A female pelvic floor multi-target curve medical image cooperative registration and alignment method

Abstract

This invention discloses a method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor, comprising: S1: marking the target curve, SCIPP line, and segmentation markers in each sample data; S2: using a dimensionality reduction method combining Y-Z plane projection and X-coordinate dynamic compensation to reduce the dimensionality of the three-dimensional medical data to two dimensions; S3: smoothing the curves using cubic spline interpolation before registration to avoid the influence of noise in the original data on the registration accuracy; S4: calculating the translation matrix based on the reference point of the SCIPP line at the lower edge of the pubic symphysis, and aligning the lower edge of the first reference point at the lower edge of the pubic symphysis. S5: Move to the origin; S6: Use the SCIPP line as a reference vector and calculate the rotation matrix using the difference between the reference vector and the target angle; S7: Apply the corresponding translation and rotation matrices to the corresponding sample curve data to establish a unified coordinate system; S8: Calculate the average value of the curve segment based on the marker points, and then combine the average values ​​calculated segment by segment into the average value of the entire curve; S9: Calculate the centroid, local centroid, and global centroid of each curve, and translate the centroid of the corresponding curve to coincide with the local and global centroids of the corresponding curve to achieve alignment. This invention realizes the standardized registration, two-dimensional alignment, and anatomical segmented averaging analysis of female pelvic organ contour curve data, providing accurate quantitative basis for the diagnosis and clinical research of urological diseases.

Description

Technical Field

[0001] This invention belongs to the field of medical image data processing technology, specifically relating to a method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor, which is particularly suitable for synchronous morphological comparison analysis of multiple related curves in multiple patients and multiple physiological states. Background Technology

[0002] In medical imaging analysis, morphological changes in the female pelvic floor organs are an important basis for assessing physiological function or disease status. The morphology of these organs is usually described by multiple related curves, such as the contour curves of the bladder, urethra, and rectum. Clinically, the assessment of pelvic floor organ morphology mainly relies on doctors' visual observation and subjective comparison of MRI or CT images, lacking quantitative and standardized methods.

[0003] In existing technologies, the curve registration of pelvic floor organ morphology has the following main problems: Independent registration of a single curve: Registering multiple related curves separately without considering the inherent relative spatial relationships (such as distance and angle) between the curves may lead to distortion of the overall morphological analysis; Single registration dimension: It only eliminates some differences by translation or rotation, without combining centroid alignment to optimize overall shape consistency; Three-dimensional data processing is complex: the contours of pelvic organs are mostly three-dimensional data, and direct analysis is easily affected by noise, acquisition errors and pose differences, and there is a lack of effective dimensionality reduction and standardization methods. Insufficient anatomical segmentation analysis: Existing methods usually only focus on the overall outline or a single indicator, lacking segmentation analysis based on anatomical landmarks, and cannot perform fine quantification of different regions of organs, thus limiting the accuracy of disease diagnosis and functional assessment.

[0004] Therefore, a collaborative registration method that can simultaneously process multiple related curves is needed to standardize each curve under a unified coordinate system, while preserving the inherent relative spatial relationships between curves. This method, combined with techniques such as two-dimensional dimensionality reduction, centroid alignment, and piecewise averaging analysis, can provide reliable data for the quantitative analysis, clinical diagnosis, and scientific research of the multi-structural morphology of female pelvic floor organs. Summary of the Invention

[0005] This invention aims to overcome the shortcomings of existing technologies, such as distortion of relative relationships, single registration dimension, and lack of collaborative verification caused by independent registration of multiple curves. It provides a collaborative method that combines rigid registration and centroid alignment to achieve high-precision synchronous standardization of multiple related curves under multiple patients and multiple physiological states.

[0006] To address the aforementioned technical problems, this invention provides a method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor, comprising the following steps: S1: In each sample data, mark the target curve, SCIPP line, and segmentation markers; S2: A dimensionality reduction method combining YZ plane projection and X coordinate dynamic compensation is used to reduce three-dimensional medical data to two dimensions; S3: Before registration, the curve is smoothed by cubic spline interpolation to avoid the impact of noise in the original data on the registration accuracy; S4: Calculate the translation matrix based on the reference point of the SCIPP line, the lower edge of the pubic symphysis, and move the first reference point, the lower edge of the pubic symphysis, to the origin; S5: Use the SCIPP line as a reference vector and calculate the rotation matrix using the difference between the reference vector and the target angle; S6: Apply the corresponding translation and rotation matrices to the corresponding sample curve data to establish a unified coordinate system; S7: Calculate the average value of the curve segment by segment based on the marker points, and then combine the average values ​​calculated segment by segment into the average value of the entire curve. S8: Calculate the centroid of each curve, its local centroid, and its global centroid, and translate the centroid of the corresponding curve to coincide with its local and global centroids to achieve alignment.

[0007] Preferably, the specific steps of S1 in marking the target curve, SCIPP line, and segmentation markers in each sample data are as follows: Step 1: In 3D medical images (such as MRI, CT), manually or automatically label two types of data, including target curve data and SCIPP line data, and place markers at key locations in the target curve data. Multiple markers can be placed for each curve. Step 2: After annotation, save according to the format. Name the target curve as "Sample ID - Curve Name - State" and the SCIPP line file as "Sample ID - SCIPP - State". Ensure that the curves of the same sample in the same physiological state (such as resting state) are associated with the corresponding SCIPP lines.

[0008] Preferably, step S2 employs a dimensionality reduction method combining YZ plane projection and X-coordinate dynamic compensation to reduce the dimensionality of three-dimensional medical data to two dimensions. The specific steps are as follows: Step 1: Extract the 3D coordinates of all target curves and SCIPP lines; Step 2: Analyze the variation of the X coordinate, divide the standard deviation of the X coordinate by the standard deviation of the Y coordinate to obtain the variation ratio, calculate the average variation ratio of all curves, and evaluate the influence of X on the morphology; Step 3: Determine the value of the compensation factor k based on the calculated average variation ratio; Step 4: Convert the three-dimensional coordinates of the target curve and SCIPP line into two-dimensional coordinates.

[0009] Preferably, the specific steps for smoothing the curve using cubic spline interpolation before S3 registration to avoid the influence of original data noise on registration accuracy are as follows: Step 1: Sort the output two-dimensional discrete point coordinates by point index to ensure curve continuity; Step 2: For the sorted point set, calculate the distance between adjacent points, accumulate the chord length L = [0, d_1, d_1+d_2,..., Σd_i], and normalize it to [0,1] as the interpolation parameter t; Step 3: Construct a cubic spline function for the Y and Z coordinates to generate 300 uniformly distributed interpolation points. Then calculate the smoothed curve coordinates and output the smoothed two-dimensional curve. Preferably, the specific steps of S4, which calculates the translation matrix based on the reference point of the SCIPP line at the lower edge of the pubic symphysis and moves the first reference point at the lower edge of the pubic symphysis to the origin, are as follows: Step 1: Extract the coordinates of the first reference point P1 (lower edge of the pubic symphysis) from the smoothed SCIPP line 2D data; Step 2: Construct a translation matrix T to translate P1 to the origin (0,0) of the coordinate system; Step 3: Substitute the coordinates of the first reference point P1 (lower edge of the pubic symphysis) of the SCIPP line for each sample and each state into the translation matrix T to obtain the corresponding translation matrix.

[0010] Preferably, the specific steps of S5, which uses the SCIPP line as a reference vector and calculates the rotation matrix using the difference between the reference vector and the target angle, are as follows: Step 1: Extract the coordinates of the second reference point P2 (e.g., the end of the pubic ramus) from the 2D data of the SCIPP line; Step 2: Calculate the SCIPP reference vector a, and use this vector as the initial vector. Step 3: Calculate the angle α between the initial vector and the positive x-axis: Step 4: Calculate the rotation angle θ and the rotation matrix R: Rotate the vector clockwise to a horizontal position and use it as the Y-axis of the coordinate system. Generally, a rotation of 32° to 36° is required, depending on the actual sample effect. The corresponding angle for clockwise rotation is negative.

[0011] Preferably, the specific steps of S6 in applying the corresponding translation and rotation matrices to the corresponding sample curve data to establish a unified coordinate system are as follows: Step 1: First calculate the combined transformation matrix M, which is M = T × R; Step 2: Multiply each target curve data of the same sample and the same physiological state by its corresponding combination transformation matrix, perform coordinate transformation, and obtain the registered coordinates.

[0012] Preferably, the specific steps of S7, which calculates the average value of the curve segment by segment based on the marker points and then combines the average values ​​calculated segment by segment into the average value of the entire curve, are as follows: Step 1: Determine the segment boundaries, which are determined by the indices of the anatomical markers on the interpolated curve; Step 2: Then determine the number of segments. The number of segments is the number of marker points plus one. Step 3: Calculate the average value of each segment of each curve; Step 4: Finally, stitch the average point sequence of each segment into a complete curve. When stitching, the duplicate endpoints of adjacent segments need to be removed.

[0013] Preferably, the specific steps of S8 in calculating the centroid of each curve, its local centroid, and its global centroid, and then translating the centroid of the corresponding curve to coincide with its local and global centroids, are as follows: Step 1: Calculate the centroid of a single curve; Step 2: Calculate the global centroid; Step 3: Calculate the local centroid; Step 4: Move the centroid of the device to coincide with the local centroid, so that all samples of all curves are aligned under the same physiological state; move the centroid of the device to coincide with the global centroid, so that all samples of all curves (regardless of physiological state) are aligned.

[0014] The beneficial effects of this invention are as follows: A method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor, compared with the prior art, has the following advantages: Simultaneous standardization of multiple curves: By using the method of "cooperative rigid registration + individual centroid alignment", the position and orientation differences of each curve are eliminated under a unified coordinate system, while preserving the relative spatial relationship between curves; Improve the reliability of multi-structure analysis: avoid distortion of the relationship between curves caused by independent registration, and be applicable to the overall morphological analysis of multiple related curves, such as the assessment of the coordinated changes of multiple structures of pelvic floor organs; Higher registration accuracy: Combining rigid transformation (correcting global offset) and centroid alignment (optimizing local morphology) results in a better fit to the curve's morphological characteristics than a single registration method; It has a collaborative verification mechanism: by verifying the centroid spacing deviation, it ensures that the inherent relationship between the registered curves is not destroyed and the results are traceable; Adaptable to multiple scenarios: Curve sets for different patients and different physiological states can achieve standardized comparisons across patients and states, providing support for dynamic functional analysis. Attached Figure Description

[0015] Figure 1 This is a flowchart of a method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor provided by an embodiment of the present invention; Figure 2 This is a schematic diagram of the SCIPP line and target curve contour of the first patient in the Wardwell state, manually annotated in mimics, provided in an embodiment of the present invention. Figure 3 This is a schematic diagram showing the alignment of all samples of all curves under the same physiological state in an embodiment of the present invention; Figure 4 This is a schematic diagram showing the alignment of all samples (regardless of physiological state) of all curves in an embodiment of the present invention. Detailed Implementation

[0016] In order to clearly and completely describe the technical solution and technical effects of the present invention, the following embodiments are provided in detail.

[0017] Example 1 See attached document Figure 1 As shown, a method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor includes the following steps: S1. In each sample data, mark the target curve, SCIPP line, and segmentation markers; In this embodiment, Mimics software is used to mark the bladder-urethra, levator lamina, rectum, uterus-vaginal contour curves, SCIPP lines, and segmented marker points in the midsagittal plane. Each state—resting state, anal sphincter contraction state, and Warburg state—requires marking. (See attached diagram.) Figure 2 As shown in the diagram. The bladder urethral contour curve uses two marker points: the uppermost point of the trigone and the bladder neck. The levator palpebrae superioris contour curve uses the midpoint of the puborectalis muscle as a marker point. The rectal contour curve uses the vesicorectal junction as a marker point. The uterine-vaginal contour curve uses the internal and external cervical ossicles as marker points. The target curves are named "Sample ID - Curve Name - Status," and the SCIPP line files are named "Sample ID - SCIPP - Status," ensuring that curves for the same sample and physiological state are associated with the SCIPP line for that sample and physiological state. After marking, the data is exported in CSV format. S2. A dimensionality reduction method combining YZ plane projection and X coordinate dynamic compensation is used to reduce the dimensionality of three-dimensional medical data to two dimensions; In this embodiment, it is first necessary to analyze the variation of the X coordinate, calculate the standard deviation of the X coordinate and the standard deviation of the Y coordinate for each target curve, divide the standard deviation of the X coordinate by the standard deviation of the Y coordinate to obtain the variation ratio, then calculate the average variation ratio of all curves, evaluate the influence of the X coordinate on the shape, and then determine the value of the compensation factor k based on the calculated average variation ratio. If the average variation ratio is < 0.05 (X has a minimal impact): k = 0; If 0.05 ≤ mean variation ratio < 0.1 (X has a negligible effect): k = 0.05; If 0.1 ≤ average variation ratio < 0.2 (X has a moderate impact): k = 0.1; If the average variation ratio is ≥ 0.2 (X has a significant impact): k = 0.15; According to the formula: Y′=Y+(X norm×k×ystd) Where Xnorm=(X−Xmin) / (Xmax−Xmin), ystd is the standard deviation of the Y coordinate of the curve, and the three-dimensional coordinates (X,Y,Z) of the target curve and SCIPP line are converted into two-dimensional coordinates (Y', Z).

[0018] S3. Before registration, the curve is smoothed by cubic spline interpolation to avoid the influence of noise in the original data on the registration accuracy. In this embodiment, the output two-dimensional discrete point coordinates (Y', Z) are sorted by point index (to ensure curve continuity). For the sorted point set, the distance between adjacent points is calculated, and the cumulative chord length L = [0, d_1, d_1+d_2, ..., Σd_i] is normalized to [0,1] as the interpolation parameter t. A cubic spline function is constructed for the Y' and Z coordinates to generate 300 uniformly distributed interpolation points. Then, the smoothed curve coordinates are calculated, and the smoothed two-dimensional curve is output.

[0019] S4. Calculate the translation matrix based on the reference point of the SCIPP line at the lower edge of the pubic symphysis, and move the lower edge of the first reference point of the pubic symphysis to the origin. In this embodiment, the specific method for aligning the lower edge of the pubic symphysis with the origin includes the following steps: From the smoothed SCIPP line two-dimensional data, extract the coordinates (y1, z1) of the first reference point P1 (lower edge of the pubic symphysis); Construct a translation matrix T such that P1 is translated to the origin (0,0) of the coordinate system, as shown in the formula: T= Substitute the coordinates of the first reference point P1 (lower edge of the pubic symphysis) of the SCIPP line for each sample and each state into the translation matrix T to obtain the corresponding translation matrix.

[0020] S5. Use the SCIPP line as a reference vector and calculate the rotation matrix using the difference between the reference vector and the target angle. In this embodiment, the specific method for calculating the rotation matrix includes the following steps: Extract the coordinates (y2, z2) of the second reference point P2 (such as the end of the pubic ramus) from the two-dimensional data of the SCIPP line; Calculate the SCIPP reference vector 'a', and use this vector as the initial vector. a = [y2 - y1, z2 - z1] Calculate the angle α between the initial vector and the positive x-axis: α = arctan2( z2 - z1, y2 - y1) Calculate the rotation angle θ and the rotation matrix: The vector is rotated clockwise to a horizontal position and used as the Y-axis of the coordinate system (generally, a rotation of 32°~36° is required, depending on the actual sample effect). In this embodiment, a clockwise rotation of 34° is chosen, so the rotation angle θ = α−34° (the corresponding angle for clockwise rotation is negative). Construct the rotation matrix R, formula: R= S6. Apply the corresponding translation and rotation matrices to the corresponding sample curve data to establish a unified coordinate system; In this embodiment, the combined transformation matrix M is calculated first. M=T×R Right now M= Then, for each target curve data (smoothed two-dimensional point set) of the same sample and the same physiological state, multiply it by its corresponding combination transformation matrix, perform coordinate transformation, and obtain the registered coordinates.

[0021] S7. Calculate the average value of the curve segment by segment based on the marker points, and then combine the average values ​​calculated segment by segment into the average value of the entire curve. In this embodiment, the specific method for calculating the average value of the curve segmentally includes the following steps: Step 1: Determining the segment boundaries; the segment boundaries are determined by the indices of the anatomical markers on the interpolated curve. Then, determine the number of segments using the formula: M=n+1 Where M is the number of segments and n is the number of anatomical markers (e.g., 2 markers correspond to 3 segments).

[0022] Step 2: Formula for calculating the average point sequence of a single segment; For the p-th segment (p=1,2,...,M), the formula for calculating the coordinates of its average point sequence is: in, : The coordinates of the t-th average point in the p-th segment (t=0,1,...,L) p -1, L p K: The number of points in the p-th segment); K: The number of curves in the same group (e.g., if there are 12 curves in the resting bladder and urethra curve group, K=12); The y-coordinate of the j-th curve after alignment at the t-th point in the p-th segment; The z-coordinate of the j-th curve after alignment at the t-th point in the p-th segment; Step 3: Patch the average point sequence of each segment into a complete curve; when patching, the duplicate endpoints of adjacent segments need to be removed (only the starting endpoint of the first segment and the ending endpoint of the last segment are kept).

[0023] S8. Calculate the centroid of each curve, its local centroid, and its global centroid, and translate the centroid of the corresponding curve to coincide with its local and global centroids to achieve alignment. In this embodiment, the specific method for calculating the centroid of the curve and shifting the centroid of the corresponding curve to coincide with the local and global centroids of the corresponding curve includes the following steps: Step 1: Calculate the centroid of a single curve; for a curve of length n, its y The formulas for calculating the centroids of the z-axis and z-axis are as follows: in, : The centroid coordinates of the j-th curve; : y and z coordinates of the t-th point after interpolation of the j-th curve; n : uniform length of the interpolated curve.

[0024] Step 2: Calculation of the global centroid; For K curves of the same curve type, the formula for calculating the global centroid is: C Where C: global centroid coordinates of the curve type dimension; : The y and z coordinates of the centroid of the j-th curve; K: The total number of curves of the same type.

[0025] Step 3: Calculation of local centroids; K for a specific physiological state grouping of a certain curve type. g The formula for calculating the local centroid of a curve is: C l Among them, C l : Local centroid coordinates of the group; K g : The number of curves in this group.

[0026] Step 4: Align the centroids; refer to the appendix. Figure 3 As shown, the centroid is moved to coincide with the local centroid, so that all samples of all curves are aligned under the same physiological state; refer to the appendix. Figure 4 As shown, the centroid of the curve is moved to coincide with the global centroid, thus aligning all samples of all curves.

[0027] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor, characterized in that, Includes the following steps: S1: In each sample data, mark the target curve, SCIPP line, and segmentation markers; S2: A dimensionality reduction method combining YZ plane projection and X-coordinate dynamic compensation is used to reduce three-dimensional medical data to two dimensions; S3: Before registration, the curve is smoothed by cubic spline interpolation to avoid the impact of noise in the original data on the registration accuracy; S4: Calculate the translation matrix based on the reference point of the SCIPP line, the lower edge of the pubic symphysis, and move the first reference point, the lower edge of the pubic symphysis, to the origin; S5: Use the SCIPP line as a reference vector and calculate the rotation matrix using the difference between the reference vector and the target angle; S6: Apply the corresponding translation and rotation matrices to the corresponding sample curve data to establish a unified coordinate system; S7: Calculate the average value of the curve segment by segment based on the marker points, and then combine the average values ​​calculated segment by segment into the average value of the entire curve. S8: Calculate the centroid of each curve, its local centroid, and its global centroid, and translate the centroid of the corresponding curve to coincide with its local and global centroids to achieve alignment.

2. The method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor according to claim 1, characterized in that, The specific steps of S1 are as follows: Step 1: In the 3D medical image, manually or automatically label two types of data, including target curve data and SCIPP line data, and place markers at key locations in the target curve data, with multiple markers for each curve; Step 2: After annotation, save according to the format. Name the target curve as "Sample ID-Curve Name-State" and the SCIPP line file as "Sample ID-SCIPP-State" to ensure that the curves of the same sample and the same physiological state are associated with the corresponding SCIPP lines.

3. The method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor according to claim 1, characterized in that, The specific steps for converting three-dimensional to two-dimensional in S2 are as follows: Step 1: Extract the 3D coordinates of all target curves and SCIPP lines; Step 2: Analyze the variation of the X-coordinate, divide the standard deviation of the X-coordinate by the standard deviation of the Y-coordinate to obtain the variation ratio, calculate the average variation ratio of all curves, and evaluate the influence of X on the morphology. Step 3: Determine the value of the compensation factor k based on the average variation ratio calculated in Step 2. The selection rule is as follows: If the average variation ratio is <0.05, then k=0; If 0.05 ≤ average variation ratio < 0.1, then k = 0.05; If 0.1 ≤ average variation ratio < 0.2, then k = 0.1; If the average variation ratio is ≥0.2, then k=0.15; Step 4: Convert the three-dimensional coordinates of the target curve and SCIPP line into two-dimensional coordinates.

4. The method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor according to claim 1, characterized in that, The specific steps for interpolating the target curve data in S3 are as follows: Step 1: Sort the output two-dimensional discrete point coordinates by point index to ensure curve continuity; Step 2: For the sorted point set, calculate the distance between adjacent points and the cumulative chord length L, where L = [0, d1, d1+d2, ..., Σd i ], normalized to [0,1] as the interpolation parameter t; Step 3: Construct a cubic spline function for the Y and Z coordinates to generate 300 uniformly distributed interpolation points, then calculate the coordinates of the smoothed curve and output the smoothed two-dimensional curve.

5. The method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor according to claim 1, characterized in that, The specific steps for calculating the translation matrix in S4 are as follows: Step 1: Extract the coordinates of the first reference point P1, the lower edge of the pubic symphysis, from the smoothed SCIPP line 2D data; Step 2: Construct a translation matrix T to translate P1 to the origin (0,0) of the coordinate system; Step 3: Substitute the coordinates of the first reference point P1 of the SCIPP line at the lower edge of the pubic symphysis for each sample and each state into the translation matrix T to obtain the corresponding translation matrix.

6. The method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor according to claim 1, characterized in that, The specific steps for calculating the rotation matrix in S5 are as follows: Step 1: Extract the coordinates of the second reference point P2 from the SCIPP line 2D data; Step 2: Calculate the SCIPP reference vector a, and use this vector as the initial vector; Step 3: Calculate the angle α between the initial vector and the positive x-axis; Step 4: Calculate the rotation angle θ and the rotation matrix R; Rotate the vector clockwise to a horizontal position and use it as the Y-axis of the coordinate system. The rotation angle is 32°~36°, and the corresponding angle for clockwise rotation is negative.

7. The method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor according to claim 1, characterized in that, The specific steps for unifying the coordinate system in S6 are as follows: Step 1: First calculate the combined transformation matrix M, which is M = T × R; Step 2: Multiply each target curve data of the same sample and the same physiological state by its corresponding combination transformation matrix, perform coordinate transformation, and obtain the registered coordinates.

8. The method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor according to claim 1, characterized in that, The specific steps for calculating the average value of the segmented curve data in S7 are as follows: Step 1: Determine the segment boundaries, which are determined by the indices of the anatomical markers on the interpolated curve; Step 2: Then determine the number of segments. The number of segments is the number of marker points plus one. Step 3: Calculate the average value of each segment of each curve; Step 4: Finally, stitch the average point sequence of each segment into a complete curve. When stitching, the duplicate endpoints of adjacent segments need to be removed.

9. The method for collaborative registration and alignment of multi-target curve medical images of the female pelvic floor according to claim 1, characterized in that, The specific steps for centroid alignment in S8 are as follows: Step 1: Calculate the centroid of a single curve; Step 2: Calculate the global centroid; Step 3: Calculate the local centroid; Step 4: Align the centroids; Move its own centroid to coincide with the local centroid to align all samples of all curves under the same physiological state; move its own centroid to coincide with the global centroid to align all samples of all curves.

Images