Adipose tissue space structure quantification method and system based on three-dimensional reconstruction

Abstract

The invention relates to the technical field of structure quantification, and discloses an adipose tissue space structure quantification method and system based on three-dimensional reconstruction, and the method comprises the steps: carrying out the three-dimensional space dynamic alignment of the anatomical boundary of a structure image in an adipose tissue and the metabolic characteristics of a functional image, and obtaining fusion body data; performing curvature continuity verification on the fat tissue curved surface based on the fusion body data to obtain a fat topological curved surface; carrying out space vector fusion on adjacent fragments in the fat topology curved surface to obtain a closed biological topology boundary; performing density gradient direction coupling on space grid nodes in the closed biological topology boundary and the tissue density gradient to obtain a connection vector; performing spatial orientation aggregation on the connection vectors to obtain a spatial heterogeneity distribution field; mapping the spatial heterogeneity distribution field to a standard anatomical coordinate system to obtain a three-dimensional vector map; according to the method, the efficiency of quantifying the adipose tissue space structure based on three-dimensional reconstruction can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of structure quantification, and particularly relates to a fat tissue spatial structure quantification method and system based on three-dimensional reconstruction. BACKGROUND

[0002] In the prior art, the fusion processing of structural images and functional images in the fat tissue spatial structure quantification process is not accurate enough, and an effective three-dimensional spatial dynamic alignment mechanism is lacking, resulting in a low matching degree of anatomical boundaries and metabolic characteristics, and it is difficult to form complete and accurate fusion data, thereby affecting the basic reliability of subsequent quantification analysis.

[0003] Meanwhile, the prior art is rough in processing aspects such as curved surface reconstruction, boundary fusion and spatial mapping, and key factors such as curvature continuity and density gradient coupling are not fully considered, so that the completeness of the generated spatial structure quantification result is insufficient, and the spatial heterogeneity distribution of the fat tissue cannot be accurately reflected, and it is difficult to meet the efficient and accurate quantification requirements. SUMMARY

[0004] The present application provides a fat tissue spatial structure quantification method and system based on three-dimensional reconstruction, which mainly aims to solve the problems raised in the background technology.

[0005] To achieve the above purpose, the fat tissue spatial structure quantification method based on three-dimensional reconstruction provided by the present application comprises:

[0006] S1. The anatomical boundaries of the structural images and the metabolic characteristics of the functional images in the fat tissue are dynamically aligned in three-dimensional space to obtain the fusion data of the fat tissue;

[0007] S2. The curvature continuity of the fat tissue curved surface is verified based on the fusion data to obtain the fat topological curved surface of the fat tissue;

[0008] S3. The adjacent segments in the fat topological curved surface are fused by spatial vectors to obtain the closed biological topological boundary of the fat tissue;

[0009] S4. The density gradient direction coupling of the spatial grid nodes and the tissue density gradient in the closed biological topological boundary is performed to obtain the connection vector of the closed biological topological boundary;

[0010] S5. The spatial orientation aggregation of the connection vector is performed to obtain the spatial heterogeneity distribution field of the fat tissue;

[0011] S6. The spatial heterogeneity distribution field is mapped to a standard anatomical coordinate system to obtain a three-dimensional vector atlas of the spatial structure quantification.

[0012] In a preferred embodiment, the anatomical boundary of the structural image of the adipose tissue is dynamically aligned with the metabolic feature of the functional image in three-dimensional space to obtain the fusion data of the adipose tissue, comprising:

[0013] The structural image of the adipose tissue is subjected to spatial noise suppression to obtain a clear structural image of the adipose tissue;

[0014] The metabolic activity region of the clear structural image is profile labeled to obtain the metabolic feature profile of the functional image;

[0015] The metabolic feature profile and the profile boundary of the clear structural image are matched and optimally registered to obtain the fusion data of the adipose tissue.

[0016] In a preferred embodiment, the curvature continuity of the adipose tissue surface is verified based on the fusion data to obtain the adipose topological surface of the adipose tissue, comprising:

[0017] The fusion data is subjected to surface topological reconstruction to obtain the initial surface of the adipose tissue;

[0018] The initial surface is subjected to global curvature feature analysis to obtain the curvature distribution profile of the initial surface;

[0019] The continuity of the adipose tissue surface is modified based on the curvature distribution profile to obtain the adipose topological surface of the adipose tissue.

[0020] In a preferred embodiment, the spatial vector fusion of adjacent segments in the adipose topological surface is performed to obtain the closed biological topological boundary of the adipose tissue, comprising:

[0021] The adjacent segment set of the adipose topological surface is obtained by segment spatial adjacency analysis of the adipose topological surface;

[0022] The calibration spatial vector group of the adipose topological surface is obtained by spatial vector registration calibration of the adjacent segment set;

[0023] The closed biological topological boundary of the adipose tissue is obtained by spatial direction consistency fusion of the calibration spatial vector group.

[0024] In a preferred embodiment, the calibration spatial vector group of the adipose topological surface is obtained by spatial vector registration calibration of the adjacent segment set, comprising:

[0025] The curvature gradient tensor of the adjacent segment set is obtained by differential manifold analysis of the boundary feature point set of each segment in the adjacent segment set;

[0026] Optimize the geometric constraint registration of the spatial position difference of the boundary feature point set based on the curvature continuity index, to obtain an initial calibration vector of the fat topological surface;

[0027] Align the initial calibration vector with the reference coordinate system of the fat topological surface in space to obtain an intermediate calibration vector group of the fat topological surface;

[0028] Integrate the intermediate calibration vector group in space to obtain a calibration space vector group of the fat topological surface.

[0029] In a preferred embodiment, the density gradient direction coupling of the spatial grid node in the closed biological topological boundary with the tissue density gradient is to obtain the connection vector of the closed biological topological boundary, comprising:

[0030] Map the spatial grid node to the spatial position field of the tissue density gradient to obtain the density gradient representation of the spatial grid node;

[0031] Based on the density gradient representation, the spatial coordinates of the spatial grid node are directionally corrected to obtain the corrected spatial node of the spatial grid node;

[0032] Connect adjacent corrected spatial nodes to obtain the connection vector of the closed biological topological boundary.

[0033] In a preferred embodiment, the spatial orientation aggregation of the connection vector is to obtain the spatial heterogeneity distribution field of the adipose tissue, comprising:

[0034] Divide the connection vector into a preset orientation interval to obtain a vector grouping set of the adipose tissue;

[0035] Differential statistics of the vector density of each group in the vector grouping set is to obtain a spatial density distribution matrix of the adipose tissue;

[0036] Map the gradient change characteristics of the spatial density distribution matrix to the spatial heterogeneity distribution field of the adipose tissue.

[0037] In a preferred embodiment, the differential statistics of the vector density of each group in the vector grouping set is to obtain a spatial density distribution matrix of the adipose tissue, comprising:

[0038] Map the vector grouping set to the spatial coordinate-density value correlation sequence of the adipose tissue;

[0039] Perform spatial interpolation processing on the spatial coordinate-density value correlation sequence to obtain a continuous density distribution surface of the adipose tissue;

[0040] The spatial location point density value is calculated based on the continuous density distribution surface, wherein the formula for calculating the spatial location point density value is as follows:

[0041]

[0042] In the formula, Here, h represents the spatial density value of the spatial location points, n is the number of vectors in the current group, and h is the number of vectors in the current group. 2 Let K be the two-dimensional Gaussian kernel function, x be the abscissa of the spatial coordinates, y be the ordinate of the spatial coordinates, h be the adaptive bandwidth parameter, d be the density value, and i be the vector index.

[0043] The spatial density distribution matrix of the adipose tissue is obtained by performing grid topology reconstruction on the density values ​​of the spatial location points.

[0044] In a preferred embodiment, mapping the spatial heterogeneity distribution field to a standard anatomical coordinate system to obtain a three-dimensional vector map of the quantified spatial structure includes:

[0045] Spatial correspondence analysis was performed on the spatial heterogeneous distribution field and the reference marker set to obtain the spatial location difference data of the adipose tissue;

[0046] A non-uniform deformation field of the adipose tissue is constructed based on the spatial location difference data;

[0047] Spatial proximity weight correction is applied to the geometrically distorted region in the non-uniform deformation field to obtain the calibrated deformation field in the non-uniform deformation field.

[0048] The spatial heterogeneity distribution field is transformed by the calibration deformation field to obtain a three-dimensional vector map of the quantized spatial structure.

[0049] To address the aforementioned problems, this invention also provides a system for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction, the system comprising:

[0050] A three-dimensional dynamic alignment module is used to dynamically align the anatomical boundaries of the structural images in adipose tissue with the metabolic features of the functional images in three-dimensional space to obtain the fused data of the adipose tissue.

[0051] The curvature verification module is used to verify the curvature continuity of the adipose tissue surface based on the fused body data, and to obtain the adipose tissue topological surface.

[0052] The vector fusion module is used to perform spatial vector fusion on adjacent segments in the fat topological surface to obtain the closed biological topological boundary of the fat tissue.

[0053] The gradient coupling module is used to couple the spatial grid nodes and tissue density gradient in the closed biological topological boundary with the density gradient direction to obtain the connection vector of the closed biological topological boundary.

[0054] The orientation aggregation module is used to perform spatial orientation aggregation on the connection vectors to obtain the spatial heterogeneous distribution field of the adipose tissue;

[0055] The coordinate mapping module is used to map the spatial heterogeneity distribution field to a standard anatomical coordinate system to obtain a three-dimensional vector map of the quantified spatial structure.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] 1. This invention achieves precise capture of the spatial structural features of adipose tissue by dynamically aligning the anatomical boundaries of structural images with the metabolic features of functional images in three-dimensional space, constructing adipose tissue topological surfaces by combining curvature continuity verification, and then forming closed biological topological boundaries through spatial vector fusion. This improves the accuracy and completeness of the quantitative results.

[0058] 2. Simultaneously, the connection vector is obtained through density gradient directional coupling, and a spatial heterogeneous distribution field is generated by spatial orientation aggregation. This field is then mapped to a standard anatomical coordinate system to obtain a three-dimensional vector map, which realizes the refined quantification of the spatial structure of adipose tissue, effectively improves the quantification efficiency, and provides reliable spatial structure data support for adipose tissue-related research. Attached Figure Description

[0059] Figure 1 This is a flowchart illustrating a method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction, provided in an embodiment of the present invention.

[0060] Figure 2 A functional block diagram of a three-dimensional reconstruction-based adipose tissue spatial structure quantification system provided in an embodiment of the present invention;

[0061] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0062] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0063] This application provides a method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction. The execution entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0064] Reference Figure 1 The diagram shown is a flowchart illustrating a method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction, according to an embodiment of the present invention. In this embodiment, the method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction includes:

[0065] S1. The anatomical boundaries of the structural images in adipose tissue are dynamically aligned with the metabolic features of the functional images in three-dimensional space to obtain the fusion data of the adipose tissue;

[0066] In this embodiment of the invention, the step of dynamically aligning the anatomical boundaries of the structural images and the metabolic features of the functional images in three-dimensional space to obtain the fused data of the adipose tissue includes:

[0067] Spatial noise suppression is performed on the structural image to obtain a clear structural image of the adipose tissue;

[0068] The metabolic activity region contours of the clear structural image are annotated to obtain the metabolic feature contours of the functional image.

[0069] The metabolic feature contours are matched and registered with the contour boundaries of the clear structural image to obtain the fusion data of the adipose tissue.

[0070] Specifically, the adipose tissue was from a goose, not a mouse. Spatial noise suppression was performed on the structural images to obtain clear structural images of the adipose tissue. First, the structural images of the adipose tissue were scanned to identify noise regions in the images caused by equipment imaging errors and tissue motion interference.

[0071] Furthermore, these areas appear as spots or stripes with irregular abrupt changes in grayscale values. Spatial neighborhood smoothing is used, with each pixel as the center, selecting a certain range of neighboring pixels, calculating the average grayscale value of these pixels, and replacing the original grayscale value of the center pixel with this average value.

[0072] Furthermore, for pixels at the anatomical boundary, only the average value of adjacent pixels on the same boundary side is calculated to avoid blurring the boundary. After processing, the gray values ​​of the noise area tend to be flat, and the gray contrast of the anatomical boundary remains clear, resulting in a clear structural image of the adipose tissue.

[0073] Furthermore, the metabolically active regions of the clear structural images are delineated to obtain the metabolic feature contours of the functional images. First, the functional images of adipose tissue are obtained, in which the metabolically active regions are represented by regions with specific signal intensities.

[0074] Furthermore, a preliminary spatial correspondence was established between functional images and clear structural images to determine the corresponding anatomical locations of highly metabolically active regions in functional images within clear structural images.

[0075] Furthermore, the contouring tool is used to mark along the anatomical boundary edge of the anatomical location, following the natural anatomical division of adipose tissue in the clear structural image, ensuring that the contour line completely surrounds the metabolically active area. After marking, the contour line is extracted as the metabolic feature contour of the functional image.

[0076] Furthermore, the metabolic feature contour and the contour boundary of the clear structural image are matched and optimized to obtain the fusion data of adipose tissue. First, the key control points of the contour boundaries of the metabolic feature contour and the clear structural image are extracted in three-dimensional space. These control points include the inflection points, endpoints and points with significant curvature changes of the contour.

[0077] Furthermore, the spatial distance between the corresponding control points is calculated, and the spatial position of the metabolic feature contour is adjusted according to the distance to reduce the distance between each corresponding control point. The adjustment process is repeated until the distance between all corresponding control points is less than the set minimum range. At this point, the contour boundaries of the two are highly coincident in three-dimensional space.

[0078] Furthermore, the anatomical information from clear structural images and the metabolic feature information from functional images are superimposed and integrated along overlapping contour boundaries to form a dataset that simultaneously contains three-dimensional anatomical boundaries and metabolic feature distributions, thus obtaining fusion data of adipose tissue.

[0079] In summary, spatial noise suppression of structural images yields clear structural images of adipose tissue, removing noise and preserving clear anatomical boundaries, thus providing an accurate basis for subsequent annotation.

[0080] In summary, by annotating the contours of metabolically active regions on clear structural images, the metabolic feature contours of functional images can be obtained, which can link metabolic regions with anatomical locations and clarify their spatial extent.

[0081] In summary, by optimizing the matching degree of metabolic feature contours with the contour boundaries of clear structural images, fusion data of adipose tissue can be obtained. This can integrate structural and metabolic information, providing a precise basis for spatial structure quantification and improving efficiency.

[0082] S2. Based on the fused body data, the curvature continuity of the adipose tissue surface is verified to obtain the adipose tissue topological surface;

[0083] In this embodiment of the invention, the step of verifying the curvature continuity of the adipose tissue surface based on the fused body data to obtain the adipose tissue topological surface includes:

[0084] The fused body data is subjected to surface topology reconstruction to obtain the initial surface of the adipose tissue;

[0085] A global curvature feature analysis is performed on the initial surface to obtain the curvature distribution profile of the initial surface;

[0086] Based on the curvature distribution contour, the adipose tissue surface is continuously spatially modified to obtain the adipose tissue topological surface.

[0087] Specifically, surface topology reconstruction is performed on the fused body data to obtain the initial surface of the adipose tissue. Three-dimensional spatial coordinate points of the adipose tissue are extracted from the fused body data. These points cover the surface locations of the anatomical boundaries and metabolic feature regions of the adipose tissue.

[0088] Furthermore, these points are connected into a triangular mesh according to their spatial proximity. The three vertices of each triangle are adjacent spatial coordinate points, ensuring that the mesh completely covers the surface range of the adipose tissue and that the mesh edges fit the anatomical boundary contours in the fusion data, forming a three-dimensional surface that can initially reflect the surface morphology of the adipose tissue, thus obtaining the initial surface of the adipose tissue.

[0089] Furthermore, a global curvature feature analysis is performed on the initial surface to obtain the curvature distribution profile of the initial surface, and each triangular mesh vertex of the initial surface is traversed.

[0090] Furthermore, the curvature value at each vertex is calculated. The magnitude of the curvature is determined by comparing the differences in the normal vector directions of the adjacent triangles around that vertex. The greater the change in the normal vector direction, the greater the curvature value, and vice versa.

[0091] Furthermore, the curvature values ​​of all vertices are recorded and marked according to their spatial positions on the surface. Then, regions with the same or similar curvature values ​​are divided into the same curvature interval, and different intervals are distinguished by different labels, forming a distribution graph that can intuitively show the degree of curvature of each region of the initial surface, thus obtaining the curvature distribution contour of the initial surface.

[0092] Furthermore, based on the curvature distribution contour, the adipose tissue surface is continuously spatially modified to obtain the adipose tissue topological surface. Based on the curvature distribution contour, regions with abrupt curvature changes in the initial surface are identified. These regions are characterized by curvature value differences between adjacent vertices that exceed the set reasonable range.

[0093] Furthermore, the positions of the grid vertices in this region are adjusted, moving the vertices to positions where the curvature values ​​of adjacent vertices transition smoothly.

[0094] Furthermore, during adjustment, the connection between the vertex and the surrounding triangular mesh remains unchanged, only the spatial coordinates are changed, and the adjustment is repeated until the curvature values ​​of all adjacent vertices are within a reasonable range, so that the curvature of the entire surface changes continuously. The resulting surface is the fat topology surface of the adipose tissue.

[0095] In summary, surface topology reconstruction of the fused body data yields an initial surface, which can transform three-dimensional coordinate points into a mesh surface, providing a basic basis for curvature analysis.

[0096] In summary, global curvature analysis of the initial surface yields the curvature distribution profile, which can identify bending changes and pinpoint areas of curvature abrupt changes that require adjustment.

[0097] In summary, fat topological surfaces obtained by curvature distribution contour trimming can eliminate discontinuities, smooth curvature changes, provide a consistent basis for subsequent fusion, and improve quantification accuracy.

[0098] S3. Spatial vector fusion is performed on adjacent segments in the adipose tissue topology surface to obtain the closed biological topological boundary of the adipose tissue;

[0099] In this embodiment of the invention, the step of fusing adjacent segments in the adipose tissue topology surface to obtain the closed biological topological boundary of the adipose tissue includes:

[0100] The fat topological surface is subjected to fragment spatial adjacency analysis to obtain the set of adjacent fragments of the fat topological surface;

[0101] Spatial vector registration and calibration are performed on the adjacent segment set to obtain the calibration spatial vector group of the fat topological surface;

[0102] Spatial orientation consistency fusion is performed on the calibration spatial vector group to obtain the closed biological topological boundary of the adipose tissue.

[0103] The step of performing spatial vector registration and calibration on the adjacent segment set to obtain the calibration spatial vector group of the fat topological surface includes:

[0104] Differential manifold analysis is performed on the boundary feature point set of each segment in the adjacent segment set to obtain the curvature gradient tensor of the adjacent segment set;

[0105] Based on the curvature continuity index, the spatial position difference of the boundary feature point set is geometrically constrained and registered to obtain the initial calibration vector of the fat topological surface.

[0106] The initial calibration vector is spatially aligned with the reference coordinate system of the fat topology surface to obtain the intermediate calibration vector group of the fat topology surface.

[0107] Spatial integration of the intermediate calibration vector group yields the calibration spatial vector group of the fat topological surface.

[0108] Specifically, the fat topological surface is parsed using segment spatial adjacency analysis to obtain a set of adjacent segments of the fat topological surface. The fat topological surface is first divided into multiple independent surface segments according to a preset grid cell size. Each segment contains a continuous triangular grid, and the segments are separated by boundary lines. The boundary lines of each surface segment are traversed.

[0109] Furthermore, check whether the boundary line intersects or overlaps with the boundary lines of other surface segments. If the boundary lines of two segments intersect or overlap for a length exceeding a preset minimum contact length, then the two segments are determined to be adjacent segments.

[0110] Furthermore, all adjacent segments are recorded in pairs to form a set containing multiple pairs of adjacent segments, thus obtaining the set of adjacent segments of the fat topological surface.

[0111] Furthermore, spatial vector registration and calibration are performed on the adjacent segment sets to obtain the calibration spatial vector group of the fat topological surface for each pair of adjacent segments in the adjacent segment set.

[0112] Furthermore, feature points on the boundary of each segment are extracted. These feature points are the endpoints and midpoints of the boundary line. The position vector of each feature point in three-dimensional space is calculated. The starting point of the vector is the reference center point inside the segment, and the ending point is the feature point.

[0113] Furthermore, using the feature point position vector of one segment as a reference, the feature point position vector of another segment is adjusted by translating or rotating the other segment so that the angle between the position vectors of the corresponding feature points in space is less than a preset angle threshold. After all the position vectors of the corresponding feature points are adjusted,

[0114] Furthermore, these adjusted position vectors are collected to form a calibration space vector group for the fat topological surface.

[0115] Furthermore, spatial orientation consistency fusion is performed on the calibration spatial vector group to obtain the closed biological topological boundary of adipose tissue, and the orientation of all vectors in the calibration spatial vector group is checked.

[0116] Furthermore, the direction difference between adjacent vectors is calculated. If the direction difference exceeds a preset consistency threshold, the direction of one of the vectors is fine-tuned so that its direction difference with that of the adjacent vectors is within the threshold range.

[0117] In detail, repeat this process until the direction difference of all adjacent vectors meets the requirements, ensuring that the direction of the entire vector group shows a continuous changing trend. Then, connect these vectors with the same direction in sequence according to their positions in space, with the starting and ending points of the vectors connecting to each other to form a closed contour without gaps. This contour is the closed biological topological boundary of adipose tissue.

[0118] Specifically, differential manifold analysis is performed on the boundary feature point set of each segment in the adjacent segment set to obtain the curvature gradient tensor of the adjacent segment set.

[0119] Furthermore, the boundary feature point set of each segment is extracted from the set of adjacent segments. These point sets include the endpoints, midpoints and points with significant curvature changes on the segment boundary. The curvature changes of the neighboring points around each feature point are observed to determine the trend and speed of curvature changes in different directions in space.

[0120] Furthermore, for each feature point, the curvature variation information in three spatial directions is recorded. This information is then organized into a structured set according to the spatial location of the feature point. This set can comprehensively reflect the curvature gradient variation at the boundary of each segment, which is the curvature gradient tensor of the adjacent segment set.

[0121] Furthermore, based on the curvature continuity index, geometric constraint registration optimization is performed on the spatial position difference of the boundary feature point set to obtain the initial calibration vector of the fat topological surface. The curvature continuity index is a standard for measuring whether the curvature change of adjacent feature points is continuous.

[0122] Furthermore, the positional difference of the corresponding boundary feature points of adjacent segments in three-dimensional space is calculated, that is, the numerical difference between the two points in coordinates. These positional differences are constrained according to the curvature continuity index, and the position of the feature point of one of the segments is adjusted to reduce the positional difference.

[0123] Furthermore, to ensure that the curvature changes of adjacent feature points after adjustment meet the continuity requirement and do not have abrupt changes, the spatial position of each feature point after adjustment is represented by a vector, with the starting point of the vector being the reference center of the segment and the ending point being the adjusted feature point position, thus obtaining the initial calibration vector of the fat topological surface.

[0124] Furthermore, the initial calibration vector is spatially aligned with the reference coordinate system of the fat topology surface to obtain the intermediate calibration vector group of the fat topology surface. The reference coordinate system of the fat topology surface is a preset three-dimensional reference coordinate system, the origin of which is set at the geometric center of the adipose tissue, and the coordinate axis direction is consistent with the human anatomical direction.

[0125] Furthermore, the coordinate values ​​of each initial calibration vector are transformed to the reference coordinate system. Through translation and rotation operations, the starting point and direction of the initial calibration vector are made to correspond with the coordinate axes of the reference coordinate system, ensuring that all initial calibration vectors are in the same coordinate system. After the transformation is completed, all vectors that have undergone coordinate system transformation are collected to form the intermediate calibration vector group of the fat topology surface.

[0126] Furthermore, the intermediate calibration vector group is spatially integrated to obtain the calibration space vector group of the fat topology surface, and the spatial relationship between the vectors in the intermediate calibration vector group is examined.

[0127] Furthermore, the angle and distance between adjacent vectors are calculated. If there is a situation where the vector direction deviation is too large or the distance exceeds a reasonable range, these vectors are fine-tuned.

[0128] Furthermore, the adjustment is based on the principle of maintaining the coherence of the overall vector group, so that adjacent vectors can transition smoothly. The process of checking and adjusting is repeated until all vectors form a coordinated distribution in space. The integrated vector group can accurately reflect the spatial positional relationship of adjacent segments, which is the calibration spatial vector group of the fat topological surface.

[0129] In summary, by performing segment spatial adjacency analysis on the fat topological surface, the set of adjacent segments of the fat topological surface can be obtained. This can accurately identify segments with spatial adjacency relationships in the surface, clarify the connection objects of each segment, define the scope for subsequent vector fusion, avoid interference from irrelevant segments, and ensure the accuracy of the fusion object.

[0130] In summary, spatial vector registration and calibration of adjacent fragment sets yields a calibration spatial vector group of fat topological surfaces. By adjusting the spatial vectors of adjacent fragments, positional deviations and directional differences between fragments can be eliminated, ensuring that the vector group maintains a consistent coordinate reference in space. This provides a precise vector basis for directional consistency fusion.

[0131] In summary, spatial orientation consistency fusion of the calibration spatial vector group yields the closed biological topological boundary of adipose tissue. This allows the vector group to transition continuously in direction, forming a closed contour without gaps, fully presenting the biological topological boundary of adipose tissue. This provides a continuous and complete boundary reference for subsequent density gradient coupling and spatial structure quantization, improving the completeness and accuracy of the quantization results.

[0132] In summary, differential manifold analysis of the feature point set of adjacent segments yields the curvature gradient tensor, which can capture changes in curvature space, providing a precise basis for registration and ensuring a close fit to the natural shape of the surface.

[0133] In summary, optimizing the spatial position difference of boundary points based on the curvature continuity index yields an initial calibration vector that balances position matching and curvature continuity, laying the foundation for vector calibration.

[0134] In summary, aligning the initial calibration vector with the reference coordinate system to obtain the intermediate calibration vector group can unify the coordinate reference, eliminate directional deviations, and provide a consistent basis for spatial integration.

[0135] In summary, spatial integration of intermediate calibration vector groups yields a calibration space vector group, which coordinates vector relationships, enables smooth transitions between adjacent vectors, provides a high-quality vector group for subsequent fusion, and improves boundary accuracy.

[0136] S4. Couple the spatial grid nodes and tissue density gradient in the closed biological topological boundary with the density gradient direction to obtain the connection vector of the closed biological topological boundary;

[0137] In this embodiment of the invention, the step of coupling the spatial grid nodes and the tissue density gradient in the closed biological topological boundary to obtain the connection vector of the closed biological topological boundary includes:

[0138] The spatial grid nodes are mapped to the spatial location field of the tissue density gradient to obtain the density gradient representation of the spatial grid nodes;

[0139] Based on the density gradient characterization, the spatial coordinates of the spatial grid nodes are oriented to obtain the corrected spatial nodes of the spatial grid nodes;

[0140] Connecting adjacent correction space nodes yields the connection vector of the closed biological topological boundary.

[0141] Specifically, spatial grid nodes are mapped to the spatial location field of tissue density gradient to obtain the density gradient representation of spatial grid nodes. Spatial grid nodes are discrete grid points on closed biological topological boundaries, and each node has a clear three-dimensional spatial coordinate.

[0142] Furthermore, the spatial location field of the tissue density gradient is a pre-constructed field containing information on density changes at various spatial locations of adipose tissue. Each location in this field records the direction in which the density changes most rapidly in three-dimensional space and the degree of change in that direction.

[0143] Furthermore, during mapping, the position of the three-dimensional coordinates of each spatial grid node in the spatial location field of the tissue density gradient is found one by one.

[0144] Furthermore, the direction and degree of density change recorded at this location are extracted, and this information is organized in the order of spatial grid nodes. The resulting set is the density gradient representation of the spatial grid nodes.

[0145] Furthermore, the spatial coordinates of the spatial grid nodes are oriented based on the density gradient representation to obtain the corrected spatial nodes of the spatial grid nodes. For each spatial grid node, the direction of density change in its density gradient representation is used.

[0146] Furthermore, it is determined whether the current spatial coordinates of the node are distributed along this direction. If the node coordinates deviate from the density change direction, the three-dimensional coordinates of the node are finely adjusted along the density change direction.

[0147] Furthermore, during the adjustment, the relative positions of the nodes on the closed biological topological boundary remain unchanged, and only the coordinate values ​​are changed to better fit the direction of density change. The nodes obtained after the adjustment are the corrected spatial nodes.

[0148] Furthermore, by connecting adjacent correction space nodes, the connection vector of the closed biological topological boundary is obtained, and the adjacent node pairs in the correction space nodes are determined. The adjacency relationship is determined based on the connection relationship of the nodes in the original mesh of the closed biological topological boundary. That is, nodes connected by edges in the original mesh are still regarded as adjacent nodes after correction.

[0149] Furthermore, each pair of adjacent correction space nodes is connected by directed line segments, with the starting point of one correction space node and the ending point of another adjacent correction space node. All such directed line segments together constitute the connection vector of the closed biological topological boundary.

[0150] In summary, mapping spatial grid nodes to the tissue density gradient field to obtain a density gradient characterization can correlate node and density change information, providing a basis for correction.

[0151] In summary, using density gradient to characterize and correct spatial grid node coordinates to obtain corrected spatial nodes enables nodes to conform to the direction of density change, thereby improving positional accuracy.

[0152] In summary, connecting adjacent correction spatial nodes to obtain connection vectors can accurately reflect the spatial connection relationships of the boundary, providing a reliable vector basis for subsequent orientation aggregation.

[0153] S5. Perform spatial orientation aggregation on the connection vectors to obtain the spatial heterogeneous distribution field of the adipose tissue;

[0154] In this embodiment of the invention, the step of spatial orientation aggregation of the connection vectors to obtain the spatial heterogeneity distribution field of the adipose tissue includes:

[0155] The connection vectors are divided into preset directional intervals to obtain a vector group set of the adipose tissue;

[0156] The spatial density distribution matrix of the adipose tissue is obtained by performing differential statistics on the vector density of each group in the vector grouping set.

[0157] The gradient variation characteristics of the spatial density distribution matrix are mapped to the spatial heterogeneity distribution field of the adipose tissue.

[0158] The step of performing differential statistics on the vector densities of each group in the vector grouping set to obtain the spatial density distribution matrix of the adipose tissue includes:

[0159] Map the vector group set to the spatial coordinate-density value association sequence of the adipose tissue;

[0160] Spatial interpolation is performed on the spatial coordinate-density value correlation sequence to obtain the continuous density distribution surface of the adipose tissue;

[0161] The spatial location point density value is calculated based on the continuous density distribution surface, wherein the formula for calculating the spatial location point density value is as follows:

[0162]

[0163] In the formula, Here, h represents the spatial density value of the spatial location points, n is the number of vectors in the current group, and h is the number of vectors in the current group. 2 Let K be the two-dimensional Gaussian kernel function, x be the abscissa of the spatial coordinates, y be the ordinate of the spatial coordinates, j be the adaptive bandwidth parameter, d be the density value, and i be the vector index.

[0164] The spatial density distribution matrix of the adipose tissue is obtained by performing grid topology reconstruction on the density values ​​of the spatial location points.

[0165] Specifically, the connecting vectors are divided into preset orientation intervals to obtain a set of vector groups for adipose tissue. The preset orientation intervals are multiple continuous intervals in three-dimensional space divided according to the direction angle, such as the origin of the spatial coordinate system as the vertex.

[0166] Furthermore, the horizontal direction angle from 0 to 360 degrees is divided into intervals of 60 degrees, and the vertical direction elevation angle from -90 to 90 degrees is divided into intervals of 30 degrees. Each interval corresponds to a specific spatial orientation range, and each connecting vector is traversed.

[0167] Furthermore, by calculating the matching degree between its direction and each preset interval, the azimuth interval to which the vector belongs is determined, and all connecting vectors belonging to the same interval are grouped together to form a set containing multiple vector groups, which is the vector grouping set of adipose tissue.

[0168] Furthermore, differential statistics are performed on the vector densities of each group in the vector grouping set to obtain the spatial density distribution matrix of adipose tissue. The vector density is the number of connection vectors contained in each group per unit volume of space.

[0169] Furthermore, first determine the spatial range it covers, which is the smallest cubic region formed by the coordinates of the start and end points of all vectors within the group. Calculate the volume of this cubic region, then count the total number of vectors within the group, and divide the total number of vectors by the volume to obtain the vector density of the group.

[0170] Furthermore, according to the arrangement order of the directional intervals of each group in three-dimensional space, the vector densities of all groups are sequentially filled into the corresponding positions of the matrix, and the resulting matrix is ​​the spatial density distribution matrix of adipose tissue.

[0171] Furthermore, the gradient change characteristics of the spatial density distribution matrix are mapped to the spatial heterogeneity distribution field of adipose tissue, and the gradient change characteristics of the spatial density distribution matrix are the density differences between adjacent elements in the matrix.

[0172] Furthermore, the density difference between each element in the matrix and its adjacent elements (up, down, left, right, front, and back) is calculated. These differences reflect the rate of density change in different spatial directions. Using the three-dimensional spatial coordinates of adipose tissue as a reference, the density value and gradient change characteristics corresponding to each matrix element are assigned to the spatial position corresponding to that element, so that each position in space contains density value and density change information.

[0173] Furthermore, by integrating this information from all spatial locations, a field is formed that can intuitively display the differences and trends in density distribution in different regions of adipose tissue, which is the spatial heterogeneity distribution field of adipose tissue.

[0174] Specifically, the vector group set is mapped to the spatial coordinate-density value association sequence of adipose tissue. Each group in the vector group set corresponds to a specific spatial orientation interval. The starting coordinates of all connecting vectors in each group are extracted, and the average value of these starting coordinates is calculated as the central spatial coordinate of the group. At the same time, the vector density value of the group is recorded.

[0175] Furthermore, the central spatial coordinates of each group are mapped one-to-one with the corresponding vector density value. These correspondences are arranged in the order of the spatial coordinates, and the resulting ordered sequence is the spatial coordinate-density value association sequence of adipose tissue.

[0176] Furthermore, spatial interpolation is performed on the spatial coordinate-density value correlation sequence to obtain a continuous density distribution surface of adipose tissue, where the coordinates in the spatial coordinate-density value correlation sequence are discrete points.

[0177] Furthermore, multiple interpolation points are selected between any two adjacent discrete points. The coordinates of each interpolation point are located on the line connecting the two discrete points. The weight is determined based on the distance between the interpolation point and the two discrete points. The closer the distance, the greater the weight. The density value of the interpolation point is calculated using the density values ​​of the two discrete points according to the weight.

[0178] Furthermore, this process is repeated until interpolation points are filled between all adjacent discrete points. The coordinates of all discrete points and interpolation points are integrated with the density values ​​to form a continuous surface covering the entire spatial range of adipose tissue, which is the continuous density distribution surface of adipose tissue.

[0179] Furthermore, the density values ​​of spatial location points are calculated based on the continuous density distribution surface. Spatial location points are selected at fixed intervals in the three-dimensional space of adipose tissue, and the coordinates of each location point cover the entire range of the continuous density distribution surface. For each spatial location point...

[0180] Furthermore, locate its corresponding position on the continuous density distribution surface, and directly read the density value at that position as the spatial location point density value, ensuring that each selected spatial location point has a corresponding density value record.

[0181] Furthermore, the density values ​​of spatial location points are reorganized into a grid topology to obtain the spatial density distribution matrix of adipose tissue. The three-dimensional space of adipose tissue is divided into cubic grids of equal size according to the spatial coordinate direction, and each grid is a spatial unit corresponding to a matrix element.

[0182] Furthermore, the spatial density values ​​of all spatial locations within each grid are statistically analyzed, and the average of these density values ​​is calculated as the matrix element value corresponding to that grid.

[0183] Furthermore, according to the arrangement order of the grid in the spatial coordinate direction, all matrix element values ​​are filled into the corresponding positions of the matrix, and the resulting matrix is ​​the spatial density distribution matrix of adipose tissue.

[0184] Specifically, n represents the number of vectors within the current group, derived from the total number of connected vectors in the currently processed group within the vector group set, which is obtained by dividing the connected vectors according to a preset directional interval. h is an adaptive bandwidth parameter, automatically determined based on the spatial distribution characteristics of the vectors within the current group, such as calculated based on the dispersion of vector coordinates, used to adjust the influence range of spatial distance. K is a two-dimensional Gaussian kernel function, a preset function used to measure the distance weight between a spatial location point and the vector coordinates; the closer the distance, the greater the weight.

[0185] Furthermore, x and y are the horizontal and vertical coordinates of the spatial coordinates, derived from the coordinates of the spatial location point from which the density value needs to be calculated. (x i ,y i ) represents the coordinates of the i-th vector within the current group, derived from the two-dimensional coordinates of the start or end point of the i-th connecting vector within the current group in the vector grouping set. d i The density value of the i-th vector is derived from the density statistics of the i-th connecting vector in the vector group set, obtained from the previous differential statistics.

[0186] Furthermore, the meaning of this formula is to calculate the density value of a spatial location point (x,y), which is achieved by integrating the density information of all vectors in the current group. Specifically, it is as follows: first, the spatial distance between each vector and the location point is calculated, and a two-dimensional Gaussian kernel function is used to assign a corresponding weight to the density value of the vector according to the distance. The closer the distance, the greater the weight. Then, all the weighted density values ​​are summed, and finally, the average is obtained by dividing by the product of the number of vectors in the current group and the square of the bandwidth, so as to obtain a density value that reflects the comprehensive influence of the density of the surrounding vectors on the location point.

[0187] Furthermore, when the spatial location point (x, y) and the coordinates of a certain vector (x, y) i ,y i The closer the distance, the larger the value of the two-dimensional Gaussian kernel function K, and the larger the density value d of the vector. i The greater the influence on the density value of spatial location points, the more significant the contribution of the nearest distance vector to the density of location points.

[0188] Furthermore, as the bandwidth h increases, the effective range of the two-dimensional Gaussian kernel function expands, and more vectors will affect the density of the location points, resulting in smoother calculation results and reduced influence from a few local vectors. When h decreases, only vectors very close to the location points will have a significant impact, and the results will better reflect subtle changes in local density.

[0189] Furthermore, the more vectors n are in the current group, the larger the denominator, and the more the calculation result will be affected by the averaging effect of more vectors. The overall result is more stable and less affected by the abnormal density of individual vectors.

[0190] In summary, dividing the connecting vectors into preset directional intervals yields a vector group set, which can classify vectors by direction, providing ordered grouping for density statistics.

[0191] In summary, density difference statistics are performed on the vector group set to obtain the spatial density distribution matrix, which can quantify the differences in vector distribution in different directions and form structured data.

[0192] In summary, mapping the gradient features of the spatial density distribution matrix to the spatial heterogeneity distribution field can intuitively present the differences in the spatial structure of adipose tissue, providing a basis for subsequent coordinate mapping.

[0193] In summary, mapping vector group sets to spatial coordinate-density value association sequences can associate vector groups with spatial locations, laying the foundation for data.

[0194] In summary, spatial interpolation of spatial coordinate-density value correlation sequences yields continuous density distribution surfaces, which can complete discrete data and form a complete density shape.

[0195] In summary, calculating the density values ​​of spatial locations based on a continuous density distribution surface can accurately obtain the density at each location, reflecting local characteristics.

[0196] In summary, reorganizing the density values ​​of spatial locations into a grid topology to obtain a spatial density distribution matrix can structure the density information, making it easier for subsequent processing.

[0197] S6. Map the spatial heterogeneity distribution field to a standard anatomical coordinate system to obtain a three-dimensional vector map of the quantified spatial structure.

[0198] In this embodiment of the invention, mapping the spatial heterogeneity distribution field to a standard anatomical coordinate system to obtain the three-dimensional vector map of the quantified spatial structure includes:

[0199] Spatial correspondence analysis was performed on the spatial heterogeneous distribution field and the reference marker set to obtain the spatial location difference data of the adipose tissue;

[0200] A non-uniform deformation field of the adipose tissue is constructed based on the spatial location difference data;

[0201] Spatial proximity weight correction is applied to the geometrically distorted region in the non-uniform deformation field to obtain the calibrated deformation field in the non-uniform deformation field.

[0202] The spatial heterogeneity distribution field is transformed by the calibration deformation field to obtain a three-dimensional vector map of the quantized spatial structure.

[0203] Specifically, spatial correspondence analysis is performed on the spatial heterogeneous distribution field and the reference marker set to obtain spatial location difference data of adipose tissue. The reference marker set consists of predefined landmarks with fixed spatial locations in the standard anatomical coordinate system, which correspond to the anatomical features around or inside the adipose tissue.

[0204] Furthermore, feature points corresponding to the reference marker set are extracted from the spatially heterogeneous distribution field. By comparing the three-dimensional coordinates, the coordinate differences of each pair of corresponding points in the three dimensions are determined. These differences are then organized according to the spatial location of the corresponding points, and the resulting dataset is the spatial location difference data of adipose tissue.

[0205] Furthermore, a non-uniform deformation field of adipose tissue is constructed based on spatial location difference data, according to the coordinate difference value of each location in the spatial location difference data.

[0206] Furthermore, the degree and direction of deformation that need to occur at this location are determined. For locations with a difference of zero, the deformation is set to zero. For locations with a difference, the corresponding deformation is assigned according to the magnitude and direction of the difference, so that the deformation of any point in the deformation field is directly related to the spatial position difference data of that point.

[0207] Furthermore, by integrating these deformation information according to their spatial distribution, a field is formed that can describe the deformation patterns of different regions of adipose tissue, which is the non-uniform deformation field of adipose tissue.

[0208] Furthermore, spatial proximity weight correction is applied to the geometric distortion region in the non-uniform deformation field to obtain the calibrated deformation field in the non-uniform deformation field. The geometric distortion region is the region in the non-uniform deformation field where the difference in deformation between adjacent points exceeds a reasonable range. After identifying these regions...

[0209] Furthermore, select reasonable neighboring points around the area where the deformation has changed, and set weights based on the spatial distance between the neighboring points and the points in the distorted area. The closer the distance, the greater the weight. Use the deformation of the neighboring points to calculate the corrected deformation of the points in the distorted area according to the weights, and replace the original unreasonable deformation. Repeat this process until the deformation of all distorted areas is within a reasonable range. The resulting field is the calibration deformation field in the non-uniform deformation field.

[0210] Furthermore, by calibrating the deformation field to perform coordinate transformation on the spatial heterogeneous distribution field, a three-dimensional vector map of spatial structure quantification is obtained. Based on the deformation and direction of each position in the calibration deformation field, the three-dimensional coordinates of all points in the spatial heterogeneous distribution field are adjusted.

[0211] Furthermore, the original coordinates of each point are transformed to the standard anatomical coordinate system according to the corresponding deformation information, ensuring that the position of the transformed point in the standard anatomical coordinate system is consistent with the spatial relationship of the reference marker point set. The spatial structure information and quantitative data of all transformed points are integrated to form a three-dimensional vector map that can intuitively display the quantitative results of the spatial structure of adipose tissue.

[0212] In summary, the analysis of spatially heterogeneous distribution fields and reference marker sets yields data on spatial location differences, which can clarify the positional deviations between the two and provide a basis for the construction of deformation fields.

[0213] In summary, constructing a non-uniform deformation field based on spatial location difference data can accurately describe the changing patterns of spatial location.

[0214] In summary, correcting the geometrically distorted regions in a non-uniform deformation field yields a calibrated deformation field, which can eliminate the effects of distortion and ensure the accuracy of the deformation field.

[0215] In summary, by calibrating the deformation field to transform the spatial heterogeneous distribution field to obtain a three-dimensional vector map, the data can be unified to a standard coordinate system, thus realizing the quantification of spatial structure.

[0216] like Figure 2 The diagram shown is a functional block diagram of a three-dimensional reconstruction-based spatial structure quantification system for adipose tissue provided in an embodiment of the present invention.

[0217] The three-dimensional reconstruction-based adipose tissue spatial structure quantification system 100 of this invention can be installed in an electronic device. Depending on the functions implemented, the three-dimensional reconstruction-based adipose tissue spatial structure quantification system 100 may include a three-dimensional dynamic alignment module 101, a curvature verification module 102, a vector fusion module 103, a gradient coupling module 104, an orientation aggregation module 105, and a coordinate mapping module 106. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0218] In this embodiment, the functions of each module / unit are as follows:

[0219] The three-dimensional dynamic alignment module is used to dynamically align the anatomical boundaries of the structural images in adipose tissue with the metabolic features of the functional images in three-dimensional space to obtain the fusion data of the adipose tissue.

[0220] The curvature verification module is used to verify the curvature continuity of the adipose tissue surface based on the fusion data, and obtain the adipose tissue topological surface.

[0221] The vector fusion module is used to perform spatial vector fusion on adjacent segments in the fat topological surface to obtain the closed biological topological boundary of the fat tissue.

[0222] The gradient coupling module is used to couple the spatial grid nodes and tissue density gradient in the closed biological topological boundary with the density gradient direction to obtain the connection vector of the closed biological topological boundary.

[0223] The orientation aggregation module is used to perform spatial orientation aggregation on the connection vector to obtain the spatial heterogeneous distribution field of the adipose tissue.

[0224] The coordinate mapping module is used to map the spatial heterogeneity distribution field to a standard anatomical coordinate system to obtain a three-dimensional vector map of the quantified spatial structure.

[0225] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0226] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0227] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0228] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0229] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0230] Finally, it should be noted that the above 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 preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction, characterized in that, The method includes: S1. The anatomical boundaries of the structural images in adipose tissue are dynamically aligned with the metabolic features of the functional images in three-dimensional space to obtain the fusion data of the adipose tissue; S2. Based on the fused body data, the curvature continuity of the adipose tissue surface is verified to obtain the adipose tissue topological surface; S3. Spatial vector fusion is performed on adjacent segments in the adipose tissue topology surface to obtain the closed biological topological boundary of the adipose tissue; S4. Couple the spatial grid nodes and tissue density gradient in the closed biological topological boundary with the density gradient direction to obtain the connection vector of the closed biological topological boundary; S5. Perform spatial orientation aggregation on the connection vectors to obtain the spatial heterogeneous distribution field of the adipose tissue; S6. Map the spatial heterogeneity distribution field to a standard anatomical coordinate system to obtain a three-dimensional vector map of the quantified spatial structure.

2. The method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction as described in claim 1, characterized in that, The process of dynamically aligning the anatomical boundaries of structural images and the metabolic features of functional images in adipose tissue in three-dimensional space to obtain fused data of the adipose tissue includes: Spatial noise suppression is performed on the structural image to obtain a clear structural image of the adipose tissue; The metabolic activity region contours of the clear structural image are annotated to obtain the metabolic feature contours of the functional image. The metabolic feature contours are matched and registered with the contour boundaries of the clear structural image to obtain the fusion data of the adipose tissue.

3. The method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction as described in claim 1, characterized in that, The process of verifying the curvature continuity of the adipose tissue surface based on the fused data to obtain the adipose tissue topological surface includes: The fused body data is subjected to surface topology reconstruction to obtain the initial surface of the adipose tissue; A global curvature feature analysis is performed on the initial surface to obtain the curvature distribution profile of the initial surface; Based on the curvature distribution contour, the adipose tissue surface is continuously spatially modified to obtain the adipose tissue topological surface.

4. The method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction as described in claim 1, characterized in that, The step of fusing spatial vectors between adjacent segments in the adipose tissue topological surface to obtain the closed biological topological boundary of the adipose tissue includes: The fat topological surface is subjected to fragment spatial adjacency analysis to obtain the set of adjacent fragments of the fat topological surface; Spatial vector registration and calibration are performed on the adjacent segment set to obtain the calibration spatial vector group of the fat topological surface; Spatial orientation consistency fusion is performed on the calibration spatial vector group to obtain the closed biological topological boundary of the adipose tissue.

5. The method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction as described in claim 4, characterized in that, The step of performing spatial vector registration and calibration on the adjacent segment set to obtain the calibration spatial vector group of the fat topological surface includes: Differential manifold analysis is performed on the boundary feature point set of each segment in the adjacent segment set to obtain the curvature gradient tensor of the adjacent segment set; Based on the curvature continuity index, the spatial position difference of the boundary feature point set is geometrically constrained and registered to obtain the initial calibration vector of the fat topological surface. The initial calibration vector is spatially aligned with the reference coordinate system of the fat topology surface to obtain the intermediate calibration vector group of the fat topology surface. Spatial integration of the intermediate calibration vector group yields the calibration spatial vector group of the fat topological surface.

6. The method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction as described in claim 1, characterized in that, The process of coupling the spatial grid nodes and tissue density gradient in the closed biological topological boundary to obtain the connection vector of the closed biological topological boundary includes: The spatial grid nodes are mapped to the spatial location field of the tissue density gradient to obtain the density gradient representation of the spatial grid nodes; Based on the density gradient characterization, the spatial coordinates of the spatial grid nodes are oriented to obtain the corrected spatial nodes of the spatial grid nodes; Connecting adjacent correction space nodes yields the connection vector of the closed biological topological boundary.

7. The method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction as described in claim 1, characterized in that, The step of spatial orientation aggregation of the connection vectors to obtain the spatial heterogeneity distribution field of the adipose tissue includes: The connection vectors are divided into preset directional intervals to obtain a vector group set of the adipose tissue; The spatial density distribution matrix of the adipose tissue is obtained by performing differential statistics on the vector density of each group in the vector grouping set. The gradient variation characteristics of the spatial density distribution matrix are mapped to the spatial heterogeneity distribution field of the adipose tissue.

8. The method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction as described in claim 1, characterized in that, The step of performing differential statistics on the vector densities of each group in the vector grouping set to obtain the spatial density distribution matrix of the adipose tissue includes: Map the vector group set to the spatial coordinate-density value association sequence of the adipose tissue; Spatial interpolation is performed on the spatial coordinate-density value correlation sequence to obtain the continuous density distribution surface of the adipose tissue; The spatial location point density value is calculated based on the continuous density distribution surface, wherein the formula for calculating the spatial location point density value is as follows: In the formula, Here, h represents the spatial density value of the spatial location points, n is the number of vectors in the current group, and h is the number of vectors in the current group. 2 Let K be the two-dimensional Gaussian kernel function, x be the abscissa of the spatial coordinates, y be the ordinate of the spatial coordinates, h be the adaptive bandwidth parameter, d be the density value, and i be the vector index. The spatial density distribution matrix of the adipose tissue is obtained by performing grid topology reconstruction on the density values ​​of the spatial location points.

9. The method for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction as described in claim 1, characterized in that, The process of mapping the spatial heterogeneity distribution field to a standard anatomical coordinate system to obtain a three-dimensional vector map of the quantified spatial structure includes: Spatial correspondence analysis was performed on the spatial heterogeneous distribution field and the reference marker set to obtain the spatial location difference data of the adipose tissue; A non-uniform deformation field of the adipose tissue is constructed based on the spatial location difference data; Spatial proximity weight correction is applied to the geometrically distorted region in the non-uniform deformation field to obtain the calibrated deformation field in the non-uniform deformation field. The spatial heterogeneity distribution field is transformed by the calibration deformation field to obtain a three-dimensional vector map of the quantized spatial structure.

10. A system for quantifying the spatial structure of adipose tissue based on three-dimensional reconstruction, characterized in that, The system includes: A three-dimensional dynamic alignment module is used to dynamically align the anatomical boundaries of the structural images in adipose tissue with the metabolic features of the functional images in three-dimensional space to obtain the fused data of the adipose tissue. The curvature verification module is used to verify the curvature continuity of the adipose tissue surface based on the fused body data, and to obtain the adipose tissue topological surface. The vector fusion module is used to perform spatial vector fusion on adjacent segments in the fat topological surface to obtain the closed biological topological boundary of the fat tissue. The gradient coupling module is used to couple the spatial grid nodes and tissue density gradient in the closed biological topological boundary with the density gradient direction to obtain the connection vector of the closed biological topological boundary. The orientation aggregation module is used to perform spatial orientation aggregation on the connection vectors to obtain the spatial heterogeneous distribution field of the adipose tissue; The coordinate mapping module is used to map the spatial heterogeneity distribution field to a standard anatomical coordinate system to obtain a three-dimensional vector map of the quantified spatial structure.

Images