CN116228727A - Image-based detection method and device for epidermal spine line of human back - Google Patents

Abstract

This application relates to the field of scoliosis detection, in particular to an image-based detection method and device for the epidermal spine line of the human back. The method includes acquiring the back depth image of the subject to be detected, and determining the target point cloud data based on the back depth image ; Determine the preset number of feature point coordinates of the subject's spine in the back depth image according to the target point cloud data; perform curve fitting according to all the feature point coordinates to generate the target epidermal spine curve Determine the target scoliosis angle according to the target epidermal spine curve, determine whether the subject's spine is scoliosis according to the angle value of the target scoliosis angle, this method has no radiation in the whole process, and is proficient for operators The requirements are low, and the result of scoliosis can be quickly and efficiently obtained through deep learning and curve fitting. It is suitable for scoliosis detection of large groups, especially primary and middle school students.

Description

An image-based detection method and device for human back skin spine line

technical field

The present application relates to the field of scoliosis detection, in particular to an image-based detection method and device for human back epidermal spine lines.

Background technique

The epidermal spine line on the back of the human body is an important indicator of scoliosis. Scoliosis is a three-dimensional deformity of the spine, including sequence abnormalities in the coronal, sagittal and axial positions. It mostly occurs in children and adolescents. 2%-4% of adolescents suffer from scoliosis. At present, there are more than 3 million patients with scoliosis in China, and the number of scoliosis patients is increasing at a rate of 300,000 per year. The national policy clearly indicates that in order to pay attention to and ensure the healthy development of the spine of primary and middle school students, the side of the spine Scoliosis examination is included in the physical examination items of primary and middle school students. Therefore, it is imminent to use artificial intelligence technology to efficiently detect the spine line of the human back surface and judge its scoliosis.

At present, the detection of scoliosis in the physical examination of primary and middle school students is mainly carried out by professional inspectors using the scoliosis measuring instrument (curvature ruler) to screen scoliosis. This method is inefficient and requires relatively high levels of technique and proficiency High, so there is a large deviation in the results measured by personnel with different proficiency levels. At present, the domestic and foreign detection methods for scoliosis mainly include X-ray, EOS, ultrasound, and moire.

Although X-ray is low in cost, simple and easy to perform, and more accurate in measurement, the radiation dose is large, which is harmful to the human body of the target primary and secondary school population; EOS is costly and still carries a small amount of radiation; ultrasound has high requirements for doctors' techniques and large errors , low efficiency, not conducive to large-scale screening; Moiré pattern can only roughly judge the situation of the target.

Contents of the invention

In view of the above problems, this application is proposed to provide an image-based method and device for detecting the epidermal spine line of the back of the human body, which overcome the above problems or at least partially solve the above problems.

An image-based detection method of human back skin spine line, said method comprising:

Obtaining the depth image of the back of the subject, and determining the target point cloud data according to the depth image of the back;

determining a preset number of feature point coordinates of the subject's spine in the back depth image according to the target point cloud data;

Carrying out curve fitting according to all the feature point coordinates to generate the target epidermal spine curve;

A target scoliosis angle is determined according to the target epidermal spine curve, and whether the subject's spine is scoliosis is determined according to the angle value of the target scoliosis angle.

Preferably, the acquisition of the back depth image of the subject, and determining the target point cloud data according to the back depth image include:

collecting a depth image of the subject's back through a depth camera, and acquiring the depth image of the back from the depth camera;

Acquiring the calibrated internal parameters of the depth image, and determining the three-dimensional point cloud coordinates corresponding to each pixel in the back depth image according to the internal parameters,

The three-dimensional point cloud data is determined according to all the three-dimensional coordinates of the point cloud, and a preset filtering algorithm is used to perform noise reduction processing on the three-dimensional point cloud data to obtain the target point cloud data.

Preferably, the determining a preset number of feature point coordinates of the subject's spine in the back depth image according to the target point cloud data includes:

Invoking a preset stacked hourglass network model, inputting the target point cloud data into the stacked hourglass network model, and training the stacked hourglass network model;

Outputting the preset number of feature point coordinates through the trained stacked hourglass network model.

Preferably, the calling preset stacked hourglass network model includes:

Determining the number of output neurons of each layer in the stacked hourglass network model and the scale coefficient after weight normalization;

According to the number of neurons and the proportional coefficient, weights are initialized in the stacked hourglass network model through a Kaiming network weight initialization method.

Preferably, the curve fitting is performed according to all the feature point coordinates to generate the target epidermal spine curve, including:

The interparc curve interpolation method is used to perform polynomial interpolation fitting processing on all the feature point coordinates to generate the target epidermal spine curve.

Preferably, the interparc curve interpolation method is used to perform polynomial interpolation fitting processing on all the feature point coordinates, including:

Obtain the initial epidermal spine curve according to the Interparc curve interpolation method and all the characteristic point coordinate fittings;

Determining the length value of the initial epidermal spine curve, and equally dividing the initial epidermal spine curve according to the length value;

Perform linear interpolation processing on the initial epidermal spine curve after the equalization process to obtain the coordinates of all target feature points;

According to all the target feature point coordinates and the Interparc curve interpolation method, the target epidermis spine curve is generated by fitting.

Preferably, determining the target scoliosis angle according to the target epidermal spine curve, and determining whether the subject's spine is scoliosis according to the angle value of the target scoliosis angle include:

determining the target normal vector according to the target skin spine curve;

determining a target scoliosis angle according to the target normal vector;

When the angle value of the target scoliosis angle is less than the preset scoliosis angle value, it means that the detected person’s spine is normal; when the angle value of the target scoliosis angle is greater than or equal to the preset scoliosis angle value When the angle value is smaller, it means that the detected person's spine is scoliosis.

Also provided is an image-based detection device for epidermal spine lines on the back of a human body, said device comprising:

The point cloud determination module is used to obtain the back depth image of the detected person, and determine the target point cloud data according to the back depth image;

A coordinate determination module, configured to determine a preset number of feature point coordinates of the subject's spine in the back depth image according to the target point cloud data;

A curve generation module, configured to perform curve fitting according to all the feature point coordinates to generate the target epidermal spine curve;

A scoliosis judging module, configured to determine a target scoliosis angle according to the target epidermal spine curve, and determine whether the subject's spine is scoliosis according to the angle value of the target scoliosis angle.

An apparatus comprising a processor, a memory, and a computer program stored on the memory and capable of running on the processor, the computer program being executed by the processor to implement an image-based The steps of the method for detecting the epidermal spine line of the back of the human body.

A computer-readable storage medium, storing a computer program on the computer-readable storage medium, when the computer program is executed by a processor, the steps of the image-based method for detecting the epidermal spine line on the back of a human body as described above are realized.

This application has the following advantages:

In an embodiment of the present application, by acquiring the back depth image of the subject, determining the target point cloud data according to the back depth image, and determining the depth of the subject's spine in the back according to the target point cloud data A preset number of feature point coordinates in the image; perform curve fitting according to all the feature point coordinates to generate a target epidermal spine curve; determine the target scoliosis angle according to the target epidermal spine curve, and determine the target scoliosis angle according to the target spine side The angle value of the bending angle determines whether the subject’s spine is bent sideways; by determining the subject’s spine feature points and coordinates in the point cloud data of the back depth image, and then performing curve fitting on all these feature point coordinates directly determine the target scoliosis angle from the target epidermal spine curve generated by fitting, so as to judge whether the subject’s spine is scoliosis according to the target scoliosis angle; the above scheme has no radiation in the whole process and is very skilled for operators Low degree requirements, through deep learning and curve fitting can quickly and efficiently obtain the result of scoliosis, suitable for scoliosis detection of large groups, especially primary and middle school students.

Description of drawings

In order to illustrate the technical solution of the present application more clearly, the accompanying drawings that need to be used in the description of the present application will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. Ordinary technicians can also obtain other drawings based on these drawings without paying creative labor.

Fig. 1 shows a flow chart of the steps of an image-based method for detecting the epidermal spine line on the back of a human body provided by an embodiment of the present application;

Fig. 2 shows a schematic structural diagram of an image-based detection device for epidermal spine lines on the back of a human body provided by an embodiment of the present application;

FIG. 3 shows a schematic structural diagram of a computer device for an image-based method for detecting the epidermal spine line on the back of a human body according to the present invention.

Detailed ways

In order to make the purpose, features and advantages of the present application more obvious and understandable, the present application will be further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Apparently, the described embodiments are some of the embodiments of the present application, but not all of them. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of this application.

In view of the current high incidence of scoliosis among primary and middle school students, existing detection methods such as X-ray, EOS, ultrasound, and moiré patterns have application limitations. This application proposes a method based on computer vision images combined with a deep learning model to analyze the Detect the scoliosis of the spine, so as to achieve the goal of zero radiation, low error and high efficiency in scoliosis detection.

Please refer to FIG. 1 , which shows a flow chart of steps of an image-based detection method for epidermal spine lines on the back of a human body provided by an embodiment of the present application. The method includes the following steps:

S110, acquiring a depth image of the back of the subject, and determining target point cloud data according to the depth image of the back;

S120. Determine a preset number of feature point coordinates of the subject's spine in the back depth image according to the target point cloud data;

S130, performing curve fitting according to all the feature point coordinates to generate a target epidermal spine curve;

S140. Determine a target scoliosis angle according to the target epidermal spine curve, and determine whether the subject's spine is scoliosis according to an angle value of the target scoliosis angle.

It should be noted that the image features of the back of the human body are diverse, and the coordinates of the feature points will change with the change of the standing posture of the human body. For example, actions such as shrugging, bending over, or twisting the hips of the human body will make the fitted epidermal spine curves show different shapes.

Therefore, in order to reduce the difficulty of feature point prediction, when collecting the depth image of the subject’s back, it is necessary for the subject to stand with his arms drooping naturally; since this application involves depth images, it is more sensitive to the shooting distance, and the subject needs to stand At a distance of 70cm±5cm from the front of the camera, and keep the back plane parallel to the camera shooting plane.

In the embodiment of the present application, by acquiring the depth image of the back of the subject, the target point cloud data is determined according to the depth image of the back, and the spine of the subject is determined in the depth image of the back according to the point cloud data of the target. The preset number of feature point coordinates in; Curve fitting is performed according to all the feature point coordinates to generate the target epidermal spine curve; the target scoliosis angle is determined according to the target epidermal spine curve, and the target scoliosis angle is determined according to the target scoliosis The angle value of the angle determines whether the subject’s spine is scoliosis; by determining the subject’s spine feature points and coordinates in the point cloud data of the back depth image, and then performing curve fitting on all these feature point coordinates , determine the target scoliosis angle directly from the target epidermal spine curve generated by fitting, so as to judge whether the subject’s spine is scoliosis according to the target scoliosis angle; The requirements are low, and the result of scoliosis can be obtained quickly and efficiently through deep learning and curve fitting. It is suitable for scoliosis detection of large groups, especially primary and middle school students.

Next, the above-mentioned image-based method for detecting the epidermal spine line of the back of the human body will be further described through the following embodiments.

As described in step S110, the depth image of the back of the subject is acquired, and the target point cloud data is determined according to the depth image of the back.

In an embodiment of the present invention, the specific process of "determining target point cloud data based on the back depth image" in step S110 can be further described in conjunction with the following description.

As described in the following steps:

collecting a depth image of the subject's back through a depth camera, and acquiring the depth image of the back from the depth camera;

Acquiring the calibrated internal parameters of the depth image, and determining the three-dimensional point cloud coordinates corresponding to each pixel in the back depth image according to the internal parameters,

The three-dimensional point cloud data is determined according to all the three-dimensional coordinates of the point cloud, and a preset filtering algorithm is used to perform noise reduction processing on the three-dimensional point cloud data to obtain the target point cloud data.

As an example, a Kinect depth camera may be used to collect a depth image of the subject when the subject is exposed, and the Kinect depth camera may be calibrated before the image is collected to obtain internal parameters of the camera. The subject's arms hang down naturally, standing 70cm in front of the camera, and the lower corner of the subject's scapula corresponds to the level of the depth camera.

Because the 3D point cloud data is affected by environmental conditions and camera quality, there are often more noise points, especially at the edge of the human body contour; the above filtering algorithm can use bilateral filtering algorithm, which can combine distance and spatial structure to denoise, and the effect is better .

As described in step S120, a preset number of feature point coordinates of the subject's spine in the back depth image are determined according to the target point cloud data.

In an embodiment of the present invention, the specific process of "determining a preset number of feature point coordinates of the subject's spine in the back depth image" in step S120 can be further described in conjunction with the following description.

As described in the following steps:

Invoking a preset stacked hourglass network model, inputting the target point cloud data into the stacked hourglass network model, and training the stacked hourglass network model;

Outputting the preset number of feature point coordinates through the trained stacked hourglass network model.

In a specific implementation, during the training process of the stacked hourglass network model, the effect of the model is verified through a test set. After the stacked hourglass network model is trained, it can output the abscissa and ordinate of the feature points of 20 epidermal spine lines, a total of 40 data.

It should be noted that the above-mentioned stacked hourglass network model (Stacked Hourglass Networks) is a kind of deep neural network model that is more important in pose estimation in the field of computer vision. In the case of attenuation and gradient disappearance, the Kaiming network weight initialization method can be used to initialize the network weights in the stacked hourglass network model.

As an example, after calling the preset stacked hourglass network model, it includes:

Determining the number of output neurons of each layer in the stacked hourglass network model and the scale coefficient after weight normalization;

According to the number of neurons and the proportional coefficient, weights are initialized in the stacked hourglass network model through a Kaiming network weight initialization method.

Wherein, the calculation formula for weight initialization in the stacked hourglass network model is as follows:

Among them, W _ij is the weight of the network, n is the number of output neurons in each layer, and U is the proportional coefficient after the normalization of the weight.

Through the above calculation formula, the gradient disappearance of the stacked hourglass network model is solved, the variance decay of the activation value is avoided, and the activation value of each layer is kept in a Gaussian distribution.

As described in step S130, curve fitting is performed according to all the feature point coordinates to generate a target epidermal spine curve.

In an embodiment of the present invention, the specific process of "carrying out curve fitting according to all the feature point coordinates to generate the target epidermal spine curve" in step S130 can be further described in conjunction with the following description.

As described in the following steps, the interparc curve interpolation method is used to perform polynomial interpolation fitting processing on all the feature point coordinates to generate the target epidermal spine curve.

As an example, the above-mentioned 20 feature point coordinates of the epidermal spine lines are subjected to multinomial interpolation fitting processing, and the calculation formula is as follows:

Among them, x is the coordinate of the feature point, and w is the polynomial coefficient.

In the embodiment of the present application, the interparc curve interpolation method is used to perform polynomial interpolation fitting processing on all the feature point coordinates, including:

Obtain the initial epidermal spine curve according to the Interparc curve interpolation method and all the characteristic point coordinate fittings;

Determining the length value of the initial epidermal spine curve, and equally dividing the initial epidermal spine curve according to the length value;

Perform linear interpolation processing on the initial epidermal spine curve after the equalization process to obtain the coordinates of all target feature points;

According to all the target feature point coordinates and the Interparc curve interpolation method, the target epidermis spine curve is generated by fitting.

As an example, the calculation formula for equalizing the initial epidermal spine curve is as follows:

As described in step S140, the target scoliosis angle is determined according to the target epidermal spine curve, and whether the subject's spine is scoliosis is determined according to the angle value of the target scoliosis angle.

In an embodiment of the present invention, the step S140 of "determining the target scoliosis angle according to the target epidermal spine curve, and determining the subject's scoliosis angle according to the angle value of the target scoliosis angle" in step S140 can be further described in conjunction with the following description. The specific process of whether your spine is scoliosis".

As described in the following steps,

determining the target normal vector according to the target skin spine curve;

determining a target scoliosis angle according to the target normal vector;

When the angle value of the target scoliosis angle is less than the preset scoliosis angle value, it means that the detected person’s spine is normal; when the angle value of the target scoliosis angle is greater than or equal to the preset scoliosis angle value When the angle value is smaller, it means that the detected person's spine is scoliosis.

As an example, the formula for determining the target scoliosis angle from the target normal vector is as follows:

Among them, v _i and v _j are target normal vectors, and θ is the angle of the target scoliosis angle.

It should be noted that the target scoliosis angle may be the Cobb angle, and its preset value of the scoliosis angle may be 12°. When the Cobb angle is less than 12°, it means that the subject's spine is normal; when the Cobb angle is greater than or equal to 12°, it means that the subject's spine is scoliosis.

As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for related parts, please refer to the part of the description of the method embodiment.

Referring to FIG. 2 , it shows a schematic structural diagram of an image-based detection device for epidermal spine lines on the back of a human body provided by an embodiment of the present application;

The devices include:

The point cloud determination module 110 is used to obtain the depth image of the back of the detected person, and determine the target point cloud data according to the depth image of the back;

A coordinate determination module 120, configured to determine a preset number of feature point coordinates of the subject's spine in the back depth image according to the target point cloud data;

The curve generation module 130 is used to perform curve fitting according to all the feature point coordinates to generate the target epidermal spine curve;

The scoliosis judging module 140 is configured to determine a target scoliosis angle according to the target epidermal spine curve, and determine whether the subject's spine is scoliosis according to the angle value of the target scoliosis angle.

With reference to Fig. 3, the structural representation of the computer equipment of the detection method of a kind of image-based human back skin spine line of the present invention is shown, specifically can comprise as follows:

The above-mentioned computer device 12 is represented in the form of a general-purpose computing device. The components of the computer device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and various system components (including system memory 28 and processing unit) connected to each other. 16) The bus 18.

The bus 18 represents one or more of several types of bus 18 structures, including a memory bus 18 or memory controller, a peripheral bus 18, an accelerated graphics port, a processor, or a bureau using any of a variety of bus 18 structures. domain bus 18. By way of example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus 18, the Micro Channel Architecture (MAC) bus 18, the Enhanced ISA bus 18, the Video Electronics Standards Association (VESA) local bus 18, and Peripheral Component Interconnect (PCI) bus 18 .

Computer device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by computer device 12 and include both volatile and nonvolatile media, removable and non-removable media.

System memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and/or cache memory 32 . The computer device 12 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from and write to non-removable, non-volatile magnetic media (commonly referred to as a "hard drive"). Although not shown in FIG. 3, a disk drive for reading and writing to removable non-volatile disks (such as "floppy disks") may be provided, as well as for removable non-volatile optical disks (such as CD-ROM, DVD-ROM or other optical media) CD-ROM drive. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. The memory may include at least one program product having a set (eg, at least one) of program modules 42 configured to perform the functions of various embodiments of the present invention.

program/utility 40 having a set (at least one) of program modules 42, such as may be stored in memory, such program modules 42 including - but not limited to - an operating system, one or more application programs, other program modules 42 and program data, each or some combination of these examples may include the implementation of the network environment. Program modules 42 generally perform the functions and/or methodologies of the described embodiments of the invention.

The computer device 12 may also communicate with one or more external devices 14 (e.g., a keyboard, pointing device, display 24, camera, etc.), and with one or more devices that enable a user to interact with the computer device 12, and/or Or communicate with any device (eg, network card, modem, etc.) that enables the computing device 12 to communicate with one or more other computing devices. Such communication may occur through input/output (I/O) interface 22 . Also, computer device 12 may communicate with one or more networks (eg, local area network (LAN)), wide area network (WAN) and/or public networks (eg, the Internet) via network adapter 20 . As shown, network adapter 20 communicates with other modules of computer device 12 via bus 18 . It should be appreciated that although not shown in FIG. 3 , other hardware and/or software modules may be used in conjunction with computer device 12, including but not limited to: microcode, device drivers, redundant processing units 16, external disk drive arrays, RAID systems, Tape drives and data backup storage systems 34 and the like.

The processing unit 16 executes various functional applications and data processing by running the programs stored in the system memory 28 , for example, implementing an image-based method for detecting the epidermal spine line of the back of the human body provided by the embodiment of the present invention.

That is to say, when the above-mentioned processing unit 16 executes the above-mentioned program, it realizes: acquiring the back depth image of the subject, determining the target point cloud data according to the back depth image; The preset number of feature point coordinates in the back depth image; curve fitting is performed according to all the feature point coordinates to generate a target epidermal spine curve; the target scoliosis angle is determined according to the target epidermal spine curve, and the target scoliosis angle is determined according to the target epidermal spine curve. The angle value of the target scoliosis angle determines whether the subject's spine is scoliotic.

In an embodiment of the present invention, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, an image-based human back epidermis as provided in all embodiments of the present application is realized. Detection method of spine line:

That is to say, when the program is executed by the processor, it is realized: acquire the back depth image of the subject, determine the target point cloud data according to the back depth image; determine the position of the subject's spine according to the target point cloud data. The preset number of feature point coordinates in the back depth image; curve fitting is performed according to all the feature point coordinates to generate the target epidermal spine curve; the target scoliosis angle is determined according to the target epidermal spine curve, and according to the The angle value of the target scoliosis angle determines whether the subject's spine is scoliotic or not.

Any combination of one or more computer readable medium(s) may be utilized. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (non-exhaustive list) of computer readable storage media include: electrical connections with one or more leads, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), Erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

A computer readable signal medium may include a data signal carrying computer readable program code in baseband or as part of a carrier wave. Such propagated data signals may take many forms, including - but not limited to - electromagnetic signals, optical signals, or any suitable combination of the foregoing. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. .

Computer program code for carrying out the operations of the present invention may be written in one or more programming languages, or combinations thereof, including object-oriented programming languages—such as Java, Smalltalk, C++, and conventional Procedural programming language - such as "C" or a similar programming language. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (for example, using an Internet service provider to connected via the Internet). Each embodiment in this specification is described in a progressive manner, each embodiment focuses on the difference from other embodiments, and the same and similar parts of each embodiment can be referred to each other.

While the preferred embodiments of the embodiments of the present application have been described, additional changes and modifications can be made to these embodiments by those skilled in the art once the basic inventive concept is understood. Therefore, the appended claims are intended to be interpreted to cover the preferred embodiment and all changes and modifications that fall within the scope of the embodiments of the application.

Finally, it should also be noted that in this text, relational terms such as first and second etc. are only used to distinguish one entity or operation from another, and do not necessarily require or imply that these entities or operations, any such actual relationship or order exists. Furthermore, the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article, or terminal equipment comprising a set of elements includes not only those elements, but also includes elements not expressly listed. other elements identified, or also include elements inherent in such a process, method, article, or terminal equipment. Without further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article or terminal device comprising said element.

The above is a detailed introduction of the image-based detection method and device for the epidermal spine line of the human body back provided by the application. In this paper, specific examples are used to illustrate the principle and implementation of the application. The description of the above embodiments It is only used to help understand the method of the present application and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation and application scope. In summary, The contents of this specification should not be understood as limiting the application.

Claims (10)

1. An image-based detection method for human back epidermis spinal lines, which is characterized by comprising the following steps:

acquiring a back depth image of a detected person, and determining target point cloud data according to the back depth image;

determining the preset number of characteristic point coordinates of the spine of the detected person in the back depth image according to the target point cloud data;

performing curve fitting according to all the characteristic point coordinates to generate a target epidermis spine curve;

and determining a target scoliosis angle according to the target epidermis spine curve, and determining whether the spine of the tested person is bent sideways according to the angle value of the target scoliosis angle.

2. The method of claim 1, wherein the acquiring the back depth image of the subject, determining target point cloud data from the back depth image, comprises:

acquiring a back depth image of the detected person through a depth camera, and acquiring the back depth image from the depth camera;

obtaining calibrated internal parameters of the depth image, determining the three-dimensional coordinates of the point cloud corresponding to each pixel point in the back depth image according to the internal parameters,

and determining three-dimensional point cloud data according to all the three-dimensional coordinates of the point cloud, and carrying out noise reduction processing on the three-dimensional point cloud data by using a preset filtering algorithm to obtain the target point cloud data.

3. The method of claim 1, wherein the determining a preset number of feature point coordinates of the subject's spine in the back depth image from the target point cloud data comprises:

invoking a preset stacking hourglass network model, inputting the target point cloud data into the stacking hourglass network model, and training the stacking hourglass network model;

and outputting the preset number of feature point coordinates through the trained stacked hourglass network model.

4. The method of claim 3, wherein the invoking the pre-set stacked hourglass network model, thereafter comprises:

determining the number of output neurons of each layer in the stacked hourglass network model and a proportionality coefficient after weight normalization;

and initializing the weight in the stacked hourglass network model by using a Kaiming network weight initialization method according to the number of the neurons and the proportionality coefficient.

5. The method of claim 1, wherein said generating a target epidermis spine curve by curve fitting based on all of said feature point coordinates comprises:

and performing polynomial interpolation fitting treatment on all the characteristic point coordinates by adopting an intersarc curve interpolation method to generate the target epidermis spine curve.

6. The method of claim 5, wherein performing polynomial interpolation fitting on all the feature point coordinates using an interpolation curve method comprises:

obtaining an initial epidermis spine curve according to the intersarc curve interpolation method and the coordinate fitting of all the characteristic points;

determining a length value of the initial epidermis spinal curve, and equally dividing the initial epidermis spinal curve according to the length value;

performing linear interpolation on the initial epidermis spine curve subjected to the equipartition treatment to obtain all target characteristic point coordinates;

and fitting and generating the target epidermis spine curve according to all the target characteristic point coordinates and an interpolation curve interpolation method.

7. The method of claim 1, wherein determining a target scoliosis angle from the target epidermis spine curve, determining whether the subject's spine is scoliotic from an angle value of the target scoliosis angle, comprises:

determining a target normal vector according to the target epidermis spine curve;

determining a target scoliosis angle according to the target normal vector;

when the angle value of the target spine side bending angle is smaller than a preset side bending angle value, the spine of the detected person is normal; and when the angle value of the target scoliosis angle is larger than or equal to a preset scoliosis angle value, representing that the spine of the detected person is scoliosis.

8. An image-based detection device for human back epidermis spinal cord, the device comprising:

the point cloud determining module is used for acquiring a back depth image of the detected person and determining target point cloud data according to the back depth image;

the coordinate determining module is used for determining the preset number of characteristic point coordinates of the spine of the detected person in the back depth image according to the target point cloud data;

the curve generation module is used for performing curve fitting according to all the characteristic point coordinates to generate a target epidermis spine curve;

and the scoliosis judging module is used for determining a target scoliosis angle according to the target epidermis spine curve and determining whether the spine of the tested person is scoliosis according to the angle value of the target scoliosis angle.

9. An apparatus comprising a processor, a memory, and a computer program stored on the memory and capable of running on the processor, which when executed by the processor, implements the method of any one of claims 1 to 7.

10. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the method according to any of claims 1 to 7.

Images