting the completion conditions of each evaluation action of the patient in the step 5) is as follows: the patient can finish one of the actions (1) and (2) and is judged as Brunnstrom III stage, otherwise, the evaluation result is output as Brunnstrom I to Brunnstrom II stage. And on the basis of judging as stage III, judging that the patient can finish one of (3), (4), (5) and (6) as Brunnstrom stage IV, otherwise, outputting the evaluation result as Brunnstrom stage III. And (3) on the basis of judging as the IV stage, judging that the patient can finish one of (7), (8), (9) and (10) as a Brunnstrom V stage, and otherwise, outputting an evaluation result as the Brunnstrom IV stage. And (3) on the basis of judging as the stage V, judging that the patient can complete the two action tasks (11) and (12), namely judging as the stage Brunnstrom VI, and otherwise, outputting an evaluation result as Brunnstrom V.

Compared with the prior art, the invention has the beneficial effects that:

the rehabilitation evaluation system formed by combining Microsoft Kinect 2.0, a computer and a monitor is low in price, so that the rehabilitation evaluation system can be widely applied to basic medical rehabilitation institutions in relatively laggard areas of economic development, and the popularization of the rehabilitation evaluation system is possible.

In the medical field, the recovery stage of a cerebral apoplexy patient is judged, and professional personnel such as a rehabilitation doctor, a therapist and the like are required to make analysis and judgment according to a rating scale. According to the technical method for judging the recovery stage of the cerebral apoplexy patient by the joint point information acquired by Microsoft Kinect 2.0, a doctor or a therapist is not required to participate in the evaluation process, so that the method has high utilization value in areas with insufficient rehabilitation professionals.

Compared with the traditional manual evaluation which is dominant by a rehabilitation therapist, the evaluation action is decomposed through an action algorithm, the obtained joint point coordinates are calculated and analyzed according to a flow, the action execution sufficiency of the patient is judged, and then the upper limb motor function recovery stage of the stroke patient is judged.

And 3) acquiring data in the step 3), namely capturing joint point data in the evaluation process of the patient in a non-contact manner by using a Kinect somatosensory camera, so that the contact between a therapist and the body of the patient in the traditional manual evaluation and the influence of the physical contact on the psychological state of the patient are avoided, interference factors in the generation process of a brain movement plan of the patient are reduced, and the evaluation result obtained by the technical scheme has higher accuracy.

In the implementation process of the method for evaluating the upper limb motor function of the cerebral apoplexy patient based on Microsoft Kinect 2.0, no sensor is required to be worn on the body of the patient, so that the use difficulty is reduced, and the potential safety hazard of the equipment worn by the patient is also reduced. Because no external equipment inputs physical sensory stimulation, the patient can more easily obtain the real feeling of the body in the device, and the attention is focused. In addition, the patient does not need extra operation and learning, only needs to finish the evaluation action under the guidance of characters, images and sounds in the screen, can realize the interaction with the computer, and has the characteristics of simplicity, practicability and good experience. This is beneficial for motivating patient involvement, improving patient compliance.

Drawings

Fig. 1 is a flowchart of a method for assessing upper limb motor function of a stroke patient.

Fig. 2 is a flowchart of determination of the recovery stage of a stroke patient.

FIG. 3 is an action decomposition flow diagram of an embodiment of the invention.

Detailed Description

The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

As shown in fig. 1, the flow of the method for assessing the motor function of the upper limb of a stroke patient based on Microsoft Kinect 2.0 of the present invention specifically includes the following steps:

1) initially defined: establishing a human body space coordinate system, calling joint point parameter definitions related to the evaluation actions, initializing various parameter values, and setting a condition algorithm of a basic posture.

Establishing a human body space coordinate system in the step 1): the HEAD, NECK, SPINESHOULDER, SPINEMID, SPINEBASE are assumed to be on a straight line. The SPINEBASE skeleton coordinate point is (0, 0), the left-right direction is (X), the up-down direction is (Y), and the front-back direction is (Z). Positive values to the right, up and forward, and negative values to the left, down and backward.

Directly calling the self-contained parameter definition of the Microsoft Kinect 2.0 system, specifically:

the testee sits right on the chair, the hip, knee and ankle joints of the lower limb flex by 90 degrees, the two upper limbs naturally droop on the body side, each parameter is initialized, and the length values of the arms and the trunk of the upper limbs of the testee are determined.