Volume bone density measurement method, device, computer equipment and storage medium

Abstract

The application relates to a volumetric bone density measurement method, device, computer equipment, storage medium and computer program product. The method comprises the following steps: acquiring a scan image of a to-be-measured object collected by a current scanner; identifying a region image of an internal reference from the scan image, and determining an actual CT value of the internal reference according to the region image; acquiring a theoretical CT value of the internal reference, determining a current effective tube voltage and a current system error of the current scanner according to an overall error between the theoretical CT value of the internal reference and the actual CT value of the internal reference; acquiring an actual CT value of a cancellous bone region of a to-be-measured part in the scan image, and determining a volumetric bone density of the cancellous bone region according to the actual CT value of the cancellous bone region, the current effective tube voltage and the current system error. The method can improve the measurement efficiency of the volumetric bone density.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method, apparatus, computer device, storage medium, and computer program product for measuring volumetric bone density. Background Technology

[0002] VBMD (Volumetric Bone Mineral Density) refers to the content of HA (Hydroxyapatite) per unit volume of bone, and its unit is mg / cm3. It reflects the loss of minerals in bone and is of great significance in osteoporosis diagnosis and fracture risk prediction.

[0003] Traditionally, volumetric bone mineral density (VBMD) measurements in the cancellous bone region are primarily performed using QCT (Quantitative Computed Tomography). However, differences exist between various CT scanners, requiring a standard VBMD phantom as an external reference for calibrating each scanner before QCT measurement. However, the fabrication of a standard VBMD phantom is demanding, complex, and time-consuming, resulting in low efficiency in volumetric bone mineral density measurement. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer device, computer-readable storage medium, and computer program product for measuring volumetric bone mineral density that can improve the measurement efficiency of volumetric bone mineral density in response to the above-mentioned technical problems.

[0005] In a first aspect, this application provides a method for measuring volumetric bone mineral density, including:

[0006] Acquire the scanned image of the object to be tested captured by the current scanner;

[0007] An internal reference region image is identified from the scanned image, and the actual CT value of the internal reference is determined based on the region image;

[0008] Obtain the theoretical CT value of the internal reference, and determine the current effective tube voltage and current system error of the current scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference.

[0009] Obtain the actual CT value of the cancellous bone region of the test site in the scanned image, and determine the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error.

[0010] In one embodiment, determining the current effective tube voltage and current system error of the current scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference includes:

[0011] Construct a loss model; the loss model is used to evaluate the overall error between the theoretical CT value of the internal reference and the actual CT value of the internal reference;

[0012] The output value of the loss model is updated based on the current scanner's predicted effective tube voltage and predicted system error until the output value of the loss model meets the preset conditions.

[0013] The predicted effective transistor voltage and the predicted system error when the output value of the loss model satisfies the preset conditions are taken as the current effective transistor voltage and the current system error, respectively.

[0014] In one embodiment, obtaining the actual CT value of the cancellous bone region of the site to be tested in the scanned image includes:

[0015] Obtain a region image of the area to be tested from the scanned image;

[0016] The image of the location is input into a pre-trained cancellous bone segmentation model to obtain the cancellous bone region in the image of the location.

[0017] The actual CT value of the cancellous bone region is determined based on its grayscale value.

[0018] In one embodiment, the pre-trained cancellous bone segmentation model is trained in the following manner:

[0019] Acquire images of the sample site and the labeled cancellous bone regions in the sample site images;

[0020] The sample area image is input into the bone segmentation model to be trained to obtain the predicted bone region in the sample area image;

[0021] Based on the difference between the predicted cancellous bone region and the labeled cancellous bone region, the cancellous bone segmentation model to be trained is iteratively trained to obtain a trained cancellous bone segmentation model, which is used as the pre-trained cancellous bone segmentation model.

[0022] In one embodiment, the cancellous bone region includes hydroxyapatite and other substances, the other substances being substances in the cancellous bone region other than hydroxyapatite, and the other substances including at least water;

[0023] The step of determining the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error includes:

[0024] Obtain the intrinsic density of the reference substance and the actual concentrations of the other substances;

[0025] The actual CT value of the cancellous bone region, the current effective tube voltage, the current system error, the intrinsic density of the reference material, and the actual concentration of the other materials are input into the volumetric bone density measurement model to obtain the actual concentration of hydroxyapatite.

[0026] The actual concentration of the hydroxyapatite was confirmed as the volumetric bone density of the cancellous bone region.

[0027] In one embodiment, the step of inputting the actual CT value of the cancellous bone region, the current effective tube voltage, the current system error, the intrinsic density of the reference material, and the actual concentration of the other substances into the volumetric bone mineral density measurement model to obtain the actual concentration of hydroxyapatite includes:

[0028] Based on the current effective tube voltage, determine the mass decay coefficient of the reference material, the mass decay coefficient of the hydroxyapatite, and the mass decay coefficients of the other materials;

[0029] The actual CT value of the cancellous bone region, the current systematic error, the mass decay coefficient and intrinsic density of the reference substance, the mass decay coefficient and actual concentration of other substances, and the mass decay coefficient of hydroxyapatite are input into the volumetric bone density measurement model to obtain the actual concentration of hydroxyapatite.

[0030] In one embodiment, obtaining the theoretical CT value of the internal reference includes:

[0031] Obtain the intrinsic density of the internal reference and the intrinsic density of the reference material;

[0032] The predicted effective tube voltage and predicted system error of the current scanner, as well as the intrinsic density of the internal reference and the intrinsic density of the reference material, are input into the theoretical CT value statistical model to obtain the theoretical CT value of the internal reference.

[0033] Secondly, this application also provides a volumetric bone mineral density measuring device, comprising:

[0034] The image acquisition module is used to acquire the scanned image of the object under test captured by the current scanner;

[0035] An image recognition module is used to identify an internal reference region image from the scanned image and determine the actual CT value of the internal reference based on the region image;

[0036] An error processing module is used to obtain the theoretical CT value of the internal reference, and determine the current effective tube voltage and current system error of the current scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference.

[0037] The density determination module is used to obtain the actual CT value of the cancellous bone region of the test site in the scanned image, and determine the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error.

[0038] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0039] Acquire the scanned image of the object to be tested captured by the current scanner;

[0040] An internal reference region image is identified from the scanned image, and the actual CT value of the internal reference is determined based on the region image;

[0041] Obtain the theoretical CT value of the internal reference, and determine the current effective tube voltage and current system error of the current scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference.

[0042] Obtain the actual CT value of the cancellous bone region of the test site in the scanned image, and determine the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error.

[0043] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0044] Acquire the scanned image of the object to be tested captured by the current scanner;

[0045] An internal reference region image is identified from the scanned image, and the actual CT value of the internal reference is determined based on the region image;

[0046] Obtain the theoretical CT value of the internal reference, and determine the current effective tube voltage and current system error of the current scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference.

[0047] Obtain the actual CT value of the cancellous bone region of the test site in the scanned image, and determine the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error.

[0048] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0049] Acquire the scanned image of the object to be tested captured by the current scanner;

[0050] An internal reference region image is identified from the scanned image, and the actual CT value of the internal reference is determined based on the region image;

[0051] Obtain the theoretical CT value of the internal reference, and determine the current effective tube voltage and current system error of the current scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference.

[0052] Obtain the actual CT value of the cancellous bone region of the test site in the scanned image, and determine the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error.

[0053] The aforementioned volumetric bone mineral density measurement method, apparatus, computer equipment, storage medium, and computer program product acquire a scanned image of the object to be measured by the current scanner, identify an internal reference region image from the scanned image, and determine the actual CT value of the internal reference based on the region image; then, acquire the theoretical CT value of the internal reference, and determine the current effective tube voltage and current systematic error of the current scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference; finally, acquire the actual CT value of the cancellous bone region of the object to be measured in the scanned image, and determine the bone mineral density based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current systematic error. The volumetric bone mineral density (BMD) of the cancellous bone region is measured. In this way, when measuring the BMD of the cancellous bone region of the test site, the current effective tube voltage and current system error of the scanner can be determined solely by the overall error between the theoretical CT value and the actual CT value of the internal reference. Simultaneously, combined with the actual CT value of the cancellous bone region of the test site, the BMD of the cancellous bone region can be calculated. The entire measurement process uses only the internal reference, eliminating the need for external references such as standard phantoms. That is, the internal reference replaces the standard phantom to calibrate the current scanner for phantom-free self-calibration, thus simplifying the BMD measurement process and improving its efficiency. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a diagram illustrating the application environment of a volumetric bone density measurement method in one embodiment.

[0056] Figure 2 This is a flowchart illustrating a volumetric bone density measurement method in one embodiment;

[0057] Figure 3 This is a flowchart illustrating the steps for determining the current effective transistor voltage and the current system error in one embodiment;

[0058] Figure 4 This is a graph showing the relationship between overall error and effective transistor voltage in one embodiment;

[0059] Figure 5 This is a flowchart of the steps for obtaining the actual CT value of the cancellous bone region of the site to be tested in a scanned image, as described in one embodiment.

[0060] Figure 6 This is a flowchart illustrating the steps for determining the volumetric bone density of a cancellous bone region in one embodiment.

[0061] Figure 7 This is a flowchart illustrating a volumetric bone density measurement method in another embodiment;

[0062] Figure 8 This is a flowchart illustrating a method for determining vertebral volumetric bone mineral density using phantom-free self-calibrated quantitative computed tomography based on multi-parameter optimization in one embodiment.

[0063] Figure 9 This is a structural block diagram of a volumetric bone density measuring device in one embodiment;

[0064] Figure 10 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0066] The volumetric bone mineral density measurement method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, scanner 102 currently communicates with terminal 104 via a network. Specifically, refer to... Figure 1 When measuring the volumetric bone mineral density of the cancellous bone region of the test site, the current scanner 102 first acquires a scan image of the test subject and sends the acquired scan image to the terminal 104. The terminal 104 identifies the internal reference region image from the received scan image and determines the actual CT value of the internal reference based on the region image; then, it acquires the theoretical CT value of the internal reference, and determines the current effective tube voltage and current system error of the current scanner 102 based on the overall error between the theoretical CT value and the actual CT value of the internal reference; finally, it acquires the actual CT value of the cancellous bone region of the test site in the scan image, and determines the volumetric bone mineral density of the cancellous bone region of the test site based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error. Here, the current scanner 102 refers to a CT scanner; the terminal 104 can be, but is not limited to, various personal computers, laptops, smartphones, and tablets.

[0067] In one exemplary embodiment, such as Figure 2 As shown, a method for measuring volumetric bone mineral density is provided, which can be applied to... Figure 1 Taking the terminal in the example, the explanation includes the following steps S201 to S204. Wherein:

[0068] Step S201: Obtain the scanned image of the object to be tested captured by the current scanner.

[0069] The current scanner refers to the CT scanner currently scanning the object under test. In a real-world scenario, a CT scanner includes a CT tube, which consists of a filament (cathode) that emits electrons and a target (anode) that receives electron bombardment.

[0070] It should be noted that when a voltage is applied to the cathode and anode of a CT tube, the electrons emitted from the cathode are suddenly accelerated due to the potential difference (i.e., tube voltage), colliding with atoms in the target material to form X-rays. X-rays have good penetrating power, and their energy attenuates to varying degrees as they penetrate the material. These varying degrees of energy attenuation are reflected in the CT image, which is the CT value. Because the time and conditions for the electrons emitted from the cathode to reach the target material differ, the energy of the generated X-rays is not a fixed value, but rather a continuous energy distribution around the center, known as the X-ray continuum. The peak value in the X-ray continuum is the KVp value set when manually operating the CT scanner; the average effect of the mixed-energy X-rays, i.e., the effective tube voltage KVe, determines the CT value of the CT image. Due to differences in manufacturing processes among different manufacturers and CT scanners, the relationship between KVp and KVe varies between CT scanners (i.e., due to the uniqueness of different CT scanners, the KVe of each CT scanner is not the same for a given KVp), and it will drift slightly over time; therefore, it is necessary to periodically calibrate the effective tube voltage KVe of the CT scanner.

[0071] The subject to be tested refers to the object whose volumetric bone density in the cancellous bone region needs to be measured, such as a patient.

[0072] Here, scanned images refer to CT images, such as whole-body CT images. In practical scenarios, scanned images include at least an image of the internal reference area and an image of the area to be tested.

[0073] For example, the current scanner responds to the scanning command for the object to be tested, scans the object to be tested, obtains a scanned image of the object to be tested, and sends the scanned image of the object to be tested to the corresponding terminal; in this way, the terminal can obtain the scanned image of the object to be tested acquired by the current scanner, which facilitates a series of analyses on the scanned image to obtain the volumetric bone density of the cancellous bone region of the object to be tested.

[0074] Step S202: Identify the internal reference region image from the scanned image, and determine the actual CT value of the internal reference based on the region image.

[0075] Internal references refer to tissues with relatively stable density, such as fat, muscle, and spleen. It should be noted that there are many internal references, and their selection can be flexibly combined based on the specific circumstances of the subject. For example, for subject A, fat and muscle can be chosen as internal references; for subject B, muscle and spleen can be chosen.

[0076] It should be noted that the densities of tissues such as fat, muscle, and spleen vary little among individuals, while their mass decay coefficients vary to some extent under different effective tube voltages. Therefore, tissues such as fat, muscle, and spleen can be selected as internal references for phantom-free self-calibration. Furthermore, the densities (i.e., intrinsic densities) of tissues such as fat, muscle, and spleen are known constants and can be obtained by querying a first database containing the densities of multiple tissues. The first database also stores the intrinsic densities of other substances (such as water). Additionally, this application can establish a second database, independent of the CT scanner, that reflects only the relationship between the mass decay coefficient of a substance (such as tissue) and the effective tube voltage; that is, the second database stores the mass decay coefficients of different substances under different effective tube voltages. Moreover, the mass decay coefficient of a substance is only related to the effective tube voltage and the substance's inherent properties.

[0077] The internal reference region image refers to the image of the region containing the internal reference, such as an image of the fat region, muscle region, or spleen region. Each pixel in the internal reference region image corresponds to a CT value (i.e., grayscale value). The actual CT value of the internal reference refers to the actual grayscale value of the internal reference, specifically the mean or peak value of the CT values ​​of all pixels in the internal reference region image. It should be noted that since the scanned image of the object under test is obtained through scanning with the current scanner, the CT value of the internal reference determined based on the internal reference region image in the scanned image is called the actual CT value of the internal reference.

[0078] For example, the terminal performs segmentation processing on the scanned image to obtain an internal reference region image; for instance, the scanned image is input into a pre-trained internal reference segmentation model for segmentation processing to obtain the internal reference region image. Then, the terminal calculates the actual CT value of the internal reference based on the CT values ​​of each pixel in the internal reference region image. For example, the terminal uses the average CT values ​​of all pixels in the internal reference region image as the actual CT value of the internal reference; or, the terminal uses the peak value of the CT values ​​of all pixels in the internal reference region image as the actual CT value of the internal reference.

[0079] Furthermore, the terminal can select the corresponding target internal reference marker (such as fat marker or muscle marker) based on the object information of the object to be tested (such as physical condition), input the target internal reference marker and the scanned image into the pre-trained internal reference segmentation model, and perform targeted segmentation processing on the scanned image based on the target internal reference marker to obtain the region image of the internal reference corresponding to the target internal reference marker, and calculate the actual CT value of the internal reference based on the region image.

[0080] Step S203: Obtain the theoretical CT value of the internal reference, and determine the current effective tube voltage and current system error of the current scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference.

[0081] For a single substance i in a CT image, the CT value is calculated as follows:

[0082]

[0083] Among them, CT value i (KV e () represents the effective tube voltage (KV) of a single substance i in a CT scanner. e CT value under f i (KV e ) and f H2O (KV e ) represent the mass decay coefficients of single substance i and water, respectively, and are only related to KV. e The properties of a single substance i and water are related to KV. e The function is expressed in g / cm³. 2 Used to calibrate the current KV e The response of grayscale values ​​in CT images to the concentrations of a single substance i and water. i (KV e ) and f H2O (KV e () represent the mass decay coefficients of single substance i and water, respectively, in relation to KV. e The relationship data can be obtained by querying a second database. ρ i and ρ H2O Let i and d represent the intrinsic densities of a single substance i and water, respectively; both are constants.

[0084] According to Formula 1, if substance i is water, then the corresponding CT value is... i (KV e The CT value is 0. However, CT scanners themselves may have a slight systematic error; that is, when scanning water, the corresponding CT value will be slightly different from 0. This systematic error is called drift, and it is relatively fixed within a certain time range for a given CT scanner. Therefore, CT scanners need to periodically scan water phantoms to correct for this systematic error. Considering the existence of systematic error, Equation 1 can be adjusted to:

[0085]

[0086] Where ε represents the systematic error of the CT scanner.

[0087] For a CT image of a mixture composed of N substances, the CT value is calculated as follows:

[0088]

[0089] Among them, CT value N (KV e () represents the effective tube voltage (kV) of a mixture of N substances in a CT scanner. e The CT value below. f i (KV e ρ represents the mass decay coefficient of substance i in the mixture. i 'Indicates the actual concentration of substance i in the mixture, that is, the content of substance i in the mixture, rather than the intrinsic density of substance i.

[0090] Based on Formula 2, the theoretical CT value of substance i is established. i (KV e ) and the effective tube voltage (KV) of the CT scanner e Relationship:

[0091] theory

[0092] Where i represents internal references for fat, muscle, spleen, etc.; f i (KV e In fact, it is a universal material i that is independent of the CT scanner and varies with the effective tube voltage KV. e A changing database (i.e., a second database). It should be noted that this application will store a general database of material i varying with the effective tube voltage KV. e A changing database is called a self-calibrating database.

[0093] The theoretical CT value of the internal reference refers to the theoretical grayscale value of the internal reference. Based on Formula 4, the theoretical CT value of the internal reference is related to the systematic error of the CT scanner, the mass attenuation coefficient and intrinsic density of the internal reference, and the mass attenuation coefficient and intrinsic density of water. The mass attenuation coefficient of the internal reference is related to its own characteristics and the effective tube voltage of the CT scanner, while the mass attenuation coefficient of water is related to the own characteristics of water and the effective tube voltage of the CT scanner. Since the intrinsic density of the internal reference and the intrinsic density of water are constants, the theoretical CT value of the internal reference can be calculated after determining the effective tube voltage and systematic error of the CT scanner.

[0094] The overall error refers to the error obtained by combining the differences between the theoretical CT values ​​and the actual CT values ​​of the identified internal references. Specifically, it refers to SSE (the sum of squares due to error), which is the sum of the squares of the differences between the theoretical CT values ​​and the actual CT values ​​of each internal reference. The current effective tube voltage and current system error of the scanner refer to the effective tube voltage estimate and system error estimate when the overall error between the theoretical CT values ​​and the actual CT values ​​of the internal references is minimized.

[0095] It should be noted that determining the current effective tube voltage and current systematic error of the scanner is equivalent to calibrating the scanner. Moreover, the core and key to determining the volumetric bone mineral density of the cancellous bone region lies in accurately obtaining the current effective tube voltage and current systematic error of the scanner during measurement.

[0096] For example, the terminal obtains the predicted effective tube voltage and predicted systematic error (e.g., randomly determined effective tube voltage and systematic error) of the current scanner. Based on the predicted effective tube voltage, it queries the correspondence between the effective tube voltage and the mass decay coefficient of the internal reference, and the correspondence between the effective tube voltage and the mass decay coefficient of water, respectively, to obtain the mass decay coefficient of the internal reference and the mass decay coefficient of water. It then queries a first database storing the intrinsic densities of substances to obtain the intrinsic density of the internal reference and the intrinsic density of water. Finally, it inputs the mass decay coefficient and intrinsic density of the internal reference, the mass decay coefficient and intrinsic density of water, and the predicted systematic error into the theoretical CT value statistical model to obtain the theoretical CT value of the internal reference. For example, the terminal calculates the theoretical CT values ​​of internal references such as fat, muscle, and spleen according to Formula 4.

[0097] Next, the terminal inputs the theoretical CT value and the actual CT value of the internal reference into the overall error statistical model to obtain the overall error between the theoretical CT value and the actual CT value of the internal reference. For example, the terminal calculates the sum of the squares of the differences between the theoretical CT value and the actual CT value of each internal reference as the overall error between the theoretical CT value and the actual CT value of the internal reference. Furthermore, the terminal continuously adjusts the predicted effective transistor voltage and the predicted system error to update the overall error. When the overall error meets preset conditions (such as minimizing the overall error), the predicted effective transistor voltage and the predicted system error at this time are used as the current effective transistor voltage and the current system error of the current scanner.

[0098] Step S204: Obtain the actual CT value of the cancellous bone region of the area to be tested in the scanned image, and determine the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error.

[0099] The area to be tested can refer to all bones in the body, such as the vertebral body, which includes cancellous bone regions. These can be segmented using a pre-trained cancellous bone segmentation model. The actual CT value of the cancellous bone region refers to the actual grayscale value of the cancellous bone region, specifically the mean or peak value of the CT values ​​of all pixels in the cancellous bone region.

[0100] Among them, the volumetric bone mineral density of the cancellous bone region refers to the actual concentration of HA in the cancellous bone, that is, the content of HA in the cancellous bone.

[0101] In this context, cancellous bone (such as vertebral cancellous bone) can be approximated as a mixture of HA and water, meaning that cancellous bone contains HA and other substances, with the other substances including at least water. Based on Formula 3, the actual concentration of HA can be determined to be related to the actual CT value of the cancellous bone region, the current systematic error, the mass decay coefficient and intrinsic density of the reference substance (such as water), the mass decay coefficients and actual concentrations of other substances, and the mass decay coefficient of HA. The mass decay coefficient of the reference substance is related to its own characteristics and the current effective tube voltage. The mass decay coefficients of other substances are related to their own characteristics and the current effective tube voltage. The mass decay coefficient of HA is related to its own characteristics and the current effective tube voltage. Since the intrinsic density of the reference substance and the actual concentrations of other substances are constants, the actual concentration of HA in the cancellous bone can be calculated after determining the actual CT value, current systematic error, and current effective tube voltage of the cancellous bone region.

[0102] For example, the terminal performs segmentation processing on the scanned image to obtain a region image of the area to be tested; for instance, the scanned image is input into a pre-trained segmentation model (such as a vertebral body segmentation model) for segmentation processing to obtain a region image of the area to be tested. Next, the terminal performs cancellous bone region identification on the region image to obtain the cancellous bone region within the region image; for instance, the region image is input into a pre-trained cancellous bone segmentation model to obtain the cancellous bone region within the region image. Then, the terminal calculates the actual CT value of the cancellous bone region based on the CT values ​​of each pixel within the cancellous bone region. For example, the terminal uses the average CT values ​​of all pixels in the cancellous bone region as the actual CT value of that region; or, the terminal uses the peak value of the CT values ​​of all pixels in the cancellous bone region as the actual CT value of that region.

[0103] Next, the terminal queries the first database to obtain the intrinsic density of the reference substance, and queries the third database, which stores the actual concentrations of multiple substances, to obtain the actual concentrations of other substances. Then, it inputs the intrinsic density of the reference substance, the actual concentrations of other substances, the actual CT value of the cancellous bone region, the current system error, and the current effective tube voltage into the HA concentration statistical model to obtain the actual concentration of HA in the cancellous bone, which is used as the volumetric bone mineral density of the cancellous bone region. For example, the terminal calculates the mass decay coefficient of the reference substance, the mass decay coefficient of other substances, and the mass decay coefficient of HA based on the current effective tube voltage through the HA concentration statistical model. Then, based on the actual CT value of the cancellous bone region, the current system error, the mass decay coefficient and intrinsic density of the reference substance, the mass decay coefficient and actual concentration of other substances, and the mass decay coefficient of HA, it calculates the actual concentration of HA in the cancellous bone, which is used as the volumetric bone mineral density of the cancellous bone region.

[0104] For example, the volumetric bone mineral density of the cancellous bone region = the actual concentration of HA in the cancellous bone = [((actual CT value of the cancellous bone region - current systematic error) / 1000+1) × mass decay coefficient of the reference material at the current effective tube voltage × intrinsic density of the reference material - mass decay coefficient of other materials at the current effective tube voltage × actual concentration of other materials] / (mass decay coefficient of HA at the current effective tube voltage).

[0105] It should be noted that, assuming the vertebral bone cancellous tissue includes N substances, the first substance is HA, and the second to Nth substances are other substances, then the mass decay coefficient of other substances under the current effective tube voltage × the actual concentration of other substances = the mass decay coefficient of the second substance under the current effective tube voltage × the actual concentration of the second substance + the mass decay coefficient of the third substance under the current effective tube voltage × the actual concentration of the third substance + ... + the mass decay coefficient of the Nth substance under the current effective tube voltage × the actual concentration of the Nth substance.

[0106] In the above-mentioned method for measuring volumetric bone mineral density, when measuring the volumetric bone mineral density of the cancellous bone region of the test site, the current effective tube voltage and current system error of the scanner can be determined by using only the overall error between the theoretical CT value and the actual CT value of the internal reference. At the same time, combined with the actual CT value of the cancellous bone region of the test site, the volumetric bone mineral density of the cancellous bone region can be calculated. The entire measurement process uses only the internal reference and does not require the creation of external references such as standard phantoms. That is, the internal reference is used instead of the standard phantom to calibrate the current scanner for phantom-free self-calibration, thereby simplifying the volumetric bone mineral density measurement process and improving the measurement efficiency.

[0107] In one exemplary embodiment, such as Figure 3As shown, step S203 above, which determines the current effective tube voltage and current system error of the scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference, specifically includes steps S301 to S303. Wherein:

[0108] Step S301: Construct a loss model; the loss model is used to evaluate the overall error between the theoretical CT value of the internal reference and the actual CT value of the internal reference.

[0109] Step S302: Update the output value of the loss model based on the current predicted effective tube voltage and predicted system error of the scanner until the output value of the loss model meets the preset conditions.

[0110] Step S303: The predicted effective transistor voltage and predicted system error when the output value of the loss model meets the preset conditions are used as the current effective transistor voltage and current system error.

[0111] The loss model refers to an overall error statistical model, such as the SSE statistical model. Of course, the loss model can also refer to models corresponding to other loss functions; this application does not limit the specific model. The output value of the loss model refers to the overall error between the theoretical CT value and the actual CT value of the internal reference, such as SSE. Furthermore, the output value of the loss model is related to the current scanner's predicted effective tube voltage and prediction system error. By adjusting the current scanner's predicted effective tube voltage and prediction system error, the theoretical CT value of the internal reference can be adjusted, thereby updating the output value of the loss model.

[0112] The output value of the loss model satisfies a preset condition, meaning the output value of the loss model is at its minimum, such as the minimum SSE. It should be noted that when the output value of the loss model is at its minimum, it means that the predicted effective transistor voltage and predicted system error when calculating the theoretical CT value of the internal reference are infinitely close to the effective transistor voltage and system error when calculating the actual CT value of the internal reference. Therefore, the predicted effective transistor voltage and predicted system error at this time can be used as the corresponding current effective transistor voltage and current system error.

[0113] Here, the current effective transistor voltage and the current system error refer to the predicted effective transistor voltage (i.e., the effective transistor voltage estimate) and the predicted system error (i.e., the system error estimate) when the output value of the loss model is minimized.

[0114] For example, the terminal first determines a model for the overall error between the theoretical CT value and the actual CT value used for statistical internal reference, as a loss model; it adjusts the predicted effective tube voltage and prediction system error of the current scanner to obtain the output value of the loss model; if the output value of the loss model does not meet a preset condition, it jumps to the step of adjusting the predicted effective tube voltage and prediction system error of the current scanner to obtain the output value of the loss model, until the output value of the loss model meets the preset condition, such as the minimum output value of the loss model. When the output value of the loss model meets the preset condition, the predicted effective tube voltage and prediction system error at this time are taken as the current effective tube voltage and current system error of the current scanner.

[0115] For example, the overall error between the theoretical CT value and the actual CT value of the internal reference is related to the effective tube voltage and system error of the current scanner, involving a multi-parameter optimization problem. This involves finding the effective tube voltage and system error of the current scanner such that the overall error between the theoretical CT value and the actual CT value of the internal reference, calculated based on the effective tube voltage and system error of the current scanner and Equation 4, is minimized. The specific implementation process is as follows: The terminal first determines the loss function for evaluating the overall error between the theoretical CT value and the actual CT value of the internal reference, such as SSE as shown in Equation 5; then, using methods such as least squares, it solves for the effective tube voltage estimate and system error estimate when the overall error between the theoretical CT value and the actual CT value of the internal reference is minimized. (See the attached diagram for details.) Figure 4 . Figure 4 In the diagram, the horizontal axis represents the effective transistor voltage KVe, and the vertical axis represents the sum of squared errors (SSE).

[0116]

[0117] Among them, theoretical CT value i (KV e This represents the theoretical CT value for internal references such as fat, muscle, and spleen, specifically calculated using Formula 4. The actual CT value... i This represents the actual CT value of internal references such as fat, muscle, and spleen, which is calculated from the CT values ​​of each pixel in the region image of the internal reference.

[0118] In this embodiment, the current effective tube voltage and current system error of the current scanner are determined by utilizing the overall error between the theoretical CT value and the actual CT value of the internal reference. This achieves the purpose of calibrating the current scanner without the need for calibration using a standard phantom, thus simplifying the calibration process and improving the calibration efficiency of the current scanner. This further improves the efficiency of subsequent vertebral volumetric bone mineral density measurements. Simultaneously, comprehensively considering the overall error between the theoretical CT value and the actual CT value of the internal reference helps to improve the accuracy of determining the current effective tube voltage and current system error.

[0119] In one exemplary embodiment, such as Figure 5 As shown, step S204 above, which obtains the actual CT value of the cancellous bone region of the area to be tested in the scanned image, specifically includes the following steps S501 to S503. Wherein:

[0120] Step S501: Obtain the part image of the part to be tested in the scanned image.

[0121] Step S502: Input the site image into the pre-trained cancellous bone segmentation model to obtain the cancellous bone region in the site image.

[0122] Step S503: Determine the actual CT value of the cancellous bone region based on the gray value of the cancellous bone region.

[0123] The scanned images include images of the regions to be tested within the object being tested. A region image refers to an image of the area containing the region to be tested, such as an image of the vertebral body to be tested.

[0124] The process may further include, after step S501: the terminal performs a first preprocessing on the part image to obtain a preprocessed part image, and inputs the preprocessed part image into a pre-trained cancellous bone segmentation model to obtain the cancellous bone region in the preprocessed part image.

[0125] The first preprocessing step includes vertebral center point detection, spinal centerline fitting, vertebral extraction and rotation, resolution standardization, and grayscale standardization. The preprocessed site image refers to the image obtained after performing the first preprocessing step on the site image, such as a preprocessed vertebral body image. The pre-trained cancellous bone segmentation model is a deep learning model used to segment cancellous bone regions; it is obtained through multiple training iterations.

[0126] The grayscale value of the cancellous bone region includes the grayscale value (i.e., CT value) of each pixel in the cancellous bone region.

[0127] For example, the terminal inputs the scanned image into a pre-trained segmentation model (such as a vertebral segmentation model) for segmentation processing to obtain a region image of the area to be tested. The segmentation model is a deep learning model used to segment the corresponding image, and the vertebral segmentation model is a deep learning model used to segment the vertebral body image. Next, the terminal performs a first preprocessing on the region image to obtain a preprocessed region image. This preprocessing includes vertebral center point detection, spinal centerline fitting, vertebral extraction and rotation, resolution standardization, and grayscale value standardization. Then, the terminal inputs the preprocessed region image into a pre-trained cancellous bone segmentation model. The pre-trained cancellous bone segmentation model segments the cancellous bone region in the preprocessed region image to obtain the cancellous bone region in the preprocessed region image. Finally, the terminal obtains the grayscale value (i.e., CT value) of each pixel in the cancellous bone region and uses the peak value of all grayscale values ​​in the cancellous bone region as the actual CT value of that region. For example, the terminal performs grayscale histogram analysis on the CT values ​​of each pixel in the cancellous bone region and takes the peak value as the actual CT value of that cancellous bone region.

[0128] In this embodiment, the area image of the test site in the scanned image is preprocessed, and the cancellous bone region in the preprocessed area image is segmented using a pre-trained cancellous bone segmentation model. Finally, the actual CT value of the cancellous bone region is determined based on its grayscale value. In this way, after preprocessing the area image of the test site, the cancellous bone region is segmented from the preprocessed area image using the pre-trained cancellous bone segmentation model, which helps to improve the segmentation accuracy of the cancellous bone region, thereby improving the accuracy of determining the actual CT value of the cancellous bone region.

[0129] In an exemplary embodiment, step S501, after acquiring the region image of the region to be tested in the scanned image, further includes a step of performing a first preprocessing on the region image to obtain a preprocessed region image. Assuming the region image refers to a vertebral body image, the first preprocessing of the vertebral body image to obtain a preprocessed vertebral body image specifically includes the following: identifying the center points of each vertebral segment in the vertebral body image; fitting the center points of each vertebral segment to obtain the spinal center curve in the vertebral body image; determining the tangent vectors at the locations of the center points of each vertebral segment based on the spinal center curve; cropping the vertebral body image based on the center points of each vertebral segment; and rotating the cropped vertebral body image to a preset size based on the tangent vectors to obtain a rotated vertebral body image that satisfies the three-dimensional spatial coordinate system; and performing resolution normalization and grayscale normalization on the rotated vertebral body image to obtain the preprocessed vertebral body image.

[0130] The spinal center curve refers to the curve obtained by fitting the center points of each vertebral segment, specifically a curve that runs through the center of the spine. The cropped vertebral image that meets the preset size refers to a cropped vertebral image with the same dimensions as the preset size. The cropped vertebral image contains one complete, independent vertebra and a small amount of surrounding soft tissue.

[0131] The rotation of the vertebral body image satisfies a three-dimensional coordinate system, meaning that the tangent vector at the center point of the vertebra is parallel to the z-axis, and the upper and lower surfaces of the vertebra are parallel to the xy-plane. Resolution normalization refers to adjusting the resolution of the rotation of the vertebral body image to a preset resolution, ensuring that each rotation of the vertebral body image has the same resolution.

[0132] Gray-scale normalization refers to adjusting the gray-scale values ​​of each pixel in the vertebral body image after resolution normalization to the range of [-1, 1].

[0133] For example, the terminal uses a pre-trained deep learning point detection model to detect key points in the vertebral body image, obtaining the center points of each vertebral segment and the segment where the center point of the vertebra is located. Based on a fitting algorithm (such as the cubic B-spline algorithm), the terminal fits the center points of each vertebral segment to obtain a curve that runs through the center points of each vertebral segment, which serves as the central curve of the spine in the vertebral body image. Next, the terminal calculates the tangent vector at the location of the center point of each vertebral segment based on the central curve of the spine. Using the center point of each vertebral segment as the core, the terminal crops each segment of the vertebral body image around the center point of each vertebral segment to obtain a cropped vertebral body image that meets the preset size. Then, the terminal rotates the cropped vertebral body image according to the tangent vectors at the center points of each vertebral segment until the rotated vertebral body image satisfies the three-dimensional spatial coordinate system. Finally, the terminal performs resolution normalization processing on the rotated vertebral body image to obtain a resolution-adjusted vertebral body image, and performs grayscale value normalization processing on the resolution-adjusted vertebral body image to obtain a grayscale value-adjusted vertebral body image, which serves as the preprocessed vertebral body image.

[0134] For example, the terminal performs a first preprocessing on the vertebral body image, which mainly includes the following processes (1) to (5), wherein:

[0135] (1) Vertebral center point detection: Based on the previous deep learning point detection model, the center point of each segment of the vertebral body image is calibrated and the segment of each vertebra is determined.

[0136] (2) Spine centerline fitting: Based on the cubic B-spline algorithm, these center points are fitted to obtain the curve that runs through the center of the spine, and the tangent vector at the location of the center point of each vertebra is calculated.

[0137] (3) Vertebral extraction and rotation: Using the center point of each vertebral segment as the core, the images are cropped to a uniform size around these center points; for example, vertebral images of the same size are extracted for different heights; the cropped image contains one complete independent vertebra and a small amount of surrounding soft tissue. Based on the tangent vector of the vertebra and its center point, the cropped image is rotated so that in three-dimensional space, the tangent vector at the center point of the vertebra is parallel to the z-axis, and the upper and lower surfaces of the vertebral body are parallel to the xy plane.

[0138] (4) Resolution normalization: After resampling, the images have the same resolution.

[0139] (5) Gray value standardization: Through image standardization processing, the gray value of each pixel in the image is adjusted to the range of [-1, 1].

[0140] In this embodiment, the center points of each vertebral segment in the vertebral body image are identified and fitted into a spinal center curve. At the same time, the tangent vectors at the locations of the center points of each vertebral segment are determined. Then, based on the center points of each vertebral segment and the tangent vectors at the locations of the center points of each vertebral segment, the vertebral body image is cropped, rotated, and subjected to resolution standardization and grayscale standardization. In this way, by performing a series of preprocessing steps on the vertebral body image, it is beneficial to further optimize the vertebral body image, so that the cancellous bone region determined from the optimized vertebral body image is more accurate.

[0141] In an exemplary embodiment, the volumetric bone mineral density measurement method provided in this application further includes a training step for a cancellous bone segmentation model, specifically including the following: acquiring a sample site image and a labeled cancellous bone region in the sample site image; inputting the sample site image into the cancellous bone segmentation model to be trained to obtain a predicted cancellous bone region in the sample site image; iteratively training the cancellous bone segmentation model to be trained based on the difference between the predicted cancellous bone region and the labeled cancellous bone region to obtain a trained cancellous bone segmentation model, which serves as the pre-trained cancellous bone segmentation model.

[0142] Among them, the sample site image refers to the vertebral body image participating in the training, such as the sample vertebral body image; the labeled cancellous bone region refers to the manually labeled cancellous bone region in the sample site image (such as the sample vertebral body image), which is an elliptical cylindrical structure; while accommodating as much cancellous bone as possible, avoid touching the lesion area, the endplate edge and the vertebral foramen area.

[0143] The process includes, after acquiring the sample site image and the labeled cancellous bone region in the sample site image, the following steps: the terminal performs a first preprocessing on the sample site image to obtain a preprocessed sample site image; and inputs the preprocessed sample site image into the cancellous bone segmentation model to be trained to obtain the predicted cancellous bone region in the preprocessed sample site image.

[0144] Here, the preprocessed sample area image refers to the image obtained after performing a first preprocessing on the sample area image, such as a preprocessed sample vertebral body image. The bone segmentation model to be trained refers to a deep learning model, such as the VB-Net (V-shape bottleneck Net) model.

[0145] For example, the terminal obtains a sample site image and a labeled cancellous bone region in the sample site image from a local database; performs a first preprocessing on the sample site image to obtain a preprocessed sample site image; inputs the preprocessed sample site image into the cancellous bone segmentation model to be trained for segmentation processing to obtain a predicted cancellous bone region in the preprocessed sample site image; calculates a first loss value based on the difference between the predicted cancellous bone region and the labeled cancellous bone region; iteratively trains the cancellous bone segmentation model to be trained based on the first loss value until the training termination condition is met (e.g., the first loss value is less than a first preset threshold, or a preset number of training iterations are reached), then the trained cancellous bone segmentation model at this point is taken as the trained cancellous bone segmentation model and confirmed as the pre-trained cancellous bone segmentation model.

[0146] For example, if the first loss value is greater than or equal to the first preset threshold, the terminal adjusts the model parameters of the bone segmentation model to be trained according to the first loss value, and retrains the adjusted bone segmentation model until the first loss value calculated from the predicted bone segmentation region output by the trained bone segmentation model is less than the first preset threshold. Then, the trained bone segmentation model is taken as the trained bone segmentation model.

[0147] In this embodiment, the bone segmentation model to be trained is repeatedly trained based on the sample site image and the labeled cancellous bone region in the sample site image. This helps to improve the accuracy of the cancellous bone region output by the trained bone segmentation model, and further improves the segmentation accuracy of the cancellous bone region.

[0148] In one exemplary embodiment, the cancellous bone region includes hydroxyapatite and other substances, wherein the other substances are substances in the cancellous bone region other than hydroxyapatite, and the other substances include at least water; then, as Figure 6 As shown, step S204 above, which determines the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error, specifically includes the following steps S601 to S603, wherein:

[0149] Step S601: Obtain the intrinsic density of the reference substance and the actual concentrations of other substances.

[0150] Step S602: Input the actual CT value of the cancellous bone region, the current effective tube voltage, the current system error, the intrinsic density of the reference material, and the actual concentrations of other materials into the volumetric bone mineral density measurement model to obtain the actual concentration of hydroxyapatite.

[0151] Step S603: The actual concentration of hydroxyapatite is confirmed as the volumetric bone density of the cancellous bone region.

[0152] Further, in step S602 above, the actual CT value of the cancellous bone region, the current effective tube voltage, the current systematic error, the intrinsic density of the reference material, and the actual concentrations of other substances are input into the volumetric bone mineral density measurement model to obtain the actual concentration of hydroxyapatite. Specifically, this includes: determining the mass attenuation coefficient of the reference material, the mass attenuation coefficient of hydroxyapatite, and the mass attenuation coefficients of other substances based on the current effective tube voltage; and inputting the actual CT value of the cancellous bone region, the current systematic error, the mass attenuation coefficient and intrinsic density of the reference material, the mass attenuation coefficient and actual concentration of other substances, and the mass attenuation coefficient of hydroxyapatite into the volumetric bone mineral density measurement model to obtain the actual concentration of hydroxyapatite.

[0153] The reference material is generally water.

[0154] Other substances refer to components in the spongy bone region other than HA (hyaluronic acid), such as water, collagen, silicates (e.g., silicon dioxide, silicates), calcium phosphates (e.g., calcium phosphate, calcium carbonate), and inorganic salts (e.g., sodium, potassium, magnesium). In practice, the spongy bone region can be approximated as a mixture of HA and water.

[0155] The mass decay coefficients of the reference material, HA, and other materials can all be obtained by querying the second database. Furthermore, given the current effective tube voltage, the mass decay coefficients of the reference material, HA, and other materials can all be considered constants.

[0156] The intrinsic density of the reference substance is constant and can be obtained by querying the first database. Since the actual concentrations of other substances (such as water) in a population can be considered essentially constant, and populations can be categorized by age, gender, etc., the actual concentrations of other substances in the spongy bone region can be determined based on the population to which the test subject belongs. For example, the third database stores the actual concentrations of other substances in different populations; by querying the third database, the actual concentrations of other substances in the population to which the test subject belongs can be obtained, thus determining the actual concentrations of other substances.

[0157] Among them, the volumetric bone mineral density measurement model is a mathematical model used to statistically measure volumetric bone mineral density, which can be obtained through formula 3.

[0158] For example, based on the current effective tube voltage, the terminal queries the second database for the correspondence between the effective tube voltage and the mass decay coefficient of the reference substance, the correspondence between the effective tube voltage and the mass decay coefficient of HA, and the correspondence between the effective tube voltage and the mass decay coefficient of other substances, to obtain the mass decay coefficient of the reference substance, the mass decay coefficient of HA, and the mass decay coefficient of other substances. Next, the terminal queries the first database, which stores the intrinsic densities of multiple substances, to obtain the intrinsic density of the reference substance; simultaneously, based on the age and gender of the test subject, it determines the population to which the test subject belongs, such as young women, middle-aged men, and elderly men; and queries the third database, which stores the actual concentrations of other substances in different populations, to obtain the actual concentration of other substances (such as water) in the population to which the test subject belongs, which is used as the actual concentration of other substances in the spongy bone region. Then, the terminal inputs the actual CT value of the cancellous bone region, the current systematic error, the mass decay coefficient and intrinsic density of the reference material, the mass decay coefficient and actual concentration of other substances, and the mass decay coefficient of HA into the volumetric bone mineral density measurement model to obtain the actual concentration of HA. For example, the actual concentration of HA = [((actual CT value of the cancellous bone region - current systematic error) / 1000+1) × mass decay coefficient of the reference material at the current effective tube voltage × intrinsic density of the reference material - mass decay coefficient of other substances at the current effective tube voltage × actual concentration of other substances] / (mass decay coefficient of HA at the current effective tube voltage). Finally, the terminal confirms the actual concentration of HA as the volumetric bone mineral density of the cancellous bone region.

[0159] For example, assuming the vertebral body cancellous bone is a mixture of HA and water, and water is the reference substance, the volumetric bone mineral density of the cancellous bone region is calculated as follows:

[0160]

[0161] Wherein, VBMD represents the volumetric bone mineral density of the cancellous bone region, ρ HA ' represents the actual concentration of HA in the cancellous bone region, i.e., the content of HA in the cancellous bone region; CT value represents the actual CT value of the cancellous bone region; ε represents the current systematic error of the current scanner; f H2O (KV e This indicates the current effective tube voltage (KV) of the water in the current scanner. e The mass attenuation coefficient, ρ H2O f represents the intrinsic density of water. i (KV e ) indicates that other substances i are at the current effective tube voltage KV e The mass attenuation coefficient, ρ i 'Indicates the actual concentration of other substance i; f HA (KVe ) indicates that HA is at the current effective tube voltage KV e The mass decay coefficient.

[0162] In this embodiment, the actual concentration of hydroxyapatite is calculated based on the actual CT value of the cancellous bone region, the current systematic error, the mass decay coefficient and intrinsic density of the reference material, the mass decay coefficient and actual concentration of other substances, and the mass decay coefficient of hydroxyapatite. Since the mass decay coefficients of the reference material, other substances, and hydroxyapatite are related to the current effective tube voltage, and the current effective tube voltage and current systematic error are values ​​after self-calibration without a phantom, and the actual CT value of the cancellous bone region is a precise value determined by the model, the final determined volumetric bone density of the cancellous bone region is more accurate, further improving the measurement accuracy of volumetric bone density.

[0163] In an exemplary embodiment, step S202 above, which identifies the internal reference region image from the scanned image, specifically includes the following: performing a second preprocessing on the scanned image to obtain a preprocessed scanned image; inputting the preprocessed scanned image into a pre-trained internal reference segmentation model for segmentation processing to obtain the internal reference region image.

[0164] The second preprocessing step involves resolution normalization, size normalization, and grayscale normalization, which ensures that scanned images have the same resolution, size, and similar grayscale values. A preprocessed scanned image is the image obtained after performing this second preprocessing.

[0165] Among them, the pre-trained internal reference segmentation model is a deep learning model used to segment the region image of the internal reference, which is obtained through multiple training sessions.

[0166] For example, the terminal performs a second preprocessing on the scanned image to obtain a preprocessed scanned image, such as performing resolution normalization, size normalization, and grayscale value normalization on the scanned image. Then, the preprocessed scanned image is input into a pre-trained internal reference segmentation model, and the pre-trained internal reference segmentation model performs internal reference region segmentation on the preprocessed scanned image to obtain internal reference region images, such as fat region images, muscle region images, spleen region images, etc.

[0167] Furthermore, the pre-trained internal reference segmentation model is trained as follows: The terminal acquires a sample scan image and labeled internal reference regions within the sample scan image; wherein, the labeled internal reference regions refer to manually labeled internal reference regions in the sample scan image, that is, manually labeled internal reference regions are used as the gold standard, and each labeled internal reference region has a unique label (e.g., 1, 2, 3) corresponding to it. Next, the terminal performs a second preprocessing on the sample scan image to obtain a preprocessed sample scan image; the preprocessed sample scan image is input into the internal reference segmentation model to be trained (e.g., the VB-Net model) for segmentation processing to obtain predicted internal reference regions in the preprocessed sample scan image; based on the difference between the predicted internal reference regions and the labeled internal reference regions, the internal reference segmentation model to be trained is iteratively trained to obtain a trained internal reference segmentation model, which serves as the pre-trained internal reference segmentation model.

[0168] For example, the terminal retrieves sample scan images and labeled internal reference regions from a local database; performs a second preprocessing on the sample scan images to obtain preprocessed sample scan images; inputs the preprocessed sample scan images into the internal reference segmentation model to be trained for segmentation processing to obtain predicted internal reference regions in the preprocessed sample scan images; calculates a second loss value based on the difference between the predicted internal reference regions and the labeled internal reference regions; if the second loss value is greater than or equal to a second preset threshold, adjusts the model parameters of the internal reference segmentation model to be trained based on the second loss value, and retrains the adjusted internal reference segmentation model until the second loss value calculated based on the predicted internal reference regions output by the trained internal reference segmentation model is less than the second preset threshold. Then, the trained internal reference segmentation model is considered the completed internal reference segmentation model and is confirmed as the pre-trained internal reference segmentation model.

[0169] In this embodiment, the scanned image is first preprocessed, and then the preprocessed scanned image is segmented using a pre-trained internal reference segmentation model to obtain an internal reference region image. In this way, after preprocessing the scanned image, the internal reference region image is segmented from the preprocessed scanned image using the pre-trained internal reference segmentation model, which makes the segmented internal reference region image more accurate and thus improves the segmentation accuracy of the internal reference region image.

[0170] In an exemplary embodiment, step S202 above, obtaining the theoretical CT value of the internal reference, specifically includes the following: obtaining the intrinsic density of the internal reference and the intrinsic density of the reference material; inputting the predicted effective tube voltage and predicted system error of the current scanner, as well as the intrinsic density of the internal reference and the intrinsic density of the reference material, into the theoretical CT value statistical model to obtain the theoretical CT value of the internal reference.

[0171] The intrinsic density of the internal reference is constant and can be obtained by querying the first database. The intrinsic density of the reference substance (such as water) is also constant and can be obtained by querying the first database.

[0172] Among them, the theoretical CT value statistical model is a mathematical model used to statistically calculate theoretical CT values ​​(such as theoretical CT values ​​for internal reference), which can be obtained through formula 4.

[0173] For example, the terminal first queries a first database storing the intrinsic densities of multiple substances to obtain the intrinsic density of an internal reference (e.g., fat) and a reference substance (e.g., water). Next, based on the current scanner's predicted effective tube voltage, the terminal queries a second database for the correspondence between the effective tube voltage and the mass decay coefficient of the internal reference, and the correspondence between the effective tube voltage and the mass decay coefficient of the reference substance, to obtain the mass decay coefficient of the internal reference and the mass decay coefficient of the reference substance. Finally, the terminal inputs the mass decay coefficient and intrinsic density of the internal reference, the mass decay coefficient and intrinsic density of the reference substance, and the current scanner's prediction system error into the theoretical CT value statistical model to obtain the theoretical CT value of the internal reference. For example, the theoretical CT value of the internal reference = [((mass decay coefficient of the internal reference × intrinsic density of the internal reference) / (mass decay coefficient of the reference substance × intrinsic density of the reference substance)) - 1] × 1000 + the current scanner's prediction system error.

[0174] For example, assuming the reference substance is water, the theoretical CT value of the internal reference can be calculated using Formula 4.

[0175] In this embodiment, the predicted effective tube voltage and prediction system error of the current scanner, as well as the intrinsic density of the internal reference and the intrinsic density of the reference material, are input into the statistical model of the theoretical CT value to obtain the theoretical CT value of the internal reference. In this way, by comprehensively considering the predicted effective tube voltage and prediction system error of the current scanner, as well as the intrinsic density of the internal reference and the intrinsic density of the reference material, it is beneficial to improve the accuracy of the determination of the theoretical CT value.

[0176] In one exemplary embodiment, such as Figure 7As shown, another method for measuring volumetric bone mineral density is provided. Taking the application of this method to a terminal as an example, the method specifically includes the following steps S701 to S711, wherein:

[0177] Step S701: Obtain the scanned image of the object to be tested captured by the current scanner.

[0178] Step S702: Perform a second preprocessing on the scanned image to obtain a preprocessed scanned image; input the preprocessed scanned image into a pre-trained internal reference segmentation model for segmentation processing to obtain an internal reference region image.

[0179] Step S703: Determine the actual CT value of the internal reference based on the regional image.

[0180] Step S704: Obtain the theoretical CT value of the internal reference and construct a loss model; the loss model is used to evaluate the overall error between the theoretical CT value of the internal reference and the actual CT value of the internal reference.

[0181] Step S705: Update the output value of the loss model based on the current predicted effective tube voltage and predicted system error of the scanner until the output value of the loss model meets the preset conditions.

[0182] Step S706: The predicted effective transistor voltage and predicted system error when the output value of the loss model meets the preset conditions are used as the current effective transistor voltage and current system error.

[0183] Step S707: Obtain the part image of the part to be tested in the scanned image; perform the first preprocessing on the part image to obtain the preprocessed part image.

[0184] Step S708: Input the preprocessed area image into the pre-trained cancellous bone segmentation model to obtain the cancellous bone region in the preprocessed area image; determine the actual CT value of the cancellous bone region based on the gray value of the cancellous bone region.

[0185] Step S709: Obtain the intrinsic density of the reference substance and the actual concentration of other substances. Based on the current effective tube voltage, determine the mass decay coefficient of the reference substance, the mass decay coefficient of hydroxyapatite, and the mass decay coefficients of other substances.

[0186] Step S710: Input the actual CT value of the cancellous bone region, the current systematic error, the mass decay coefficient and intrinsic density of the reference material, the mass decay coefficient and actual concentration of other materials, and the mass decay coefficient of hydroxyapatite into the volumetric bone density measurement model to obtain the actual concentration of hydroxyapatite.

[0187] Step S711: The actual concentration of hydroxyapatite is confirmed as the volumetric bone density of the cancellous bone region.

[0188] In the above-mentioned method for measuring volumetric bone mineral density, when measuring the volumetric bone mineral density of the cancellous bone region of the test site, the current effective tube voltage and current system error of the scanner can be determined by using only the overall error between the theoretical CT value and the actual CT value of the internal reference. At the same time, combined with the actual CT value of the cancellous bone region of the test site, the volumetric bone mineral density of the cancellous bone region can be calculated. The entire measurement process uses only the internal reference and does not require the creation of external references such as standard phantoms. That is, the internal reference is used instead of the standard phantom to calibrate the current scanner for phantom-free self-calibration, thereby simplifying the volumetric bone mineral density measurement process and improving the measurement efficiency.

[0189] In an exemplary embodiment, to more clearly illustrate the volumetric bone mineral density measurement method provided in this application, the following specific embodiment will be used to describe the volumetric bone mineral density measurement method. In one embodiment, referring to... Figure 8 This application also provides a phantom-free, self-calibrated quantitative computed tomography (QCT) method for determining vertebral volumetric bone mineral density (VBMD) based on multi-parameter optimization. This method uses substances with relatively small density differences among individuals, such as air, human fat, and human spleen, as references to replace the standard VBMD phantom for calibrating the current CT scanner, thereby obtaining the effective tube voltage KVe and systematic error ∈ of the current CT scanner. Based on the scan images of the subject acquired by the current CT scanner, the actual CT value of the cancellous vertebral bone region of the subject is measured. Finally, using the effective tube voltage KVe, systematic error ∈, and the actual CT value of the cancellous vertebral bone region of the current CT scanner, the VBMD of the cancellous vertebral bone region is calculated. This method does not require the creation of a standard phantom and performs scanner calibration on the scan images of each subject. It has similar accuracy to existing QCT schemes that rely on standard phantoms, and the operation is simpler. Specifically, it includes the following:

[0190] (1) Acquire CT images of the subject using a CT scanner.

[0191] (2) Using a pre-trained internal reference segmentation model, the region images of internal references such as fat, muscle, and spleen are identified from CT images, and the actual CT values ​​of internal references such as fat, muscle, and spleen are calculated based on the region images of internal references such as fat, muscle, and spleen.

[0192] (3) According to Formula 4, the theoretical CT value of the internal reference is calculated, and the effective tube voltage KVe and system error ∈ of the current CT scanner are estimated based on the overall error between the theoretical CT value of the internal reference and the actual CT value, so as to achieve the purpose of self-calibration without phantom.

[0193] (4) Using a pre-trained cancellous bone segmentation model, the vertebral cancellous bone region is segmented from the CT image, and the actual CT value of the vertebral cancellous bone region is calculated.

[0194] (5) Input the calculated effective tube voltage KVe and system error ∈ of the current CT scanner, and the actual CT value of the vertebral cancellous bone region into the VBMD calculation formula (e.g., formula 6) to calculate the VBMD of the vertebral cancellous bone region.

[0195] The above embodiments can achieve the following technical effects: (1) Based on the construction of an internal reference database including common internal reference tissues such as fat and muscle, tissues with relatively stable density such as spleen and blood are added as internal references, which is conducive to broadening the selection range of internal references and can flexibly combine the selection of internal references according to the actual situation of the test subject. (2) Based on the basic principle of CT imaging, and using the multi-parameter optimization method, each CT image (such as the scan image of the test subject) is independently calibrated based only on the internal reference in the CT image, without using a linear regression model with other VBMD and its matching CT image data as prior knowledge, and without imposing any restrictions on the tube voltage of the CT image. (3) Internal references, vertebral bone cancellous CT values, etc. are all automatically determined by deep learning networks, which is convenient to operate and the results are reliable. (4) Compared with QCT measurement of VBMD based on phantoms, this application uses common substances or tissues as internal references, without the need to order standard phantoms, which is low in cost and has the same accuracy as QCT measurement of VBMD based on phantoms. (5) Compared with the currently commonly used phantom-free self-calibrated QCT method for determining VBMD (i.e., calculating VBMD based on a linear regression model constructed from a set of previously matched VBMD and CT data with fixed coefficients), this application has a wider range of applications and fewer restrictions on the selection of internal references and CT image tube voltages; moreover, it does not require prior accumulation of VBMD and its matching CT image data and can use it as prior knowledge as needed; at the same time, it uses deep learning networks to achieve rapid internal reference and vertebral bone cancellous CT value determination, which is both simple and convenient.

[0196] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0197] Based on the same inventive concept, this application also provides a volumetric bone mineral density measuring device for implementing the aforementioned volumetric bone mineral density measuring method. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations of one or more embodiments of the volumetric bone mineral density measuring device provided below can be found in the limitations of the volumetric bone mineral density measuring method described above, and will not be repeated here.

[0198] In one exemplary embodiment, such as Figure 9 As shown, a volumetric bone density measuring device is provided, comprising: an image acquisition module 901, an image recognition module 902, an error processing module 903, and a density determination module 904, wherein:

[0199] The image acquisition module 901 is used to acquire the scanned image of the object to be tested captured by the current scanner.

[0200] The image recognition module 902 is used to identify the internal reference region image from the scanned image and determine the actual CT value of the internal reference based on the region image.

[0201] Error processing module 903 is used to obtain the theoretical CT value of the internal reference, and determine the current effective tube voltage and current system error of the current scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference.

[0202] The density determination module 904 is used to obtain the actual CT value of the cancellous bone region of the test site in the scanned image, and determine the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage and the current system error.

[0203] In an exemplary embodiment, the error processing module 903 is further configured to construct a loss model; the loss model is used to evaluate the overall error between the theoretical CT value of the internal reference and the actual CT value of the internal reference; the output value of the loss model is updated according to the current scanner's predicted effective tube voltage and predicted system error until the output value of the loss model meets a preset condition; the predicted effective tube voltage and predicted system error when the output value of the loss model meets the preset condition are used as the current effective tube voltage and current system error.

[0204] In an exemplary embodiment, the density determination module 904 is further configured to acquire a site image of the site to be tested in the scanned image; input the site image into a pre-trained cancellous bone segmentation model to obtain the cancellous bone region in the site image; and determine the actual CT value of the cancellous bone region based on the gray value of the cancellous bone region.

[0205] In an exemplary embodiment, the volumetric bone mineral density measurement device provided in this application further includes a model training module, used to acquire a sample site image and a labeled cancellous bone region in the sample site image; input the sample site image into a cancellous bone segmentation model to be trained to obtain a predicted cancellous bone region in the sample site image; and iteratively train the cancellous bone segmentation model to be trained based on the difference between the predicted cancellous bone region and the labeled cancellous bone region to obtain a trained cancellous bone segmentation model, which serves as a pre-trained cancellous bone segmentation model.

[0206] In an exemplary embodiment, the cancellous bone region includes hydroxyapatite and other substances, wherein the other substances are substances in the cancellous bone region other than hydroxyapatite, and the other substances include at least water; the density determination module 904 is also used to obtain the intrinsic density of the reference substance and the actual concentration of the other substances; input the actual CT value of the cancellous bone region, the current effective tube voltage, the current system error, the intrinsic density of the reference substance and the actual concentration of the other substances into the volumetric bone density measurement model to obtain the actual concentration of hydroxyapatite; and confirm the actual concentration of hydroxyapatite as the volumetric bone density of the cancellous bone region.

[0207] In an exemplary embodiment, the density determination module 904 is further configured to determine the mass decay coefficient of the reference material, the mass decay coefficient of hydroxyapatite, and the mass decay coefficients of other materials based on the current effective tube voltage; and input the actual CT value of the cancellous bone region, the current system error, the mass decay coefficient and intrinsic density of the reference material, the mass decay coefficient and actual concentration of other materials, and the mass decay coefficient of hydroxyapatite into the volumetric bone density measurement model to obtain the actual concentration of hydroxyapatite.

[0208] In an exemplary embodiment, the image recognition module 902 is further configured to obtain the intrinsic density of the internal reference and the intrinsic density of the reference material; and input the predicted effective tube voltage and predicted system error of the current scanner, as well as the intrinsic density of the internal reference and the intrinsic density of the reference material, into the theoretical CT value statistical model to obtain the theoretical CT value of the internal reference.

[0209] Each module in the aforementioned volumetric bone mineral density measuring device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0210] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 10 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a method for measuring volumetric bone density. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0211] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0212] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0213] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0214] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0215] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0216] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0217] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0218] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for measuring volumetric bone mineral density, characterized in that, The method includes: Acquire the scanned image of the object to be tested captured by the current scanner; An internal reference region image is identified from the scanned image, and the actual CT value of the internal reference is determined based on the region image; The theoretical CT value of the internal reference is obtained. Based on the overall error between the theoretical CT value and the actual CT value of the internal reference, the current effective tube voltage and the current system error of the current scanner are determined. The overall error refers to the sum of the squares of the differences between the theoretical CT values ​​and the actual CT values ​​of each internal reference. The current effective tube voltage and the current system error refer to the estimated effective tube voltage and the estimated system error when the overall error between the theoretical CT value and the actual CT value of the internal reference is minimized. Obtain the actual CT value of the cancellous bone region of the test site in the scanned image, and determine the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error.

2. The method according to claim 1, characterized in that, The step of determining the current effective tube voltage and current system error of the current scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference includes: Construct a loss model; The output value of the loss model is updated based on the current scanner's predicted effective tube voltage and predicted system error until the output value of the loss model meets the preset conditions. The predicted effective transistor voltage and the predicted system error when the output value of the loss model satisfies the preset conditions are taken as the current effective transistor voltage and the current system error, respectively.

3. The method according to claim 1, characterized in that, The process of obtaining the actual CT value of the cancellous bone region of the site to be tested in the scanned image includes: Obtain a region image of the area to be tested from the scanned image; The image of the location is input into a pre-trained cancellous bone segmentation model to obtain the cancellous bone region in the image of the location. The actual CT value of the cancellous bone region is determined based on its grayscale value.

4. The method according to claim 3, characterized in that, The pre-trained cancellous bone segmentation model was trained in the following manner: Acquire images of the sample site and the labeled cancellous bone regions in the sample site images; The sample area image is input into the bone segmentation model to be trained to obtain the predicted bone region in the sample area image; Based on the difference between the predicted cancellous bone region and the labeled cancellous bone region, the cancellous bone segmentation model to be trained is iteratively trained to obtain a trained cancellous bone segmentation model, which is used as the pre-trained cancellous bone segmentation model.

5. The method according to claim 1, characterized in that, The cancellous bone region includes hydroxyapatite and other substances, wherein the other substances are substances in the cancellous bone region other than hydroxyapatite; The step of determining the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error includes: Obtain the intrinsic density of the reference substance and the actual concentrations of the other substances; The actual CT value of the cancellous bone region, the current effective tube voltage, the current system error, the intrinsic density of the reference material, and the actual concentration of the other materials are input into the volumetric bone density measurement model to obtain the actual concentration of hydroxyapatite. The actual concentration of the hydroxyapatite was confirmed as the volumetric bone density of the cancellous bone region.

6. The method according to claim 5, characterized in that, The process of inputting the actual CT value of the cancellous bone region, the current effective tube voltage, the current systematic error, the intrinsic density of the reference material, and the actual concentrations of other substances into the volumetric bone mineral density measurement model to obtain the actual concentration of hydroxyapatite includes: Based on the current effective tube voltage, determine the mass decay coefficient of the reference material, the mass decay coefficient of the hydroxyapatite, and the mass decay coefficients of the other materials; The actual CT value of the cancellous bone region, the current systematic error, the mass decay coefficient and intrinsic density of the reference substance, the mass decay coefficient and actual concentration of other substances, and the mass decay coefficient of hydroxyapatite are input into the volumetric bone density measurement model to obtain the actual concentration of hydroxyapatite.

7. The method according to any one of claims 1-6, characterized in that, The process of obtaining the theoretical CT value of the internal reference includes: Obtain the intrinsic density of the internal reference and the intrinsic density of the reference material; The predicted effective tube voltage and predicted system error of the current scanner, as well as the intrinsic density of the internal reference and the intrinsic density of the reference material, are input into the theoretical CT value statistical model to obtain the theoretical CT value of the internal reference.

8. A volumetric bone density measuring device, characterized in that, The device includes: The image acquisition module is used to acquire the scanned image of the object under test captured by the current scanner; An image recognition module is used to identify an internal reference region image from the scanned image and determine the actual CT value of the internal reference based on the region image; An error processing module is used to obtain the theoretical CT value of the internal reference, and determine the current effective tube voltage and current system error of the current scanner based on the overall error between the theoretical CT value and the actual CT value of the internal reference; the overall error refers to the sum of the squares of the differences between the theoretical CT value and the actual CT value of each internal reference; the current effective tube voltage and the current system error refer to the estimated effective tube voltage and the estimated system error when the overall error between the theoretical CT value and the actual CT value of the internal reference is minimized. The density determination module is used to obtain the actual CT value of the cancellous bone region of the test site in the scanned image, and determine the volumetric bone density of the cancellous bone region based on the actual CT value of the cancellous bone region, the current effective tube voltage, and the current system error.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-7.

Images