阅读说明：本技术 一种基于OpenPose算法的术后康复方法及其数据采集装置支架 (OpenPose algorithm-based postoperative rehabilitation method and data acquisition device support thereof ) 是由 周晋 李晶晶 曾杰 侯科宇 *** 贺易楠 于 2020-06-28 设计创作，主要内容包括：本发明公开了一种基于OpenPose算法的术后康复方法,其技术方案要点是包括以下步骤：步骤1：终端通过其内置摄像头对需要居家康复的对象进行视频采集；步骤2：将视频数据经过终端进行压缩编码,便于传输；步骤3：终端将数据上传至服务器；步骤4：服务器对数据进行解码；步骤5：将数据传输至相应终端,确定算法使用在线模式或者离线模式；步骤6：根据算法模式算出最终结果显示屏幕终端；该术后康复方法能够有效的检测患者在家的康复情况并且辅助其完成康复动作,并且能够远程分析病人的运动协调性或者康复情况,且技术投入成本较低便于实现。(The invention discloses an OpenPose algorithm-based postoperative rehabilitation method, which has the technical scheme key points that the method comprises the following steps: step 1: the terminal carries out video acquisition on an object needing home rehabilitation through a built-in camera of the terminal; step 2: the video data is compressed and encoded through a terminal, so that the transmission is convenient; and step 3: the terminal uploads the data to the server; and 4, step 4: the server decodes the data; and 5: transmitting the data to a corresponding terminal, and determining whether an algorithm uses an online mode or an offline mode; step 6: calculating a final result display screen terminal according to the algorithm mode; the postoperative rehabilitation method can effectively detect the rehabilitation condition of the patient at home and assist the patient to complete the rehabilitation action, can remotely analyze the movement coordination or the rehabilitation condition of the patient, and has low technical input cost and convenient implementation.)

1. An OpenPose algorithm-based postoperative rehabilitation method is characterized by comprising the following steps: the method comprises the following steps:

step 1: the terminal carries out video acquisition on an object needing home rehabilitation through a built-in camera of the terminal;

step 2: the video data is compressed and encoded through a terminal, so that the transmission is convenient;

and step 3: the terminal uploads the data to the server;

and 4, step 4: the server decodes the data;

and 5: transmitting the data to a corresponding terminal, and determining whether an algorithm uses an online mode or an offline mode;

step 6: and calculating a final result according to the algorithm mode and displaying the screen terminal.

2. The OpenPose algorithm-based postoperative rehabilitation method according to claim 1, characterized in that: and 5, selecting an online mode: detecting key points of bones and recognizing postures;

step 61: downloading the identification data to the terminal;

step 62: performing logic judgment according to the task terminal;

and step 63: and displaying the gesture recognition result on a screen terminal by the terminal.

3. The OpenPose algorithm-based postoperative rehabilitation method according to claim 1, characterized in that: step 5, selecting off-line Mohs: detecting key points of bones and recognizing postures;

step 64: evaluating the movement coordination;

step 65: the doctor assesses the subject's rehabilitation status and adjusts the rehabilitation strategy;

and step 66: and sending the updated rehabilitation strategy to a terminal display screen terminal.

4. The OpenPose algorithm-based postoperative rehabilitation method according to claim 1, 2 or 3, characterized in that: the OpenPose algorithm identification comprises the following steps:

and 7: inputting videos/pictures;

and 8: predicting a neural network;

and step 9: finding out bone joint coordinates;

step 10: determining an articulation mode;

step 11: splicing the joints into a human body;

step 12: outputting the coordinates of each joint of each person;

step 13: and obtaining a final result display screen terminal.

5. A data acquisition device holder based on the openpos algorithm of claim 1, characterized in that: including display screen (2) and mobilizable support (1), can dismantle between display screen (2) and the support (1) and be connected, support (1) liftable, support (1) below is provided with and is used for supporting piece (3) of contradicting with ground, still be provided with on support (1) and be used for promoting support (1) rolling subassembly (4), rolling subassembly (4) can be propped up and keep away from ground.

6. The OpenPose algorithm-based data acquisition device holder according to claim 5, wherein: support (1) is provided with connecting block (5) including branch (11) of at least two between branch (11), connecting block (5) are used for linking together many branch (11), branch (11) are including upper boom (12) and lower beam (13), lower beam (13) outside is located to upper boom (12) cover, upper boom (12) can be along the reciprocal sliding of length direction of lower beam (13), be fixed with on upper boom (12) and insert piece (6), it can be along with upper boom (12) removal to insert piece (6), be provided with slot (61) in inserting piece (6), slot (61) are used for supplying outside anchorage nail (62) to insert fixedly.

7. The OpenPose algorithm-based data acquisition device holder according to claim 6, wherein: roll subassembly (4) and set up on lower beam (13), roll subassembly (4) including first rotary rod (41) and second rotary rod (42), the below of first rotary rod (41) is provided with rolls wheel (43), enclose between first rotary rod (41), second rotary rod (42) and lower beam (13) and establish and be the triangle structure setting.

8. The openpos algorithm-based data acquisition device holder according to claim 7, wherein: be provided with dead lever (7) on lower beam (13), be provided with on second rotary rod (42) and supply dead lever (7) male fixed slot (71), when dead lever (7) tip inserted in fixed slot (71), lower beam (13), dead lever (7) and second rotary rod (42) enclose to establish and form the triangle-shaped structure.

Technical Field

The invention relates to an OpenPose algorithm-based postoperative rehabilitation method and a data acquisition device support thereof.

Background

After lower limb surgery or sports strain, special rehabilitation training is required, and professional rehabilitation training can ensure that the injured lower limb recovers to the original sports level; conversely, if the rehabilitation is not in place, irreversible motor function damage is formed. However, the main problem of developing rehabilitation training in hospitals is that hospital sites and rehabilitation therapists are limited, require appointment and have short training time, and cannot fully cover the rehabilitation period; meanwhile, the hospital rehabilitation cost is high, the universal charge is 300-500 yuan/hour, tens of thousands of yuan of cost is needed if the whole rehabilitation period is completed, the rehabilitation cost is high, under the common condition, a doctor recommends the patient to walk at home or carry out rehabilitation training of specific actions, however, the rehabilitation effect and the movement effect of the patient cannot be timely fed back to the doctor, and therefore timely rehabilitation effect evaluation is carried out.

Disclosure of Invention

Aiming at the defects in the prior art, the invention aims to provide an OpenPose algorithm-based postoperative rehabilitation method and a data acquisition device support thereof.

In order to achieve the purpose, the invention provides the following technical scheme: an OpenPose algorithm-based postoperative rehabilitation method comprises the following steps:

step 1: the terminal carries out video acquisition on an object needing home rehabilitation through a built-in camera of the terminal;

step 2: the video data is compressed and encoded through a terminal, so that the transmission is convenient;

and step 3: the terminal uploads the data to the server;

and 4, step 4: the server decodes the data;

and 5: transmitting the data to a corresponding terminal, and determining whether an algorithm uses an online mode or an offline mode;

step 6: and calculating a final result according to the algorithm mode and displaying the screen terminal.

Further, an online mode is selected in step 5: detecting key points of bones and recognizing postures;

step 61: downloading the identification data to the terminal;

step 62: performing logic judgment according to the task terminal;

and step 63: and displaying the gesture recognition result on a screen terminal by the terminal.

Further, step 5 selects the offline mode: detecting key points of bones and recognizing postures;

step 64: evaluating the movement coordination;

step 65: the doctor assesses the subject's rehabilitation status and adjusts the rehabilitation strategy;

and step 66: and sending the updated rehabilitation strategy to a terminal display screen terminal.

Further, openpos algorithm identification includes the following steps:

and 7: inputting videos/pictures;

and 8: predicting a neural network;

and step 9: finding out bone joint coordinates;

step 10: determining an articulation mode;

step 11: splicing the joints into a human body;

step 12: outputting the coordinates of each joint of each person;

step 13: and obtaining a final result display screen terminal.

The utility model provides a data acquisition device support based on OpenPose algorithm, includes display screen and mobilizable support, can dismantle between display screen and the support and be connected, the support liftable, the support below is provided with the contact piece that is used for contradicting with ground, still be provided with on the support and be used for promoting the rolling roll subassembly of support, the roll subassembly can be propped up and keep away from ground.

Further, the support is provided with the connecting block including two at least branches between the branch, the connecting block is used for linking together many branches, branch includes upper boom and lower beam, the upper boom cover is located the lower beam outside, the upper boom can slide along the length direction reciprocating of lower beam, be fixed with on the upper boom and insert the piece, it can be along with the upper boom removal to insert the piece, it is provided with the slot in the piece to insert, the slot is used for supplying outside binding nail insert fixed.

Further, the rolling assembly sets up on the lower beam, the rolling assembly includes first rotary rod and second rotary rod, the below of first rotary rod is provided with the wheel that rolls, enclose between first rotary rod, second rotary rod and the lower beam and establish and be the triangle structure setting.

Further, be provided with the dead lever on the lower beam, be provided with on the second rotary rod and supply dead lever male fixed slot, when the dead lever tip inserted to the fixed slot in, lower beam, dead lever and second rotary rod enclose to establish and form the triangle-shaped structure.

The invention has the beneficial effects that: 1. the online mode and the offline mode are coordinated with each other, and the training of the rehabilitation action of the patient at home can be effectively assisted.

2. Through the setting of data acquisition device support, can be convenient and effectual removal data acquisition device, provide the convenience for the patient.

Drawings

FIG. 1 is a block diagram of a postoperative rehabilitation method;

FIG. 2 is a schematic diagram of an OpenPose algorithm block;

FIG. 3 is a diagram of the connection between the display screen and the stand;

FIG. 4 is a schematic view of a specific structure of the stent;

fig. 5 is a structural schematic diagram of the bottom of the bracket.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings and examples. In which like parts are designated by like reference numerals. It should be noted that the terms "front," "back," "left," "right," "upper" and "lower" used in the following description refer to directions in the drawings, and the terms "bottom" and "top," "inner" and "outer" refer to directions toward and away from, respectively, the geometric center of a particular component.

Referring to fig. 1, the post-operation rehabilitation method based on the openpos algorithm of the present embodiment includes the following steps:

step 1: the terminal carries out video acquisition on an object needing home rehabilitation through a built-in camera of the terminal; step 2: the video data is compressed and encoded through a terminal, so that the transmission is convenient; and step 3: the terminal uploads the data to the server; and 4, step 4: the server decodes the data; and 5: transmitting the data to a corresponding terminal, and determining whether an algorithm uses an online mode or an offline mode; step 6: and calculating a final result according to the algorithm mode and displaying the screen terminal.

The specific improvement is as follows: the offline mode refers to: after the patient carries out rehabilitation training for a period of time, the situation is certainly changed, the patient can record walking videos, then the server calculates CRP through all joint angles to analyze the movement coordination of the patient, and quantitative parameters are given. Therefore, when the doctor is available, the doctor can update the rehabilitation task by watching the walking video of the patient and adjusting the rehabilitation training means by the quantitative parameters, and the off-line mode analyzes the movement coordination and the continuous relative phase, so that the professional data is obscure and unintelligible, the patient cannot understand the data, only the doctor and the system know what the data represent, and the off-line analysis is analysis after recording the rehabilitation condition of the patient.

The off-line vermicelli mode working mode of the postoperative rehabilitation state comprises the following steps:

firstly, a terminal (exercise rehabilitation auxiliary equipment) carries out video acquisition on an object needing home rehabilitation through a built-in camera.

And secondly, the video data is compressed and encoded through the terminal, so that the transmission is convenient.

And thirdly, the terminal uploads the data to the server.

And fourthly, the server decodes the data.

And fifthly, the server identifies the postures of the video characters by using an OpenPose algorithm to obtain the data of each key point.

The server calculates a continuous relative phase (continuousLatifevelPhase) using the bone joint point data and evaluates the motion coordination of the object.

And seventhly, the doctor evaluates the postoperative rehabilitation condition and adjusts the rehabilitation strategy according to the movement coordination of the object, and issues a new rehabilitation task.

And the server sends the new rehabilitation strategy to the terminal, and the terminal updates the rehabilitation task.

As a specific embodiment of the improvement, the online mode is selected in step 5: detecting key points of bones and recognizing postures; step 61: downloading the identification data to the terminal; step 62: performing logic judgment according to the task terminal; and step 63: and displaying the gesture recognition result on a screen terminal by the terminal.

The specific improvement is as follows: the online mode refers to: the doctor gives a specific rehabilitation action task to the patient and needs the patient to complete the task on time, so that the patient needs to use a terminal (mobile phone) to open a camera, analyze the completion condition of the patient by using a posture analysis and provide a completion prompt (voice prompt can be given in the completion process), the whole process is automatically completed by the system, and the doctor does not need to participate in the whole process.

The real-time analysis mode working mode of postoperative rehabilitation training comprises the following steps:

firstly, a terminal (exercise rehabilitation auxiliary equipment) carries out video acquisition on an object needing home rehabilitation through a built-in camera.

And secondly, the video data is compressed and encoded through the terminal, so that the transmission is convenient.

And thirdly, the terminal uploads the data to the server.

And fourthly, the server decodes the data.

And fifthly, the server identifies the postures of the video characters by using an OpenPose algorithm to obtain the data of each key point.

Sixthly, the server sends the identification data back to the terminal.

And seventhly, judging whether the object completes the designated postoperative rehabilitation action or not by using the skeletal key point data according to the task requirement.

And displaying the logic judgment result on the terminal screen.

As a specific embodiment of the improvement, step 5 selects the offline mode: detecting key points of bones and recognizing postures; step 64: evaluating the movement coordination; step 65: the doctor assesses the subject's rehabilitation status and adjusts the rehabilitation strategy; and step 66: and sending the updated rehabilitation strategy to a terminal display screen terminal.

As a specific embodiment of the improvement, refer to fig. 2: the OpenPose algorithm identification comprises the following steps: and 7: inputting videos/pictures; and 8: predicting a neural network; and step 9: finding out bone joint coordinates; step 10: determining an articulation mode; step 11: splicing the joints into a human body; step 12: outputting the coordinates of each joint of each person; step 13: and obtaining a final result display screen terminal.

The specific improvement is as follows: based on openpos human body 2D pose recognition, after the video/picture input system, it first predicts the confidence map of the skeletal key points through 2-way convolutional neural network VGG-19 — branch one uses target detection technology, and branch two uses partial association domain (PartAffinityFields) technology to predict the 2D vectors of the connection modes (limbs) between the skeletal key points. And finally, analyzing the 2D human body posture from the confidence graph of the key points and the partial association domain by using greedy reasoning, and respectively and independently outputting the coordinates of each joint point of each person.

The utility model provides a data acquisition device support based on OpenPose algorithm, includes support 1 and sets up in display screen 2 of 1 top of support, can dismantle between display screen 2 and the support 1 and be connected, support 1 liftable, 1 below of support is provided with and is used for the conflict piece 3 of contradicting with ground, still be provided with on the support 1 and be used for promoting the rolling subassembly 4 of support 1, rolling subassembly 4 can be propped up and keep away from ground.

The specific improvement is as follows: referring to fig. 3 and 4: the screen is used for displaying a final logic judgment result realized by using an OpenPose algorithm, in order to facilitate carrying, the display screen 2 and the bracket 1 are detachably connected, a clamping structure is arranged below the display screen 2, the clamping structure can refer to the clamping technology of a vehicle-mounted mobile phone bracket, the clamping structure is the prior art, so that the structure is not written more, in order to ensure that the display screen 2 can be placed on the ground more stably, the bottom of the bracket 1 is provided with a contact piece 3 which is in contact with the ground and a rolling component 4 which is used for pushing the bracket 1 to roll, when the contact piece 3 is in contact with the ground, the bracket 1 can be placed on the ground stably, when the bracket 1 needs to be pushed to move, the rolling component 4 is rotated downwards, the rolling component 4 is in contact with the ground, when the rolling component 4 is in contact with the ground, the contact piece 3 can be upwards supported to be far away from the ground, the support 1 can be stably placed on the ground, and most of strength can be saved when the support 1 needs to be moved.

As a specific embodiment of the improvement, refer to fig. 5: support 1 is including the branch 11 of at least two, is provided with connecting block 5 between the branch 11, connecting block 5 is used for linking together many branches 11, branch 11 includes upper boom 12 and lower beam 13, upper boom 12 cover is established with the lower beam 13 outside, upper boom 12 can be along the reciprocal sliding of the length direction of lower beam 13, be fixed with on the upper boom 12 and insert piece 6, insert piece 6 and can remove along with upper boom 12, it is provided with slot 61 in the piece 6 to insert, slot 61 is used for supplying outside binding nail 62 to insert fixedly.

The specific improvement is as follows: support 1 includes upper boom 12 and lower beam 13, and support 1 realizes going up and down through the removal of upper boom 12 and lower beam 13, and the inside cavity of upper boom 12, lower beam 13 can be taken in to upper boom 12 inside, and after upper boom 12 and lower beam 13 confirmed the position, when needing to fix the position of upper boom 12 and lower beam 13, it can in inserting slot 61 with fastening nail 62.

As a modified specific embodiment, the rolling assembly 4 is disposed on the lower rod 13, the rolling assembly 4 includes a first rotating rod 41 and a second rotating rod 42, a rolling wheel 43 is disposed below the first rotating rod 41, and a triangular structure is defined between the first rotating rod 41, the second rotating rod 42 and the lower rod 13.

The specific improvement is as follows: when rolling assembly 4 need not be used, with first rotary rod 41 and second rotary rod 42 clockwise rotation, can pack up rolling wheel 43 this moment, make conflict pole and ground conflict be connected, so just can guarantee that support 1 can be steady stand subaerial.

As a modified specific embodiment, the fixing rod 7 is arranged on the lower rod 13, the fixing groove 71 for inserting the fixing rod 7 is arranged on the second rotating rod 42, and when the end of the fixing rod 7 is inserted into the fixing groove 71, the lower rod 13, the fixing rod 7 and the second rotating rod 42 are enclosed to form a triangular structure.

The specific improvement is as follows: when the bracket 1 needs to be moved, the first rotating rod 41 and the second rotating rod 42 can be rotated counterclockwise, so that the rolling wheel 43 can contact the ground, when the rolling wheel 43 contacts the ground, the abutting member 3 is upwards supported, at this time, the rolling wheel 43 rolls, the bracket 1 can be moved, when the rolling wheel 43 contacts the ground, in order to prevent the first rotating rod 41 and the second rotating rod 42 from rotating during movement, the fixing rod 7 is arranged on the lower rod 13, the fixing groove 71 is arranged on the second rotating rod 42, and when the fixing rod 7 is inserted into the fixing groove 71, the first rotating rod 41 and the second rotating rod 42 can be prevented from rotating during movement, thereby causing the bracket 1 to collapse.

The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may occur to those skilled in the art without departing from the principle of the invention, and are considered to be within the scope of the invention.