Artificial intelligence-based three-dimensional reconstruction method and system for meridians and collaterals of acupuncture

Abstract

The present application relates to the technical field of three-dimensional image reconstruction, and particularly relates to a three-dimensional reconstruction method and system for acupuncture meridians based on artificial intelligence, comprising: obtaining each target acupoint and tissue structure three-dimensional model in infrared image data of a user to be modeled; in the infrared image data, obtaining a temperature change curve of each pixel point in a stimulation point and a neighborhood range of each target acupoint, and determining a surface meridian distribution of each target acupoint according to the temperature change curve to analyze temperature rising; determining a deep meridian distribution according to each edge line determined in the tissue structure three-dimensional model and the surface meridian distribution of each target acupoint; and constructing a three-dimensional meridian model of the user to be modeled in combination with the surface meridian distribution and the deep meridian distribution. The present application reconstructs a three-dimensional model by determining the surface meridian distribution and the deep meridian distribution, and effectively improves the accuracy of three-dimensional reconstruction of acupuncture meridians.

Description

Technical Field

[0001] This invention relates to the field of three-dimensional image reconstruction technology, specifically to a method and system for three-dimensional reconstruction of acupuncture meridians based on artificial intelligence. Background Technology

[0002] Three-dimensional models of acupuncture meridians can be used to analyze the anatomical structure and functional characteristics of human meridians, providing precise anatomical evidence for acupuncture treatment and assisting in diagnosis and treatment plan formulation. Therefore, constructing accurate and reliable three-dimensional models of acupuncture meridians is of paramount importance.

[0003] Currently, when constructing a three-dimensional model of acupuncture meridians, the model is obtained by fitting the meridian pathways at different locations. However, existing multimodal imaging technologies, such as MRI (Magnetic Resonance Imaging), infrared, and CT (Computed Tomography), have gaps in data capture between the surface and deep layers of meridians and lack individualized dynamic functional imaging capabilities, making it impossible to construct a three-dimensional model that accurately reflects the anatomical structure and functional state of the meridians. Summary of the Invention

[0004] To address the technical problem of incomplete existing three-dimensional models of acupuncture meridians, the present invention aims to provide an artificial intelligence-based method and system for three-dimensional reconstruction of acupuncture meridians. The specific technical solution adopted is as follows:

[0005] One embodiment of the present invention provides a method for three-dimensional reconstruction of acupuncture meridians based on artificial intelligence, the method comprising the following steps:

[0006] Infrared image data of the user to be modeled is acquired, and each target acupoint in the three-dimensional model of the human body tissue structure of the user to be modeled is acquired, thereby determining each target acupoint in the infrared image data; wherein, the target acupoint is a specific acupoint marked on the skin surface;

[0007] In the infrared image data, the temperature change curves of each target acupoint stimulation point and each pixel in its neighborhood are obtained, and the surface meridian distribution of each target acupoint is determined by analyzing the temperature rise based on the temperature change curves.

[0008] In the three-dimensional model of the tissue structure, the surface meridian distribution of each edge line and each target acupoint is determined, and the deep meridian distribution is determined based on the surface meridian distribution of each edge line and each target acupoint.

[0009] A three-dimensional meridian model of the user to be modeled is constructed by combining the surface meridian distribution and the deep meridian distribution.

[0010] Further, determining each of the target acupoints in the infrared image data includes:

[0011] Image registration is performed based on infrared images and standard human acupoint diagrams, so that each standard acupoint in the standard human acupoint diagram is mapped onto the infrared image, thereby obtaining each candidate acupoint in the infrared image.

[0012] The candidate acupoints in the infrared image are matched with the three-dimensional model of the tissue structure to determine the target acupoints in the infrared image data.

[0013] Furthermore, the process of obtaining each candidate acupoint in the infrared image includes:

[0014] Obtain a standard human acupoint chart, and scale the standard human acupoint chart using a Gaussian pyramid to obtain standard human acupoint charts of different scales.

[0015] Extract each matching point pair between the infrared image and the standard human acupoint map at different scales, and calculate the affine transformation matrix between the infrared image and the standard human acupoint map based on the matching point pairs;

[0016] The standard human acupoint diagram is transformed using the affine transformation matrix to obtain a transformed image, and the transformed image is resampled to obtain an infrared image aligned to the coordinate system and a standard human acupoint diagram of a selected scale.

[0017] By mapping each acupoint in the standard human acupoint chart of a selected scale onto the infrared image, each candidate acupoint in the infrared image is obtained.

[0018] Furthermore, the acquisition of the temperature change curves of each stimulation point and its neighborhood pixels for each target acupoint includes:

[0019] For each target acupoint, the temperature data sequence of each pixel point in the stimulation point and its neighborhood range within a preset time period is obtained by stimulating the target acupoint.

[0020] Based on the temperature data sequence of each pixel in the stimulation point and its neighborhood within a preset time period, a temperature change curve of each pixel in the stimulation point and its neighborhood is constructed.

[0021] Furthermore, the step of determining the surface meridian distribution of each target acupoint by analyzing the temperature rise based on the temperature change curve includes:

[0022] For each target acupoint, the degree of temperature rise at each time point within the preset time period is determined based on the temperature change curve.

[0023] The temperature rise at each time point is compared with a preset temperature rise threshold to obtain the time points of temperature change.

[0024] The temperature change time intervals are obtained by considering the temperature change time differences between the stimulation point and each pixel in its neighborhood, and the temperature diffusion direction at the location of the stimulation point of the target acupoint is determined based on the temperature change time intervals.

[0025] Based on the temperature diffusion direction at the location of the stimulation point, a region growth analysis is performed on the infrared image to obtain the surface meridian distribution of the target acupoint.

[0026] Further, determining the degree of temperature rise at each time point within the preset time period based on the temperature change curve includes:

[0027] For any given location, obtain the temperature difference between the initial temperature and the temperature at the current time point, and obtain the slope of the temperature change curve at the current time point; wherein, the location point is the stimulus point and each pixel point within its neighborhood.

[0028] By combining the temperature difference value and the slope, the degree of temperature rise at the current time point is determined; wherein, both the temperature difference value and the slope are positively correlated with the degree of temperature rise.

[0029] Furthermore, after determining the superficial meridian distribution of each target acupoint, the process includes:

[0030] When there are breaks in the surface meridian distribution, a human meridian acupoint map is obtained. Based on the meridian connection information between acupoints in the human meridian acupoint map, spline difference fitting is performed on the surface meridian distribution with breaks to obtain the fitted surface meridian distribution.

[0031] Further, determining the deep meridian distribution based on the surface meridian distribution of each edge line and each target acupoint includes:

[0032] The meridian evaluation value of each edge line is determined based on its length, density, disorder, and distance from the distribution of each surface meridian.

[0033] Obtain the edge lines between two target acupoints, and determine the connection direction between the two target acupoints by using the edge line corresponding to the maximum meridian evaluation value between the two target acupoints.

[0034] The distribution of deep meridians is obtained on a three-dimensional model of the structure based on the connection direction of two target acupoints.

[0035] Furthermore, determining the meridian evaluation value of each edge line based on its length, density, disorder, and distance from the distribution of surface meridians includes:

[0036] For each edge line, determine the length of the edge line and count the number of edge lines within the neighborhood.

[0037] Determine the distance between the edge line and each surface meridian distribution, and calculate the slope variance of the edge line;

[0038] By integrating the length of the edge line, the number of edge lines, the distance, and the slope variance, a meridian evaluation value for the edge line is obtained.

[0039] The meridian evaluation value is positively correlated with the length and the number of edge lines, and negatively correlated with the distance and the slope variance.

[0040] Another embodiment of the present invention provides an artificial intelligence-based three-dimensional reconstruction system for acupuncture meridians, including a processor and a memory, wherein the processor is used to process instructions stored in the memory to implement an artificial intelligence-based three-dimensional reconstruction method for acupuncture meridians.

[0041] The present invention has the following beneficial effects:

[0042] This invention first acquires infrared image data to analyze temperature changes and diffusion at acupoints. The direction and manner of this diffusion correspond to the meridian pathways in Traditional Chinese Medicine (TCM) theory. Therefore, by analyzing infrared image data, the path of temperature changes can be inferred, indirectly indicating the direction of meridians, thus analyzing the distribution of superficial meridians. Next, a three-dimensional model of the tissue structure is acquired. When the meridian pathways overlap with specific organs, muscle groups, neural networks, etc., the three-dimensional model can reveal the three-dimensional spatial distribution of the relevant structures, thereby analyzing the distribution of deep meridians. The superficial and deep meridian distributions are then fused for matching and connection analysis. Subsequently, based on the matched and connected human meridian distribution, a three-dimensional model is reconstructed, effectively improving the accuracy of three-dimensional reconstruction of acupuncture meridians. Attached Figure Description

[0043] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 A flowchart illustrating the steps of an artificial intelligence-based three-dimensional reconstruction method for acupuncture meridians, as provided in one embodiment of the present invention;

[0045] Figure 2 This is a flowchart illustrating the steps involved in determining the distribution of surface meridians in an embodiment of the present invention.

[0046] Figure 3 This is an example diagram of the surface meridian distribution in an embodiment of the present invention;

[0047] Figure 4 This is a flowchart illustrating the steps involved in determining the distribution of deep meridians in an embodiment of the present invention. Detailed Implementation

[0048] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the specific implementation methods, structures, features, and effects of the technical solution proposed according to the present invention are described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0049] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0050] The purpose of this invention is to: determine the distribution of meridian information on the surface by acquiring the temperature information of the patient's body, match the acquired meridian data according to the connection between the deep and surface information of the meridians at the same location, and reconstruct the three-dimensional model based on the acquired matching information, so as to improve the construction accuracy of the three-dimensional model of acupuncture meridians.

[0051] One embodiment of the present invention provides a method for three-dimensional reconstruction of acupuncture meridians based on artificial intelligence, such as... Figure 1 As shown, it includes the following steps:

[0052] S1. Acquire infrared image data of the user to be modeled, and acquire each target acupoint in the three-dimensional model of the human body tissue structure of the user to be modeled, thereby determining each target acupoint in the infrared image data.

[0053] Here, the target acupoints are specific acupoints marked on the skin surface. The human meridians are obtained through the connections between acupoints. Therefore, in order to obtain a three-dimensional meridian model, it is necessary to analyze the characteristics of human acupoints.

[0054] As an exemplary implementation, step S1 can be achieved through steps S11 to S13 (not shown in the figure), as follows:

[0055] S11, acquire infrared image data and human body slice data of the user to be modeled.

[0056] Here, the users to be modeled refer to the population that needs to undergo three-dimensional reconstruction of acupuncture meridians; infrared image data can be used to observe temperature changes and diffusion characteristics at acupoints, and its diffusion characteristics are consistent with the meridian paths in traditional Chinese medicine theory. In other words, infrared image data can infer the path of temperature changes and indirectly indicate the direction of meridians. Therefore, infrared image data is used as data support for constructing a three-dimensional meridian model; when the meridian path overlaps with specific organs, muscle groups, neural networks, etc., the slice data here can reveal the three-dimensional spatial distribution of these related structures. Therefore, human slice data is also used as data support for constructing a three-dimensional meridian model.

[0057] As a specific example of acquiring infrared image data: Infrared image data is obtained by collecting temperature data of the user's body at different locations using a high-precision infrared sensor. The infrared sensor can be an infrared thermal imaging instrument.

[0058] It should be noted that when acquiring temperature data, a constant temperature should be maintained, direct wind should be avoided, and direct sunlight should be avoided from shining on the infrared thermal imaging instrument; the thermal sensitivity should be 0.05℃ or higher resolution, and the user should sit quietly for 5-10 minutes after removing some clothing before starting to collect infrared image data.

[0059] For example, to obtain human body slice data, the human body of the user to be modeled is scanned by CT or MRI to obtain human body slice data.

[0060] S12: Obtain the target acupoints in the three-dimensional model of the tissue structure of the user's body to be modeled based on the human body slice data.

[0061] As a specific example, firstly, human body slice data is input into artificial intelligence for subsequent processing, constructing a corresponding three-dimensional model of the tissue structure from the acquired slice data. The process from slice to three-dimensional model involves preprocessing, registration, segmentation, reconstruction, and post-processing. The specific implementation process of constructing the three-dimensional model of the tissue structure is existing technology and is not within the scope of this invention; therefore, it will not be described in detail here.

[0062] Then, nitroglycerin capsule patches are used to mark specific acupoints on the skin surface, which means marking specific acupoints in the three-dimensional tissue structure model, thus identifying the target acupoints. Subsequently, the anatomical structures related to the meridians at the target acupoints can be obtained, that is, the tissue structure of the target acupoints can be obtained. A three-dimensional tissue structure model is generated for all target acupoints, with each acupoint being only a part of the three-dimensional tissue structure model.

[0063] For the slice data of target acupoints, such as the acupoints of the Dai Mai meridian (lumbar level) corresponding to the external and internal oblique muscles, and the acupoints of the Yin Qiao Mai meridian (medial malleolus side) corresponding to the abductor muscles of the foot, etc., there are no specific restrictions on the type and number of target acupoints. Implementers can set them according to the specific actual situation.

[0064] It should be noted that external markings can ensure a precise correspondence between MRI images and acupoints, avoiding positioning errors caused by individual differences and effectively enhancing the accuracy of target acupoint determination.

[0065] S13. Based on the infrared image data of the user to be modeled, and combined with the target acupoints in the three-dimensional model of the tissue structure, determine the target acupoints in the infrared image data.

[0066] Temperature changes at acupoints can reflect the state of Qi and blood circulation and disease conditions in the body. These temperature changes are also related to the conduction function of meridians; for example, discontinuity or shortening of the high-temperature zone in the Du meridian may indicate a loss of Qi and blood flow. Therefore, to analyze temperature changes at acupoints, it is necessary to identify each target acupoint in the infrared image data.

[0067] To determine individual target acupoints in infrared image data, a specific example is as follows:

[0068] The first step is to perform image registration based on the infrared image and the standard human acupoint map, so that each standard acupoint in the standard human acupoint map is mapped into the infrared image, and thus obtain each candidate acupoint in the infrared image.

[0069] Here, candidate acupoints refer to all acupoints in the human body, which are used to select target acupoints in the subsequent process.

[0070] Because standard acupoint charts share certain similarities with the user's own meridian pathways, image registration analysis can be performed between infrared images and standard acupoint charts. However, during image registration, differences in body size lead to scale discrepancies between the infrared images and standard acupoint charts. To overcome the impact of these size differences on the accuracy of the registration results, it is necessary to determine infrared images and standard acupoint charts with aligned coordinate systems. Based on this, candidate acupoints in the infrared image can then be identified.

[0071] The first sub-step is to obtain a standard human acupoint map, and then scale the standard human acupoint map using a Gaussian pyramid to obtain standard human acupoint maps at different scales.

[0072] In this embodiment, the number and size of scales can be set by the implementer according to the specific actual situation. The implementation process of Gaussian pyramid processing is existing technology and is not within the protection scope of this invention, so it will not be described in detail here.

[0073] It is worth noting that, in order to ensure the integrity of each acupoint in the standard human acupoint diagram, the scaling of the standard human acupoint diagram in this embodiment only changes the size, without any deletions.

[0074] The second sub-step involves extracting matching point pairs between the infrared image and standard human acupoint diagrams at different scales, and calculating the affine transformation matrix between the infrared image and the standard human acupoint diagram based on the matching point pairs.

[0075] In this embodiment, corresponding points between infrared images and standard human acupoint diagrams at any scale are extracted and matched using SIFT (Scale-Invariant Feature Transform) to obtain each matching point pair. The spatial transformation relationship between the two is analyzed based on the matching point pairs to determine the affine transformation matrix. The implementation process of the SIFT algorithm is existing technology and is not within the scope of this invention; therefore, it will not be described in detail here.

[0076] It is worth noting that when extracting matching point pairs, since infrared images rely on thermal radiation while acupoint images are in visible light, preprocessing is required to unify the grayscale space. This means converting the infrared image to a grayscale image for acupoints. Figure 2 Value-enhancing processing is used to improve the outline; acupoint diagrams are mostly line drawings, which may have missing textures. Feature points can be manually added at the intersection of acupoints or the turning points of meridians to overcome the texture loss defect.

[0077] The third sub-step involves transforming the standard human acupoint map using an affine transformation matrix to obtain a transformed image, and then resampling the transformed image to obtain an infrared image aligned to the coordinate system and a standard human acupoint map at a selected scale.

[0078] In this embodiment, the standard human acupoint map and the infrared image are aligned using an affine transformation matrix. After the affine transformation, the standard human acupoint map is resampled to ensure that the coordinate systems of the standard human acupoint map at the selected scale and the infrared image are aligned. The standard human acupoint map at the selected scale is one that overcomes the influence of differences in human body shape, meaning it is the standard human acupoint map that best matches the human body of the user to be modeled. The implementation processes of the affine transformation and resampling are existing technologies and are not within the scope of this invention; therefore, they will not be described in detail here.

[0079] It should be noted that both affine transformation and resampling are used to align the coordinate systems of the infrared image and the standard human acupoint diagram, so as to facilitate subsequent image matching.

[0080] The fourth sub-step involves mapping each acupoint in the standard human acupoint chart of the selected scale onto the infrared image to obtain each candidate acupoint in the infrared image.

[0081] In this embodiment, to analyze the meridian pathways in the human body, the standard human acupoint map at a selected scale and the infrared image are matched based on the relationships between acupoints. This means mapping each acupoint in the standard human acupoint map at the selected scale onto the infrared image, resulting in candidate acupoints in the infrared image. Candidate acupoints refer to the acupoints located in the infrared image.

[0082] The second step is to match each candidate acupoint in the infrared image with the three-dimensional model of the tissue structure to determine each target acupoint in the infrared image data.

[0083] In this embodiment, the location information of each candidate acupoint in the infrared image is matched with the three-dimensional model of the tissue structure to align some candidate acupoints in the infrared image with each target acupoint in the three-dimensional model of the tissue structure, ensuring data consistency. For example, a spatial correspondence is established based on the coordinate position of the target acupoint.

[0084] Thus, this embodiment has obtained the three-dimensional models of the tissue structures corresponding to all target acupoints and the infrared images of each target acupoint.

[0085] S2, in the infrared image data, acquire the temperature change curves of the stimulation points of each target acupoint and the pixels in their neighborhood, and determine the surface meridian distribution of each target acupoint by analyzing the temperature rise based on the temperature change curves.

[0086] Infrared thermography can capture temperature changes that are typically concentrated on the skin surface and superficial tissues. However, temperature changes in deeper tissues, especially where meridians extend deeper, are difficult to measure directly. The temperature signal from deep meridians weakens as it passes through the skin layer. Therefore, infrared images can only reflect the heat propagation patterns of some surface meridians, thus determining that the meridian distribution is also a superficial meridian distribution.

[0087] The infrared temperature intensity at acupoints is higher than that of surrounding tissues, averaging 0.5-1.0℃ higher. This indicates that the temperature at the center of the acupoint is higher than the surrounding temperature, demonstrating a temperature diffusion phenomenon between the center and the external temperature. Furthermore, to amplify this temperature difference, acupuncture or massage can be used to stimulate acupoints. Stimulation activates the sympathetic nervous system, dilates blood vessels, increases local blood flow, and raises the acupoint temperature by 1-2℃. For example, the significant change in infrared radiation intensity at acupoints after acupuncture reflects adjustments in energy metabolism and blood circulation, indicating a pronounced meridian diffusion phenomenon.

[0088] First, in the infrared image data, the temperature change curves of each target acupoint stimulation point and its neighboring pixels are obtained.

[0089] As a specific example, for each target acupoint, the temperature data sequence of each pixel in the stimulation point and its neighborhood within a preset time period is obtained by stimulating the target acupoint; based on the temperature data sequence of each pixel in the stimulation point and its neighborhood within the preset time period, the temperature change curve of each pixel in the stimulation point and its neighborhood is constructed.

[0090] In this embodiment, the preset time period can be 60 seconds, which can be set by the implementer according to the specific actual situation, and no specific limitation is made here; the temperature data sequence is a time series, which can be obtained by analyzing video data of temperature changes and temperature sensor acquisition; the target acupoints can be stimulated by acupuncture, and the acupuncture position with the best effect can be determined as the stimulation point; the neighborhood range here refers to the eight-neighborhood; the curve can be obtained by fitting with the least squares method, and the implementation process of the least squares method is the prior art; the horizontal axis of the temperature change curve is the time point, and the vertical axis is the temperature value (grayscale value).

[0091] Secondly, the surface meridian distribution of each target acupoint is determined by analyzing the temperature rise based on the temperature change curve.

[0092] The temperature difference between the acupoints and their surrounding tissues typically reflects the flow of Qi and blood, that is, temperature changes along traditional meridian pathways, specifically a decreasing trend in temperature along a particular pathway. After acupuncture or thermotherapy, infrared thermography can be used to observe temperature changes and diffusion at the acupoints, and the pattern of temperature diffusion may correspond to the meridian pathways in Traditional Chinese Medicine theory. Therefore, through infrared image data, the path of temperature change can be inferred, indirectly indicating the direction of the meridians.

[0093] As an exemplary implementation, the process of determining the above-mentioned surface meridian distribution can be achieved through... Figure 2 Steps S21 to S24 shown are implemented as follows:

[0094] S21, for each target acupoint, determine the degree of temperature rise at each time point within the preset time period based on the temperature change curve.

[0095] Here, the degree of temperature rise refers to the degree of temperature change over time after a target acupoint is stimulated.

[0096] In this embodiment, the degree of temperature rise is determined by analyzing the temperature change value and slope between each time point and the initial time point.

[0097] As a concrete example, for any given location, the temperature difference between the initial temperature and the current temperature is obtained, along with the slope of the temperature change curve at the current time point. The temperature difference and slope are then combined to determine the degree of temperature rise at the current time point. Here, the stimulus point is the stimulated pixel, such as an acupuncture point, and the location point is the stimulus point and all pixels within its neighborhood. Both the temperature difference and the slope are positively correlated with the degree of temperature rise.

[0098] As an example, the formula for calculating the degree of temperature rise at location point t at time t can be:

[0099] In the formula, This indicates the degree of temperature rise at location point t at time point t, where time points are any points within a preset time period. This represents the difference between the temperature value at time point t and the temperature value at the first time point (the initial stimulation time point). The slope of the temperature change curve at the t-th time point of the location point is represented by norm, which represents the linear normalization function.

[0100] Referring to the process for determining the degree of temperature rise at the t-th time point, the degree of temperature rise at each time point within the preset time period can be obtained.

[0101] S22, compare the temperature rise at each time point with the preset temperature rise threshold to obtain the time points of temperature change.

[0102] As a specific example, the temperature rise ranges from 0 to 1. The preset temperature rise threshold can be an empirical value of 0.68. Implementers can set the preset temperature rise threshold according to the specific situation. The time point when the temperature rise is greater than the preset temperature rise threshold of 0.68 and the temperature rise is the largest is recorded as the temperature change time point. The temperature change time point corresponding to each location point can be obtained.

[0103] It should be noted that when the number of temperature change time points for a certain pixel is 0, it indicates that the temperature expansion trend of that pixel in the neighborhood is not obvious, and it is unlikely to be the direction of temperature diffusion. Therefore, no further analysis of the temperature diffusion direction will be performed on it.

[0104] S23, obtain the temperature change time intervals based on the temperature change time differences between the stimulation point and each pixel in its neighborhood, and determine the temperature diffusion direction of the stimulation point of the target acupoint based on the temperature change time intervals.

[0105] As a specific example, obtain the temperature change time points corresponding to each pixel in the stimulation point and its neighborhood range; calculate the time interval between the temperature change time points of the stimulation point and the temperature change time points of each pixel in the neighborhood range, and record it as the temperature change time interval; set the time interval threshold, take an empirical value of 0.3, and take the direction of the pixel points whose temperature change time interval is less than the time interval threshold as the temperature diffusion direction of the current stimulation point.

[0106] The smaller the time interval between temperature changes, the better the temperature diffusion effect. Therefore, the direction of the pixel point with a temperature change time interval less than the time interval threshold is taken as the temperature diffusion direction of the current stimulus point.

[0107] S24. Based on the temperature diffusion direction at the location of the stimulation point, regional growth analysis is performed in the infrared image to obtain the surface meridian distribution of the target acupoint.

[0108] In this embodiment, when analyzing the temperature diffusion of the target acupoint, the acupuncture site can be used as the optimal effect point and as the seed point. Through a similar region growth method, the meridian pathways in different directions are obtained, constituting the surface meridian distribution of the target acupoint. An example diagram of the surface meridian distribution is shown below. Figure 3 As shown.

[0109] There are breaks in the distribution of surface meridians in some areas. Firstly, the distance between two locations is too far, which prevents the temperature from spreading properly, thus creating a break. Secondly, the meridians are deep-seated, and the tissue layer causes a significant drop in temperature. The purpose of analyzing the tissue structure of the slices is to overcome the defect that the temperature cannot penetrate to the surface.

[0110] Therefore, after determining the surface meridian distribution of each target acupoint, the process includes: when there is a break in the surface meridian distribution, obtaining a human meridian acupoint map, and based on the meridian connection information between acupoints in the human meridian acupoint map, performing spline difference fitting on the surface meridian distribution with a break to obtain the fitted surface meridian distribution.

[0111] In this embodiment, when faced with a broken circuit, it is necessary to analyze the known meridian information, namely the human meridian acupoint diagram, which is the standard human acupoint diagram. The meridian connection information between acupoints can be obtained through the human meridian acupoint diagram. For cases where the temperature cannot diffuse due to the large distance between two locations, spline difference fitting of the curve is performed to fit the broken circuit curve into a single curve, which serves as the final surface meridian distribution.

[0112] By referring to the process of obtaining the surface meridian distribution of any of the target acupoints mentioned above, the surface meridian distribution of each target acupoint can be obtained.

[0113] Thus, this embodiment has obtained the surface meridian distribution of each target acupoint in the infrared image.

[0114] S3. In the three-dimensional model of the tissue structure, determine the surface meridian distribution of each edge line and each target acupoint, and determine the deep meridian distribution based on the surface meridian distribution of each edge line and each target acupoint.

[0115] From the perspective of modern medicine, meridians do not necessarily correspond to specific anatomically visible structures. Slice data itself does not directly reveal "meridians" in the traditional sense because they do not constitute independently anatomically identifiable structures. Slice data is primarily used to present the density differences of the body's hard tissues (bones) and soft tissues, but it does not directly represent dynamic or functional aspects such as blood flow and energy channels (meridians). However, meridian-related anatomical structures (such as nerves and blood vessels) may indirectly reflect the paths of certain "meridians," especially when the meridian pathways overlap with specific organs, muscle groups, or neural networks. In such cases, slice data can reveal the three-dimensional spatial distribution of these related structures. Therefore, by combining the pathways of surface meridian data with known meridian distributions, analyzing three-dimensional models of tissue structures can determine their possible meridian pathways and reveal the distribution of deep meridians.

[0116] Furthermore, by analyzing the distribution of muscles, bones, fascia, and other tissues in MRI, and combining this with a pre-known standard meridian model, the connection path between two acupoints can be inferred. Based on theoretical support, the distribution of fascial connective tissue in MRI images can be compared with traditional meridian pathways, revealing a high degree of agreement. Therefore, the edge information of the tissues in MRI can be used to quantify the meridian pathway information at a depth. The theoretical support refers to the fact that the human monitoring system composed of connective tissue fascia is the anatomical basis of acupuncture therapy.

[0117] First, the surface meridian distribution of each edge line and each target acupoint is determined in the three-dimensional model of the tissue structure.

[0118] Here, since the nodules of the tissue may contain data on the direction of meridians, it is necessary to obtain the edge lines in order to analyze the situation where each edge is a meridian.

[0119] In this embodiment, 3D Canny edge detection is used to perform edge detection on the three-dimensional model of the tissue structure, thereby determining each edge line in the three-dimensional model of the tissue structure. The three-dimensional model of the tissue structure and each target acupoint in the infrared image exhibit a one-to-one correspondence, so the surface meridian distribution of each target acupoint can be determined in the three-dimensional model of the tissue structure. The implementation process of 3D Canny edge detection is existing technology and is not within the scope of protection of this invention; therefore, it will not be described in detail here.

[0120] Secondly, based on the surface meridian distribution of each edge line and each target acupoint, the deep meridian distribution is determined.

[0121] As an exemplary implementation, the process of determining the distribution of deep meridians described above can be achieved through... Figure 4 Steps S31 to S33 are implemented as follows:

[0122] S31. Determine the meridian evaluation value of each edge line based on its length, density, disorder, and distance from the distribution of each surface meridian.

[0123] Here, the meridian assessment value indicates the probability that the edge line belongs to the edge of the meridian. The higher the meridian assessment value, the greater the probability that the edge line belongs to the edge of the meridian.

[0124] The above-mentioned data obtained from the user's infrared image information constitutes a portion of the meridian pathway data, i.e., the surface meridian distribution. For deeper meridian pathway analysis, it is necessary to determine the meridian evaluation value using meridian data obtained from the surface fracture locations. Because the gradient characteristics of meridians are quite pronounced, the meridian distribution within the tissue structure is analyzed based on the gradient of the edge lines. Furthermore, the overlapping of multiple tissues leads to relatively dense edge lines, and these densely distributed edge lines exhibit a certain directionality. Based on this, the meridian evaluation value of the edge lines is determined.

[0125] As a specific example, step S31 above may include:

[0126] The first step is to determine the length of each edge line and count the number of edge lines within the neighborhood of each edge line.

[0127] In this embodiment, the length of the edge line is equivalent to the number of pixels on the edge line. A higher number of pixels indicates better edge line extensibility and a stronger conformity to meridian distribution characteristics. The number of edge lines within the surrounding area is also counted. The surrounding area can be... It can characterize the edge density around the edge line. The greater the number of edge lines, the stronger the directional consistency around the edge line, and the more it conforms to the distribution characteristics of meridians.

[0128] The second step is to determine the distance between the edge line and the distribution of each surface meridian, and to calculate the slope variance of the edge line.

[0129] In this embodiment, the closer the edge line is to the surface meridian distribution of a target acupoint, the more likely the edge line is to connect to the surface meridian, and the greater the possibility that it belongs to the deep meridian. The minimum distance is taken, so the minimum distance is negatively correlated with the meridian evaluation value. The larger the slope variance of the edge line, the more chaotic the direction of the edge line is, and the less likely it is to show a consistent direction. Since the meridian edge has good consistency, the slope variance is negatively correlated with the meridian evaluation value.

[0130] The third step is to integrate the length, number, distance, and slope variance of the edge lines to obtain the meridian evaluation value of the edge lines.

[0131] In this embodiment, the meridian evaluation value is positively correlated with length and number of edge lines, and negatively correlated with distance and slope variance.

[0132] As an example, the formula for calculating the meridian evaluation value of the j-th edge line can be:

[0133] In the formula, This represents the meridian evaluation value of the j-th edge line. This represents the length of the j-th edge line. Let represent the distance between the j-th edge line and each surface meridian distribution, min denotes the minimum value function, and n represent the number of edge lines within the neighborhood of the edge line. Let represent the slope variance of the j-th edge line.

[0134] Referring to the calculation process of the meridian evaluation value of the j-th edge line above, the meridian evaluation value of each edge line can be obtained.

[0135] S32, obtain each edge line located between two target acupoints, and determine the connection direction of the two target acupoints through the edge line corresponding to the maximum meridian evaluation value located between the two target acupoints.

[0136] Each edge line located between two target acupoints refers to the edge line that is connected to the acupoint and is located between the two acupoints. In the theory of meridians in traditional Chinese medicine, there is usually only one meridian that directly connects two acupoints. Therefore, in this embodiment, the direction of the edge line corresponding to the maximum meridian evaluation value is taken as the connection direction of the two target acupoints.

[0137] By referring to the method for determining the connection direction between two target acupoints, the connection direction between each pair of target acupoints can be obtained.

[0138] S33, based on the connection direction of two target acupoints, obtains the deep meridian distribution on the three-dimensional model of the structural tissue.

[0139] In this embodiment, the concept of region growth is utilized. Under the premise of determining the connection direction, region growth analysis is performed in the three-dimensional model of the structural organization. Specifically, meridian edge lines are selected, and the results of the completed region growth are used as the distribution of deep meridians. The concept of region growth is existing technology and is not within the scope of this invention; therefore, it will not be elaborated upon here.

[0140] Thus, this embodiment has obtained the distribution of deep meridians in the three-dimensional model of the structural organization.

[0141] S4, combining surface meridian distribution and deep meridian distribution to construct a three-dimensional meridian model of the user to be modeled.

[0142] In this embodiment, the surface and deep meridian distributions are fitted and connected using spline interpolation to fill in any breaks, resulting in a complete human meridian distribution. Based on the connection patterns of the human meridian distribution, a three-dimensional meridian model can be constructed using voxel modeling within the built human body model, thus obtaining a three-dimensional meridian model of the user to be modeled. The implementation processes of spline interpolation and voxel modeling are existing technologies and are not within the scope of this invention; therefore, they will not be described in detail here.

[0143] Thus, this embodiment has obtained a three-dimensional meridian model of the user to be modeled with high completeness and accuracy.

[0144] Another embodiment of the present invention provides an artificial intelligence-based three-dimensional reconstruction system for acupuncture meridians, including a processor and a memory, wherein the processor is used to process instructions stored in the memory to implement an artificial intelligence-based three-dimensional reconstruction method for acupuncture meridians.

[0145] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for three-dimensional reconstruction of acupuncture meridians based on artificial intelligence, characterized in that, Includes the following steps: Infrared image data of the user to be modeled is acquired, and each target acupoint in the three-dimensional model of the human body tissue structure of the user to be modeled is acquired, thereby determining each target acupoint in the infrared image data; wherein, the target acupoint is a specific acupoint marked on the skin surface; In the infrared image data, the temperature change curves of each target acupoint stimulation point and each pixel in its neighborhood are obtained, and the surface meridian distribution of each target acupoint is determined by analyzing the temperature rise based on the temperature change curves. In the three-dimensional model of the tissue structure, the surface meridian distribution of each edge line and each target acupoint is determined, and the deep meridian distribution is determined based on the surface meridian distribution of each edge line and each target acupoint. A three-dimensional meridian model of the user to be modeled is constructed by combining the surface meridian distribution and the deep meridian distribution; The process of acquiring the temperature change curves of each stimulation point and its neighboring pixels for each target acupoint includes: For each target acupoint, the temperature data sequence of each pixel point in the stimulation point and its neighborhood range within a preset time period is obtained by stimulating the target acupoint. Based on the temperature data sequence of each pixel in the stimulation point and its neighborhood within a preset time period, construct the temperature change curve of each pixel in the stimulation point and its neighborhood. The step of determining the surface meridian distribution of each target acupoint by analyzing the temperature rise based on the temperature change curve includes: For each target acupoint, the degree of temperature rise at each time point within the preset time period is determined based on the temperature change curve. The temperature rise at each time point is compared with a preset temperature rise threshold to obtain the time points of temperature change. The temperature change time intervals are obtained by considering the temperature change time differences between the stimulation point and each pixel in its neighborhood, and the temperature diffusion direction at the location of the stimulation point of the target acupoint is determined based on the temperature change time intervals. Based on the temperature diffusion direction at the location of the stimulation point, a region growth analysis is performed on the infrared image to obtain the surface meridian distribution of the target acupoint. The determination of deep meridian distribution based on the surface meridian distribution of each edge line and each target acupoint includes: The meridian evaluation value of each edge line is determined based on its length, density, disorder, and distance from the distribution of each surface meridian. Obtain the edge lines between two target acupoints, and determine the connection direction between the two target acupoints by using the edge line corresponding to the maximum meridian evaluation value between the two target acupoints. The distribution of deep meridians is obtained on a three-dimensional model of the structural tissue based on the connection direction of two target acupoints. The process of determining the meridian evaluation value of each edge line based on its length, density, disorder, and distance from the distribution of surface meridians includes: For each edge line, determine the length of the edge line and count the number of edge lines within the neighborhood. Determine the distance between the edge line and each surface meridian distribution, and calculate the slope variance of the edge line; By integrating the length of the edge line, the number of edge lines, the distance, and the slope variance, a meridian evaluation value for the edge line is obtained. The meridian evaluation value is positively correlated with the length and the number of edge lines, and negatively correlated with the distance and the slope variance.

2. The method for three-dimensional reconstruction of acupuncture meridians based on artificial intelligence according to claim 1, characterized in that, The determination of each target acupoint in the infrared image data includes: Image registration is performed based on infrared images and standard human acupoint diagrams, so that each standard acupoint in the standard human acupoint diagram is mapped onto the infrared image, thereby obtaining each candidate acupoint in the infrared image. The candidate acupoints in the infrared image are matched with the three-dimensional model of the tissue structure to determine the target acupoints in the infrared image data.

3. The method for three-dimensional reconstruction of acupuncture meridians based on artificial intelligence according to claim 2, characterized in that, The process of obtaining each candidate acupoint in the infrared image includes: Obtain a standard human acupoint chart, and scale the standard human acupoint chart using a Gaussian pyramid to obtain standard human acupoint charts of different scales. Extract each matching point pair between the infrared image and the standard human acupoint map at different scales, and calculate the affine transformation matrix between the infrared image and the standard human acupoint map based on the matching point pairs; The standard human acupoint diagram is transformed using the affine transformation matrix to obtain a transformed image, and the transformed image is resampled to obtain an infrared image aligned to the coordinate system and a standard human acupoint diagram of a selected scale. By mapping each acupoint in the standard human acupoint chart of a selected scale onto the infrared image, each candidate acupoint in the infrared image is obtained.

4. The method for three-dimensional reconstruction of acupuncture meridians based on artificial intelligence according to claim 1, characterized in that, Determining the degree of temperature rise at each time point within the preset time period based on the temperature change curve includes: For any given location, obtain the temperature difference between the initial temperature and the temperature at the current time point, and obtain the slope of the temperature change curve at the current time point; wherein, the location point is the stimulus point and each pixel point within its neighborhood. By combining the temperature difference value and the slope, the degree of temperature rise at the current time point is determined; wherein, both the temperature difference value and the slope are positively correlated with the degree of temperature rise.

5. The method for three-dimensional reconstruction of acupuncture meridians based on artificial intelligence according to claim 1, characterized in that, After determining the superficial meridian distribution of each target acupoint, the following is included: When there are breaks in the surface meridian distribution, a human meridian acupoint map is obtained. Based on the meridian connection information between acupoints in the human meridian acupoint map, spline difference fitting is performed on the surface meridian distribution with breaks to obtain the fitted surface meridian distribution.

6. A three-dimensional reconstruction system for acupuncture meridians based on artificial intelligence, characterized in that, It includes a processor and a memory, the processor being used to process instructions stored in the memory to implement an artificial intelligence-based three-dimensional reconstruction method for acupuncture meridians as described in any one of claims 1-5.

Images