A method and device for detecting a skin spine line on a back of a human body based on an image

Abstract

The application relates to the field of scoliosis detection, in particular to a human back epidermis spine line detection method and device based on images, which comprises the following steps: acquiring a back depth image of a detected person, determining target point cloud data according to the back depth image; determining a preset number of feature 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 feature point coordinates to generate a target epidermis spine curve; determining a target scoliosis angle according to the target epidermis spine curve, and determining whether the spine of the detected person is scoliotic according to the angle value of the target scoliosis angle. The method is free of radiation throughout the whole process, has low proficiency requirement for the operator, can quickly and efficiently obtain the scoliosis result through depth learning and curve fitting, and is suitable for scoliosis detection of a large group, especially a primary and secondary school student group.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of scoliosis detection, and in particular to a method and device for detecting a human back epidermal spine line based on images. BACKGROUND

[0002] The human back epidermal spine line is an important representation for determining the scoliosis condition of the human back, and scoliosis is a three-dimensional deformity of the spine, including sequence abnormalities in the coronal, sagittal and axial positions, and is more common in children and adolescents. About 2-4% of adolescents worldwide suffer from scoliosis, and the number of scoliosis patients in China has exceeded 3 million, with an annual increase of 300,000 people. The national policy clearly indicates that attention and protection of the healthy development of primary and secondary school students' spines will be paid, and scoliosis examination will be included in the physical examination of primary and secondary school students. Therefore, the efficient detection of the human back epidermal spine line using artificial intelligence technology and the judgment of the scoliosis condition are imminent.

[0003] Currently, the detection of scoliosis in the physical examination of primary and secondary school students is mainly through professional detection personnel using a scoliosis measuring instrument (scoliosis ruler) to screen for scoliosis. This method is low in efficiency and requires high skill and proficiency of the detection personnel, so the results measured by different personnel with different proficiency levels have large deviations. The current domestic and foreign detection methods for scoliosis mainly include X-ray, EOS, ultrasound, and cloud pattern.

[0004] X-ray is low in cost, simple and easy to operate, and accurate in measurement, but has a large radiation dose and causes great harm to the human body of the target primary and secondary school population; EOS is high in cost and still has a small amount of radiation; ultrasound requires high skill of the doctor and has large errors and low efficiency, which is not conducive to large-scale screening; and cloud pattern can only roughly determine the condition of the target. SUMMARY

[0005] In view of the problems, the present application is proposed to provide a method and device for detecting a human back epidermal spine line based on images to overcome the problems or at least partially solve the problems.

[0006] A method for detecting a human back epidermal spine line based on images, the method comprising:

[0007] obtaining a back depth image of a detected person, determining target point cloud data according to the back depth image;

[0008] determining a preset number of feature point coordinates of the spine of the detected person in the back depth image according to the target point cloud data;

[0009] performing curve fitting according to all the feature point coordinates to generate a target epidermal spine curve;

[0010] Determine a target scoliosis angle according to the target skin surface spine curve, and determine whether the spine of the detected person is scoliotic according to an angle value of the target scoliosis angle.

[0011] Preferably, the back depth image of the detected person is acquired, and target point cloud data is determined according to the back depth image, including:

[0012] The back depth image of the detected person is acquired by a depth camera, and the back depth image is acquired from the depth camera;

[0013] Calibrated internal parameters of the depth image are acquired, and a point cloud three-dimensional coordinate corresponding to each pixel point in the back depth image is determined according to the internal parameters,

[0014] Three-dimensional point cloud data is determined according to all the point cloud three-dimensional coordinates, and the three-dimensional point cloud data is denoised by using a preset filtering algorithm to obtain the target point cloud data.

[0015] Preferably, the target point cloud data is used to determine a preset number of feature point coordinates of the spine of the detected person in the back depth image, including:

[0016] A preset stacked hourglass network model is called, the target point cloud data is input into the stacked hourglass network model, and the stacked hourglass network model is trained;

[0017] The preset number of feature point coordinates is output by the trained stacked hourglass network model.

[0018] Preferably, the preset stacked hourglass network model is called, and then includes:

[0019] The number of output neurons and the proportion coefficient of weight normalization in each layer of the stacked hourglass network model are determined;

[0020] According to the number of neurons and the proportion coefficient, the weight initialization of the stacked hourglass network model is performed by using a Kaiming network weight initialization method.

[0021] Preferably, the curve fitting is performed according to all the feature point coordinates to generate a target skin surface spine curve, including:

[0022] All the feature point coordinates are processed by polynomial interpolation fitting by using an Interparc curve interpolation method to generate the target skin surface spine curve.

[0023] Preferably, the polynomial interpolation fitting processing of all the feature point coordinates by using the Interparc curve interpolation method includes:

[0024] fitting an initial epidermal spinal curve according to the Interparc curve interpolation method and all the feature point coordinates;

[0025] determining a length value of the initial epidermal spinal curve, and performing equal division processing on the initial epidermal spinal curve according to the length value;

[0026] performing linear interpolation processing on the initial epidermal spinal curve after the equal division processing, to obtain all target feature point coordinates;

[0027] generating the target epidermal spinal curve according to all the target feature point coordinates and the Interparc curve interpolation method.

[0028] Preferably, the determining a target spinal scoliosis angle according to the target epidermal spinal curve, and determining whether the spine of the detected person is scoliotic according to an angle value of the target spinal scoliosis angle, comprises:

[0029] determining a target normal vector according to the target epidermal spinal curve;

[0030] determining a target spinal scoliosis angle according to the target normal vector;

[0031] when the angle value of the target spinal scoliosis angle is less than a preset scoliosis angle value, it indicates that the spine of the detected person is normal; when the angle value of the target spinal scoliosis angle is greater than or equal to the preset scoliosis angle value, it indicates that the spine of the detected person is scoliotic.

[0032] Also provided is a human back epidermal spinal line detection device based on images, the device comprising:

[0033] a point cloud determination module, configured to acquire a back depth image of a detected person, and determine target point cloud data according to the back depth image;

[0034] a coordinate determination module, configured to determine a preset number of feature point coordinates of the spine of the detected person in the back depth image according to the target point cloud data;

[0035] a curve generation module, configured to perform curve fitting according to all the feature point coordinates, to generate a target epidermal spinal curve;

[0036] a scoliosis judgment module, configured to determine a target spinal scoliosis angle according to the target epidermal spinal curve, and determine whether the spine of the detected person is scoliotic according to an angle value of the target spinal scoliosis angle.

[0037] An apparatus comprising a processor, a memory, and a computer program stored on the memory and executable on the processor, the computer program, when executed by the processor, implementing the steps of the method for detecting a skin spine line of a back of a human body based on an image as described above.

[0038] A computer readable storage medium having stored thereon a computer program, the computer program, when executed by a processor, implementing the steps of the method for detecting a skin spine line of a back of a human body based on an image as described above.

[0039] The present application has the following advantages:

[0040] In the embodiment of the present application, by acquiring a back depth image of a detected person, determining target point cloud data according to the back depth image, determining a preset number of feature point coordinates of a spine of the detected person in the back depth image according to the target point cloud data; performing curve fitting according to all the feature point coordinates to generate a target skin spine curve; determining a target scoliosis angle according to the target skin spine curve, and determining whether the spine of the detected person is scoliotic according to an angle value of the target scoliosis angle; by determining the spine feature points and coordinates of the detected person in the point cloud data of the back depth image, and then performing curve fitting on all the feature point coordinates, the target scoliosis angle is determined from the target skin spine curve generated by fitting, so as to determine whether the spine of the detected person is scoliotic according to the target scoliosis angle; the above scheme is radiation-free throughout, has low proficiency requirements for the operator, and can quickly and efficiently obtain the result of whether the spine is scoliotic through depth learning and curve fitting, and is suitable for scoliosis detection of large groups, especially primary and secondary school students. BRIEF DESCRIPTION OF DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the present application, the drawings needed to be used in the description of the present application will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0042] Figure 1 A step flowchart of a method for detecting a skin spine line of a back of a human body based on an image provided by an embodiment of the present application is shown;

[0043] Figure 2 A structure schematic diagram of a device for detecting a skin spine line of a back of a human body based on an image provided by an embodiment of the present application is shown;

[0044] Figure 3A structural schematic diagram of a computer device for a human back skin spine line detection method based on images is shown. DETAILED DESCRIPTION

[0045] In order to make the objects, characteristics and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below in combination with the drawings and specific embodiments. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0046] In view of the high incidence of scoliosis in primary and secondary school students at present, the existing detection methods such as X-ray, EOS, ultrasound and cloud pattern have application limitations, the present application proposes a method for analyzing the scoliosis of the detected spine based on computer vision images combined with a deep learning model, so as to achieve the purpose of zero radiation, low error and high efficiency of scoliosis detection.

[0047] Please refer to Figure 1 , a step flowchart of a human back skin spine line detection method based on images is shown. The method comprises the following steps:

[0048] S110, acquiring a back depth image of a detected person, determining target point cloud data according to the back depth image;

[0049] S120, determining a preset number of feature point coordinates of the spine of the detected person in the back depth image according to the target point cloud data;

[0050] S130, performing curve fitting according to all the feature point coordinates to generate a target skin spine curve;

[0051] S140, determining a target scoliosis angle of the target spine according to the target skin spine curve, and determining whether the spine of the detected person is scoliotic according to the angle value of the target scoliosis angle.

[0052] It should be noted that the image features of the human back have diversity, and the feature point coordinates will change with the change of the standing posture of the human body. For example, the actions such as hunched shoulders, bent waist or twisted hips will make the fitted skin spine curve present different shapes.

[0053] Therefore, in order to reduce the difficulty of feature point prediction, when collecting the back depth image of the detected person, the detected person needs to stand with the arms naturally drooping; since the present application involves depth images, the shooting distance is relatively sensitive, and the detected person also needs to stand at a distance of 70cm±5cm in front of the camera, and keep the back plane parallel to the camera shooting plane.

[0054] In the embodiment of the present application, the back depth image of the detected person is acquired, the target point cloud data is determined according to the back depth image, the preset number of feature point coordinates of the spine of the detected person in the back depth image is determined according to the target point cloud data, curve fitting is performed according to all the feature point coordinates, the target epidermal spine curve is generated, the target scoliosis angle is determined according to the target epidermal spine curve, and whether the spine of the detected person is scoliotic is determined according to the angle value of the target scoliosis angle. By determining the spine feature points and coordinates of the detected person in the point cloud data of the back depth image, and then performing curve fitting on all the feature point coordinates, the target scoliosis angle is determined from the target epidermal spine curve generated by fitting, so that whether the spine of the detected person is scoliotic is determined according to the target scoliosis angle. The above scheme is radiation-free throughout, has low proficiency requirements for the operator, and can quickly and efficiently obtain the result of whether the spine is scoliotic through depth learning and curve fitting, and is suitable for detecting the scoliosis of a large group of people, especially primary and secondary school students.

[0055] In the following, the above-mentioned image-based human back epidermal spine line detection method will be further described through the following embodiments.

[0056] As described in step S110, the back depth image of the detected person is acquired, and the target point cloud data is determined according to the back depth image.

[0057] In an embodiment of the present application, the specific process of “determining the target point cloud data according to the back depth image” described in step S110 can be further described in combination with the following description.

[0058] As described in the following steps:

[0059] The back depth image of the detected person is acquired by a depth camera, and the back depth image is acquired from the depth camera;

[0060] The calibrated internal parameters of the depth image are acquired, and the point cloud three-dimensional coordinates corresponding to each pixel point in the back depth image are determined according to the internal parameters,

[0061] The three-dimensional point cloud data is determined according to all the point cloud three-dimensional coordinates, the three-dimensional point cloud data is denoised by using a preset filtering algorithm, and the target point cloud data is obtained.

[0062] As an example, a Kinect depth camera can be used to acquire the depth image of the detected person when the back is bare, the Kinect depth camera is calibrated before the image is acquired, and the internal parameters of the camera are obtained. The detected person stands at 70 cm in front of the camera with the arms naturally hanging down, and the lower corner of the scapula of the detected person is horizontally corresponding to the depth camera.

[0063] Due to the three-dimensional point cloud data is affected by environmental conditions and camera quality, there are often many noise points, more significant 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.

[0064] As described in step S120, according to the target point cloud data, the preset number of feature point coordinates of the spine of the detected person in the back depth image are determined.

[0065] In an embodiment of the present application, the specific process of "determining the preset number of feature point coordinates of the spine of the detected person in the back depth image" described in step S120 can be further described in combination with the following description.

[0066] As described in the following steps:

[0067] The preset stacked hourglass network model is called, the target point cloud data is input into the stacked hourglass network model, and the stacked hourglass network model is trained;

[0068] The preset number of feature point coordinates is output by the trained stacked hourglass network model.

[0069] In a specific implementation, in the training process of the stacked hourglass network model, the model effect is verified by a test set. When the stacked hourglass network model is trained, 20 feature point coordinates of the epidermal spine line can be output, including horizontal coordinates and vertical coordinates, a total of 40 data.

[0070] It should be noted that the stacked hourglass network model (Stacked Hourglass Networks) described above is a kind of deep neural network model applied in pose estimation in the field of computer vision. In order to prevent the variance of activation value from decaying and the gradient from disappearing in the forward propagation process of the network, the Kaiming network weight initialization method can be used to initialize the network weight in the stacked hourglass network model.

[0071] As an example, after the preset stacked hourglass network model is called, the following steps are included:

[0072] The number of output neurons and the proportion coefficient after weight normalization in each layer of the stacked hourglass network model are determined;

[0073] According to the number of neurons and the proportion coefficient, the Kaiming network weight initialization method is used to initialize the weight in the stacked hourglass network model.

[0074] The calculation formula for initializing the weight in the stacked hourglass network model is as follows:

[0075]

[0076] wherein, W ij is the network weight, n is the number of output neurons of each layer, and U is the proportionality coefficient after weight normalization.

[0077] Through the above calculation formula, the gradient vanishing of the stacked hourglass network model is solved, the variance attenuation of the activation value is avoided, and the Gaussian distribution of the activation value of each layer is maintained.

[0078] As described in step S130, curve fitting is performed according to all the feature point coordinates to generate a target epidermal ridge curve.

[0079] In an embodiment of the present application, the specific process of "curve fitting according to all the feature point coordinates to generate a target epidermal ridge curve" described in step S130 can be further illustrated in combination with the following description.

[0080] 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 ridge curve.

[0081] As an example, the 20 feature point coordinates of the epidermal ridge line are subjected to polynomial interpolation fitting processing, and the calculation formula is as follows:

[0082]

[0083] wherein, x is the feature point coordinate, and w is the polynomial coefficient.

[0084] In the embodiments of the present application, the Interparc curve interpolation method is used to perform polynomial interpolation fitting processing on all the feature point coordinates, including:

[0085] An initial epidermal ridge curve is fitted according to the Interparc curve interpolation method and all the feature point coordinates;

[0086] The length value of the initial epidermal ridge curve is determined, and the initial epidermal ridge curve is evenly divided according to the length value;

[0087] Linear interpolation processing is performed on the initial epidermal ridge curve after the division processing to obtain all the target feature point coordinates;

[0088] The target epidermal ridge curve is fitted and generated according to all the target feature point coordinates and the Interparc curve interpolation method.

[0089] As an example, the calculation formula for the division processing of the initial epidermal ridge curve is as follows:

[0090]

[0091] As described in step S140, a target scoliosis angle is determined according to the target epidermal spine curve, and whether the spine of the detected person is scoliotic is determined according to an angle value of the target scoliosis angle.

[0092] In an embodiment of the present application, the specific process of "determining a target scoliosis angle according to the target epidermal spine curve, and determining whether the spine of the detected person is scoliotic according to an angle value of the target scoliosis angle" described in step S140 can be further illustrated in combination with the following description.

[0093] As described in the following steps,

[0094] a target normal vector is determined according to the target epidermal spine curve;

[0095] a target scoliosis angle is determined according to the target normal vector;

[0096] When the angle value of the target scoliosis angle is less than a preset scoliosis angle value, it indicates that the spine of the detected person is normal; when the angle value of the target scoliosis angle is greater than or equal to the preset scoliosis angle value, it indicates that the spine of the detected person is scoliotic.

[0097] As an example, the calculation formula of the target scoliosis angle determined by the target normal vector is as follows:

[0098]

[0099] Wherein, v i ,v j is the target normal vector, and θ is the angle of the target scoliosis angle.

[0100] It should be noted that the target scoliosis angle can be a Cobb angle, and the preset scoliosis angle value can be 12°. When the Cobb angle is less than 12°, it indicates that the spine of the detected person is normal; when the Cobb angle is greater than or equal to 12°, it indicates that the spine of the detected person is scoliotic.

[0101] For the device embodiment, it is basically similar to the method embodiment, so the description is relatively simple, and the related parts can be referred to the part of the method embodiment.

[0102] Referring to Figure 2 , a structure schematic diagram of a human back epidermal spine line detection device based on an image is shown, which is provided by an embodiment of the present application;

[0103] The device comprises:

[0104] The point cloud determination module 110 is configured to acquire a back depth image of a detected person, and determine target point cloud data according to the back depth image.

[0105] The coordinate determination module 120 is configured to determine a preset number of feature point coordinates of a spine of the detected person in the back depth image according to the target point cloud data.

[0106] The curve generation module 130 is configured to perform curve fitting according to all the feature point coordinates, and generate a target epidermal spine curve.

[0107] The scoliosis judgment module 140 is configured to determine a target scoliosis angle according to the target epidermal spine curve, and determine whether the spine of the detected person is scoliotic according to an angle value of the target scoliosis angle.

[0108] Referring to Figure 3 , a structure schematic diagram of a computer device of a human back epidermal spine line detection method based on images is shown, which can specifically include the following:

[0109] The computer device 12 is in the form of a general-purpose computing device, and the components of the computer device 12 can include but are not limited to one or more processors or processing units 16, a system memory 28, and a bus 18 connecting different system components including the system memory 28 and the processing unit 16.

[0110] The bus 18 represents one or more of several types of buses 18, including a memory bus 18 or memory controller, a peripheral bus 18, a graphics acceleration port, a processor or local bus 18 using any of a variety of bus 18 architectures including an industry standard architecture (ISA) bus 18, a micro-channel architecture (MAC) bus 18, an enhanced ISA bus 18, a video electronics standards association (VESA) local bus 18, and a peripheral component interconnect (PCI) bus 18. For example, these architectures include but are not limited to an industry standard architecture (ISA) bus 18, a micro-channel architecture (MAC) bus 18, an enhanced ISA bus 18, a video electronics standards association (VESA) local bus 18, and a peripheral component interconnect (PCI) bus 18.

[0111] The computer device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computer device 12, including both volatile and nonvolatile media, removable and non-removable media.

[0112] The system memory 28 can 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 can further include other mobile / non-mobile, volatile / non-volatile computer system storage media. By way of example only, a storage system 34 can be used to read from and write to a non-removable, non-volatile magnetic media (typically called a "hard drive"). Although not specifically shown, the computer device 12 can further include other removable / non-removable, volatile / non-volatile computer system storage media such as a magnetic disk drive, a solid state drive, a Blu-ray drive, and the like.Figure 3 In some embodiments, a disk drive can be provided for reading from or writing to a removable nonvolatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive can be provided for reading from or writing to a removable nonvolatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media). In these cases, each drive can be connected to the bus 18 by one or more data media interfaces. The memory can include, at a minimum, one program product having a set (e.g., at least one) of program modules 42 configured to carry out the functions of embodiments of the application.

[0113] The program / utility 40, having a set (at least one) of program modules 42, can be stored in memory, for example, RAM, ROM, EEPROM, flash or other non-volatile memory. Each of the operating system, one or more application programs, other program modules 42, and program data, or some combination thereof, can include an implementation of a networking environment. The program modules 42 generally carry out the functions and / or methodologies of embodiments of the application as described herein.

[0114] The computer device 12 can also communicate with one or more external devices 14 such as a keyboard, a pointing device, a display 24, a camera, etc.; one or more devices that enable a user to interact with the computer device 12; and / or one or more devices that enable the computer device 12 to communicate with one or more other computer devices. Such communication can be via an input / output (I / O) interface 22. Further, the computer device 12 can communicate with one or more networks such as a local area network (LAN), a general wide area network (WAN), and / or a public network (e.g., the Internet) via a network adapter 20. As shown, the network adapter 20 communicates with the other components of the computer device 12 via the bus 18. It should be appreciated that the network adapter 20 can also be connected to one or more components of the computer device 12 in some other fashion such as over a wireless connection, etc. Figure 3 Other hardware and / or software modules can be used in conjunction with the 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 archival storage systems 34, etc.

[0115] The processing unit 16 executes the various functions and data processing operations by running programs stored in the system memory 28, such as to implement an image-based detection method of a human back surface skin spine line according to embodiments of the application.

[0116] That is, the processing unit 16 implements the following when executing the program: obtaining a back depth image of a detected person, determining target point cloud data according to the back depth image; determining a preset number of feature point coordinates of a spine of the detected person in the back depth image according to the target point cloud data; performing curve fitting according to all the feature point coordinates to generate a target epidermal spine curve; determining a target scoliosis angle according to the target epidermal spine curve, and determining whether the spine of the detected person is scoliotic according to an angle value of the target scoliosis angle.

[0117] In the embodiments of the present application, the present application also provides a computer readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the image-based human back epidermal spine line detection method provided in all embodiments of the present application.

[0118] That is, the processing unit 16 implements the following when executing the program: obtaining a back depth image of a detected person, determining target point cloud data according to the back depth image; determining a preset number of feature point coordinates of a spine of the detected person in the back depth image according to the target point cloud data; performing curve fitting according to all the feature point coordinates to generate a target epidermal spine curve; determining a target scoliosis angle according to the target epidermal spine curve, and determining whether the spine of the detected person is scoliotic according to an angle value of the target scoliosis angle.

[0119] Any combination of one or more computer readable medium can be employed. The computer readable medium can be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium include an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In this document, the computer readable storage medium can be any tangible medium that contains, or stores a program for use by or in connection with an instruction execution system, apparatus, or device.

[0120] A computer readable signal medium can include a propagated data signal with computer executable code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated signal can take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium can be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate or transport programming code.

[0121] Computer program code for carrying out operations of the present application can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can 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 the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider). The embodiments of the present application, as described in the specification, exemplify the application and are not meant to be limiting, but rather a variety of changes, modifications or alterations can be made thereto by those skilled in the art without departing from the scope of the present application.

[0122] While the preferred embodiments of the methods and systems of the present application have been described, additional variations and modifications can be made by those skilled in the art once they have the benefit of the present disclosure. Therefore, it is intended that the appended claims shall cover all such changes and modifications which come within the scope of the present application.

[0123] Finally, it is to be understood that the phraseology or terminology such as "first" and "second" etc. used herein is merely intended to differentiate one entity or operation from another entity or operation, without necessarily requiring or implying any actual such relationship or order between such entities or operations. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0124] The above provides a kind of based on image human body back epidermis spinal line detection method and device provided in the application, have carried out detailed introduction, the principle and implementation mode of the application are described in this paper with specific examples, the above example is only for helping understanding the method and its core idea of the application;Meanwhile, for the general technical personnel of the art, according to the idea of the application, there will be changes in specific implementation mode and application range, as described above, the content of the specification should not be understood as the limitation of the application.

Claims

1. A method for detecting the spinal line on the human back skin based on images, characterized in that, The method includes: Acquire a depth image of the back of the subject being tested, and determine the target point cloud data based on the back depth image; The method for determining the coordinates of a predetermined number of feature points of the subject's spine in the back depth image based on the target point cloud data includes: The target point cloud data is input into the stacked hourglass network model, and the stacked hourglass network model is trained. The preset number of feature point coordinates are output through the trained stacked hourglass network model; Determine the number of output neurons in each layer of the stacked hourglass network model and the normalized weight ratio coefficients; Based on the number of neurons and the scaling factor, the weights of the stacked hourglass network model are initialized using the Kaiming network weight initialization method; Based on the coordinates of all the aforementioned feature points, curve fitting is performed to generate the target epidermal spine curve; The method for determining the target spinal scoliosis angle based on the target epidermal spinal curve includes the following: The target normal vector is determined based on the target epidermal spine curve; The target scoliosis angle is determined based on the target normal vector. The calculation formula for the target scoliosis angle using the target normal vector is as follows: ； in, , For the target normal vector, The angle for the target scoliosis angle; The scoliosis angle of the subject is used to determine whether the subject's spine is curved. When the scoliosis angle is less than the preset scoliosis angle, the subject's spine is normal. When the scoliosis angle is greater than or equal to the preset scoliosis angle, the subject's spine is curved.

2. The method according to claim 1, characterized in that, The step of acquiring the back depth image of the subject and determining the target point cloud data based on the back depth image includes: The depth image of the back of the subject is acquired by a depth camera, and the depth image of the back is obtained from the depth camera; Obtain the calibrated intrinsic parameters of the depth image, and determine the 3D point cloud coordinates corresponding to each pixel in the back depth image based on the intrinsic parameters. Based on all the three-dimensional coordinates of the point cloud, the three-dimensional point cloud data is determined, and the three-dimensional point cloud data is denoised using a preset filtering algorithm to obtain the target point cloud data.

3. The method according to claim 1, characterized in that, The step of performing curve fitting based on the coordinates of all the feature points to generate the target epidermal spine curve includes: The Interparc curve interpolation method is used to perform polynomial interpolation fitting on the coordinates of all the feature points to generate the target epidermal spine curve.

4. The method according to claim 3, characterized in that, The step of performing polynomial interpolation fitting on the coordinates of all the feature points using the Interparc curve interpolation method includes: The initial epidermal spine curve is obtained by fitting the Interparc curve interpolation method and the coordinates of all the feature points; Determine the length value of the initial epidermal spine curve, and divide the initial epidermal spine curve into equal parts based on the length value; Linear interpolation is performed on the initial epidermal spine curve after equal division to obtain the coordinates of all target feature points; Based on the coordinates of all the target feature points and the Interparc curve interpolation method, the target epidermal spine curve is fitted and generated.

5. An image-based device for detecting the spinal line on the human back, characterized in that, The device includes: The point cloud determination module is used to acquire a back depth image of the subject being tested and determine the target point cloud data based on the back depth image. The coordinate determination module is used to determine the coordinates of a preset number of feature points of the spine of the subject in the back depth image based on the target point cloud data. The curve generation module is used to perform curve fitting based on the coordinates of all the feature points to generate the target epidermal spine curve. The scoliosis detection module is used to determine the target spinal scoliosis angle based on the target epidermal spinal curve, and to determine whether the subject's spine is scoliotic based on the angle value of the target spinal scoliosis angle.

6. A device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the method as described in any one of claims 1 to 4.

Images