Image sequence motion correction method and device based on intensity corrected optical flow registration

Abstract

The present specification relates to an image sequence motion correction method and device based on intensity correction optical flow registration. The method comprises: determining a reference image of each deformed image in a deformed image sequence, performing motion correction on the deformed image by adjusting the weight of the mechanical regularization term of the reference image and the deformed image to obtain the displacement field of the deformed image relative to the reference image; based on the displacement field and the intensity correction value of the finite element node, intensity correction is performed on the reference image; adjusting the weight of the mechanical regularization term, performing motion correction on the deformed image, intensity correction is performed on the first corrected reference image based on the optimized displacement field, and the execution is repeated until the image residual function converges, the deformed image is determined to be corrected, the optimal solution of the weight value of the mechanical regularization term of the deformed image is recorded; and determining the true displacement of the deformed image according to the optimal solution of the weight value. The present specification integrates the mechanical model behind physiological motion and intensity correction technology into the image registration algorithm to reduce image noise and improve algorithm accuracy.

Description

Technical Field

[0001] This specification relates to the field of image registration technology, and in particular to a method and apparatus for motion correction of image sequences based on intensity-corrected optical flow registration. Background Technology

[0002] Chemical exchange saturation transfer (CEST) magnetic resonance imaging (MRI) is a novel in vivo molecular imaging technique based on the exchange of saturated protons with surrounding flowing water. It provides crucial evidence for the diagnosis and treatment of diseases such as tumors, neurodegenerative diseases, and cardiovascular diseases, and has enormous potential for preclinical research and clinical application. However, during the long-duration CEST MRI scan, physiological movements (such as breathing and heartbeat) and body movements (such as head movement) inevitably occur, leading to misregistration between images when scanning corresponding organs, affecting the final image quality and diagnostic results.

[0003] Optical flow conservation-based registration methods are conventional registration methods used for motion correction. These methods assume that only motion exists between image sequences while intensity remains constant. However, the intensity differences in CEST MRI images at different frequencies are significant, rendering this assumption inapplicable. Traditional registration methods introduce regularization terms (such as L1 and L2 regularization) to improve registration accuracy. However, statistically based constraints can only handle part of the effects of image noise, cannot be physically explained, and cannot solve the problem of optical flow non-conservation. Summary of the Invention

[0004] To address the problem of inaccurate image registration caused by significant intensity and motion differences during the existing image registration process, this specification provides an image sequence motion correction method and apparatus based on intensity-corrected optical flow registration.

[0005] This specification provides an image sequence motion correction method based on intensity-corrected optical flow registration. The method includes: determining a reference image for each deformed image in a deformed image sequence, wherein the deformed image sequence includes multiple frames of deformed images caused by organ and tissue motion; performing motion correction on the deformed images by adjusting the weights of the mechanical regularization terms of the reference image and the deformed images to obtain a displacement field of the deformed images relative to the reference image, wherein the mechanical regularization terms are constructed based on the principle of mechanical equilibrium; performing intensity correction on the reference image based on the displacement field using a finite element shape function and intensity correction values ​​of finite element nodes to obtain a first corrected reference image; iteratively adjusting the weights of the mechanical regularization terms to perform motion correction on the deformed images to optimize the displacement field; performing intensity correction on the first corrected reference image based on the optimized displacement field; alternately repeating the process until the image residual function constrained by the mechanical regularization terms converges to a preset value, determining that the deformed image correction is complete, and recording the optimal solution of the weight values ​​of the mechanical regularization terms of the deformed images; and determining the true displacement of each deformed image relative to the reference image based on the optimal solution of the weight values.

[0006] According to one aspect of the embodiments of this specification, before adjusting the weight of the mechanical regularization term of the deformed image, the mechanical regularization term of the image is determined by: meshing the deformed image to obtain multiple finite element meshes of the deformed image; constructing the mechanical regularization term of the deformed image based on the mechanical equilibrium principle of organ and tissue deformation and the sum of forces on nodes inside any finite element mesh, as shown in the following formula: [K] represents the rectangular stiffness matrix of the finite element mesh and material parameters, and {u} represents the required nodal displacements. This represents the mechanical regularization term of the deformed image.

[0007] According to one aspect of the embodiments of this specification, motion correction of a deformed image by adjusting the weight of a mechanical regularization term of the deformed image includes: determining an initial weight of the mechanical regularization term as a first value; and performing initial motion correction on the deformed image under the weight of the mechanical regularization term at the first value to obtain an initial motion correction displacement field.

[0008] According to one aspect of the embodiments of this specification, intensity correction of the reference image using finite element shape functions and intensity correction values ​​of finite element nodes includes: meshing the reference image after initial motion correction to obtain multiple finite element meshes of the reference image; calculating the brightness correction value and contrast correction value of each node in each finite element mesh to perform intensity correction on the reference image, so as to reduce the intensity difference between the reference image and the deformed image.

[0009]

[0010]

[0011] Where b(x) represents the overall brightness correction field of the reference image, and c(x) represents the overall contrast correction field; θ k (x) represents the finite element shape function corresponding to the k-th node, b k c represents the brightness correction value of the k-th finite element node. k This represents the contrast correction value for the k-th finite element node.

[0012] According to one aspect of an embodiment of this specification, the image residual function constrained by a mechanical regularization term is determined by the following formula: in, Represents the image residual function, ω m This represents the weight of the applied mechanical regularization term, and its value is related to l. lreg 4 Proportional, l reg Indicates the regularization length. This represents the motion residual between the reference image and the deformed image; Represents the mechanical regularization term of the image; x represents the coordinates of the spatial location, I0(x) represents the pixel value in the reference image, and I n (x) represents the pixel value in the deformed image, and u(x,{v}) represents the deformed image I. n Each pixel in (x) is moved to the corresponding position in the reference image I0(x) by a displacement field, I n (x+u(x,{v})) represents the deformed image after correction by the displacement field u(x,{v}).

[0013] According to one aspect of an embodiment of this specification, iteratively adjusting the weight of the mechanical regularization term to perform motion correction on the deformed image to optimize the displacement field, and performing intensity correction on a first correction reference image based on the optimized displacement field, and alternately repeating the process until the image residual function constrained by the mechanical regularization term converges to a preset value includes: reducing the weight value of the mechanical regularization term; performing motion correction on the deformed image based on the reduced weight value to obtain an updated displacement field; performing intensity correction on the first correction reference image based on the updated displacement field to obtain a second correction reference image; calculating the value of the image residual function and determining whether the value of the image residual function has converged to a preset value; if yes, determining that the deformed image has been corrected and recording the weight value of the mechanical regularization term; if no, repeating the steps of reducing the weight value of the mechanical regularization term and performing intensity correction on the second correction reference image until the value of the image residual function converges to a preset value.

[0014] According to one aspect of the embodiments of this specification, determining a reference image for each deformed image in a deformed image sequence includes: calculating the structural similarity between each image in the magnetic resonance image sequence and a first frame image; forming a first image sequence group from multiple images with structural similarity higher than a first preset threshold; forming a second image sequence group from multiple images with structural similarity lower than a second preset threshold; using the first frame image in the first image sequence group as a reference image for each deformed image in the first image sequence group; and using the previous frame deformed image of each deformed image in the second image sequence group as a reference image for the deformed image.

[0015] According to one aspect of the embodiments of this specification, when the deformed image belongs to a second image sequence group, determining the true displacement of each deformed image relative to a reference image includes: determining the true displacement of the last frame image in the first image sequence group relative to its reference image, and recording the displacement as a first displacement; using a forward accumulation method to sequentially determine the true displacements of all deformed images in the second image group that are located before the deformed image relative to the reference image, and recording the true displacements as second displacements; adding the first displacement and the second displacement to obtain the true displacement of each deformed image in the second image sequence group relative to the reference image.

[0016] This specification provides an image sequence motion correction device based on intensity-corrected optical flow registration. The device includes: a reference image determination unit, used to determine a reference image for each deformed image in a deformed image sequence, wherein the deformed image sequence includes multiple frames of deformed images caused by organ and tissue motion; a motion correction unit, used to perform motion correction on the deformed images by adjusting the weights of the mechanical regularization terms of the reference image and the deformed images to obtain a displacement field of the deformed images relative to the reference image, wherein the mechanical regularization terms are constructed based on the principle of mechanical equilibrium and organ and tissue motion; an intensity correction unit, used to perform intensity correction on the reference image based on the displacement field, using a finite element shape function and the intensity correction value of the finite element nodes to obtain a first corrected reference image; and an iterative adjustment unit, used to iteratively adjust the weights of the mechanical regularization terms to perform motion correction on the deformed images to optimize the displacement field, and perform intensity correction on the first corrected reference image based on the optimized displacement field, repeating the process alternately until the image residual function constrained by the mechanical regularization terms converges to a preset value, determining that the deformed image has been corrected, and recording the optimal solution of the weight values ​​of the mechanical regularization terms of the deformed images.

[0017] The true displacement determination unit is used to determine the true displacement of each deformed image relative to the reference image based on the optimal solution of the weight values.

[0018] This specification provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the image sequence motion correction method based on intensity-corrected optical flow registration.

[0019] This specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the image sequence motion correction method based on intensity-corrected optical flow registration.

[0020] This manual incorporates physiological movements of human tissues or organs (such as heartbeat, respiration, and postural changes) into a mechanical regularization term. By integrating mechanical equilibrium conditions into the image registration algorithm as an additional constraint, it not only reduces image noise but also eliminates motion components that violate physical laws, significantly improving the algorithm's accuracy and interpretability. This manual also introduces intensity correction technology, which effectively handles drastic changes in image intensity. By adjusting the image's brightness and contrast, it effectively reduces or eliminates intensity distortion caused by various factors, solving the problem of image intensity changes coupled with motion. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 The diagram shown is a flowchart of an image sequence motion correction method based on intensity-corrected optical flow registration according to an embodiment of this specification.

[0023] Figure 2 The diagram shown is a flowchart of a method for determining the mechanical regularization term of an image according to an embodiment of this specification.

[0024] Figure 3 The diagram shown is a flowchart of a motion correction method according to an embodiment of this specification.

[0025] Figure 4 The diagram shown is a flowchart of a method for intensity correction of a reference image according to an embodiment of this specification.

[0026] Figure 5 The diagram shown is a flowchart of an iterative motion correction and intensity correction method according to an embodiment of this specification.

[0027] Figure 6The diagram shown is a flowchart of a method for determining a reference image for each deformed image according to an embodiment of this specification.

[0028] Figure 7 The diagram shown is a flowchart of a method for determining the true displacement of a deformed image relative to a reference image according to an embodiment of this specification.

[0029] Figure 8 The image shown is an example of an image sequence motion correction device based on intensity-corrected optical flow registration according to an embodiment of this specification;

[0030] Figure 9 The diagram shown is a schematic representation of motion correction of CEST MRI image sequences based on intensity-corrected optical flow registration according to an embodiment of this specification.

[0031] Figure 10 The diagram shown is a structural schematic of a computer device according to an embodiment of this specification.

[0032] Explanation of symbols in the attached drawings:

[0033] 801. Reference image determination unit;

[0034] 802. Motion correction unit;

[0035] 803, Strength Correction Unit;

[0036] 804. Iterative adjustment unit;

[0037] 805. Actual displacement determination element;

[0038] 1002. Computer equipment;

[0039] 1004, Processor;

[0040] 1006. Memory;

[0041] 1008. Drive mechanism;

[0042] 1010. Input / Output Module;

[0043] 1012. Input devices;

[0044] 1014. Output devices;

[0045] 1016. Presentation device;

[0046] 1018. Graphical User Interface;

[0047] 1020. Network interface;

[0048] 1022. Communication link;

[0049] 1024. Communication bus. Detailed Implementation

[0050] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0051] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0052] This specification provides the operational steps of the methods described in the embodiments or flowcharts, but based on conventional or non-inventive labor, more or fewer operational steps may be included. The order of steps listed in the embodiments is merely one possible execution order among many and does not represent the only possible execution order. In actual system or device products, the methods shown in the embodiments or drawings can be executed sequentially or in parallel.

[0053] It should be noted that the image sequence motion correction method based on intensity-corrected optical flow registration described in this specification can be applied to the field of magnetic resonance imaging and also to the field of medical technology. This specification does not limit the application fields of the image sequence motion correction method and device based on intensity-corrected optical flow registration.

[0054] Figure 1 The diagram shown is a flowchart of an image sequence motion correction method based on intensity-corrected optical flow registration according to an embodiment of this specification, which specifically includes the following steps:

[0055] Step 101: Determine a reference image for each deformed image in the deformed image sequence, wherein the deformed image sequence includes multiple frames of deformed images caused by the movement of organs and tissues.

[0056] In the embodiments described in this specification, during the long-term scanning process of the CEST MRI sequence, physiological and physical movements of the human body are unavoidable. This causes the images obtained during the scanning process to be distorted relative to the initial images. Multiple images obtained during the scanning process constitute an image sequence. Because multiple images in the image sequence are distorted to some extent, this image sequence is also called a deformed image sequence. The deformed image sequence contains multiple deformed images caused by the movement of human organs and tissues.

[0057] In the embodiments of this specification, in order to register each frame of deformed image, it is necessary to determine the corresponding reference image for each frame of deformed image. Traditional optical flow-based registration methods only select a certain frame of image as the reference image. However, there are drastic intensity changes between CEST MRI image sequences, which makes it impossible to use traditional registration methods between some registered images and reference images. Therefore, it is necessary to change the reference image, that is, to determine the corresponding reference image for each frame of deformed image.

[0058] Step 102: By adjusting the weights of the mechanical regularization terms of the reference image and the deformed image, motion correction is performed on the deformed image to obtain the displacement field of the deformed image relative to the reference image.

[0059] The mechanical regularization term is derived from the principle of mechanical equilibrium.

[0060] In the embodiments of this specification, considering the physical deformation characteristics of human tissues or organs, the mechanical equilibrium in the finite element biomechanical model is used as a constraint condition. This allows for a better simulation and understanding of the deformation observed in the image, and constitutes the mechanical regularization terms for the reference image and the deformed image. Specifically, corresponding weights are assigned to the mechanical regularization terms. By adjusting the weights of the mechanical regularization terms, motion correction is performed on the deformed image to obtain the displacement field of the deformed image relative to the reference image, where the displacement field is the motion field.

[0061] Step 103: Based on the displacement field, the reference image is subjected to intensity correction using the finite element shape function and the intensity correction value of the finite element nodes to obtain the first corrected reference image.

[0062] In the embodiments of this specification, due to the drastic intensity variations in the CEST image sequence—for example, the reference image in the CEST image sequence is very bright, while the deformed image is typically dark—it is necessary to perform intensity correction (i.e., brightness and contrast correction) on the reference image after coarse registration of the deformed image in the CEST image sequence, while keeping the displacement field between the deformed image and the reference image unchanged, to further reduce the residual between the reference image and the deformed image.

[0063] Step 104: Iteratively adjust the weights of the mechanical regularization term, perform motion correction on the deformed image to optimize the displacement field, and perform intensity correction on the first correction reference image based on the optimized displacement field. Repeat the process alternately until the image residual function constrained by the mechanical regularization term converges to a preset value, determine that the deformed image has been corrected, and record the optimal solution of the weights of the mechanical regularization term of the deformed image.

[0064] Based on steps 102 and 103, after coarse registration of the deformed image with the reference image and obtaining the intensity of the corrected reference image, the displacement field of the coarsely registered deformed image based on the reference image and the intensity of the corrected reference image are obtained. Based on this displacement field and the corrected reference image, the weight of the mechanical regularization term is iteratively adjusted to reduce its proportion in the image residual function, and the intensity of the reference image is further adjusted. Motion correction and intensity correction are performed alternately and iteratively until the image residual function finally converges.

[0065] In the embodiments described in this specification, a biomechanical model incorporating physiological movements of human tissues or organs (such as heartbeat, respiration, and postural changes) is used, employing mechanical regularization terms. By integrating mechanical equilibrium conditions into the image registration algorithm as an additional constraint, this method not only effectively reduces image noise but also eliminates motion components that violate physical laws, significantly improving the accuracy and interpretability of the algorithm.

[0066] Step 105: Based on the optimal solution of the weight values, determine the true displacement of each deformed image relative to the reference image. The weight value of the mechanical regularization term corresponding to the convergence of the image residual function is the optimal solution of the weight values. Based on this optimal solution, the mechanical regularization term can be determined. Further, the rectangular stiffness matrix and nodal displacements constituting the mechanical regularization term are determined, thereby determining the true displacement of the deformed image relative to the reference image.

[0067] This manual explains the mechanical model behind physiological movement and the intensity correction technology, as well as the image registration algorithm to reduce image noise and improve algorithm accuracy.

[0068] Figure 2 The diagram shown is a flowchart of a method for determining the mechanical regularization term of an image according to an embodiment of this specification, which specifically includes the following steps:

[0069] Step 201: Mesh the reference image to obtain multiple finite element meshes of the reference image. In the embodiments of this specification, each frame of the reference image is meshed, thereby each frame of the reference image can be divided into multiple finite element elements.

[0070] Step 202: Based on the deformation of organs and tissues and the mechanical equilibrium principle of nodes inside arbitrary finite element meshes, construct the mechanical regularization terms of the reference image and the deformed image.

[0071] In this step, the mechanical equilibrium principle of the nodes inside the finite element mesh is: the sum of the forces acting on any node inside the finite element mesh must be 0, that is, F = [K]{u} = 0. Here, [K] represents the rectangular stiffness matrix of the finite element mesh and material parameters, and {u} represents the required nodal displacement. However, under normal circumstances, due to factors such as noise, [K]{u} - F is not equal to 0. Based on this principle, the formula for the mechanical regularization term of the deformation image is as follows:

[0072] Where [K] represents the rectangular stiffness matrix of the finite element mesh and material parameters, and {u} represents the required nodal displacements. This represents the mechanical regularization term of the deformed image. The material parameters can be Young's modulus or Poisson's ratio, etc. In this specification, it is expected that the product of the rectangular stiffness matrix of the finite element mesh and material parameters and the nodal displacements will be minimized.

[0073] Figure 3 The diagram shown is a flowchart of a motion correction method according to an embodiment of this specification, which specifically includes the following steps:

[0074] Step 301: Determine the initial weight of the mechanical regularization term as the first value.

[0075] In the embodiments of this specification, the final residual function between the reference image and the deformed image is determined based on the residual function between the reference image and the deformed image and the mechanical regularization term of the deformed image in the traditional optical flow registration method. This residual function can also be referred to as the final residual function between the reference image and the deformed image. In this residual function, the mechanical regularization term has a corresponding adjustable weight parameter.

[0076] When motion correction is initially performed on the deformed image based on the reference image, the initial weight of the mechanical regularization term is adjusted to a first value, where the first value is a large value. In this case, the mechanical considerations of tissues and organs account for a large proportion in the final residual function between the reference image and the deformed image.

[0077] In the embodiments described in this specification, the weight of the mechanical regularization term can be determined by ω. m This indicates that its value is related to the regularization length l. reg 4 Proportional. In subsequent processing steps of this specification, the regularization length l is adjusted from large to small. reg 4 The value can be used to perform motion correction on deformed images from coarse to fine.

[0078] In the embodiments of this specification, the image residual function constrained by the mechanical regularization term is determined by the following formula:

[0079] in, Represents the image residual function, ωm This represents the weight of the applied mechanical regularization term, and its value is related to l. reg 4 Proportional, l reg Indicates the regularization length. This represents the motion residual between the reference image and the deformed image; This represents the mechanical regularization term of the image.

[0080] In the embodiments described in this specification, the motion residuals of the reference image and the deformed image It is calculated using the following formula:

[0081] x represents the coordinates of the spatial location, I0(x) represents the pixel value in the reference image, and I n (x) represents the pixel value in the deformed image, and u(x,{v}) represents the deformed image I. n Each pixel in (x) is moved to the corresponding position in the reference image I0(x) by a displacement field, I n (x+u(x,{v})) represents the deformed image after correction by the displacement field u(x,{v}).

[0082] To perform global registration, the least squares method is used to evaluate across the entire region of interest (ROI). The value of is shown in the following formula:

[0083]

[0084] However, using the least squares method to calculate When the minimum value is reached, the problem of no unique solution may arise. Therefore, the finite element method is often used to parameterize the displacement field u(x,{v}) to regularize and stabilize the problem. Within this framework, the registration problem can be transformed into a linear system for solution using the Gauss-Newton method.

[0085] [M]{δu}={b};

[0086] Where [M] is the Hessian matrix, {b} is the residual vector, and {δu} is the correction for the mesh node displacement during each iteration. Therefore, the registration problem of the deformed image is transformed into the problem of adjusting the displacement of the finite element nodes to minimize the residual vector {b}, thereby finding the optimal displacement field u(x,{v}).

[0087] Step 302: Under the weight of the mechanical regularization term of the first value, perform initial motion correction on the deformed image to obtain the displacement field of the initial motion correction.

[0088] In this step, when the weight of the mechanical regularization term is determined to be the first value, motion correction is performed on the first deformed image of a given intensity in the deformed image sequence using the weight of the first value. Because the first value is relatively large, the mechanical regularization term accounts for a large proportion in the image residual function. Therefore, the image residual function converges quickly, resulting in the coarsely registered motion field {u0}.

[0089] Figure 4 The diagram shown is a flowchart of a method for intensity correction of a reference image according to an embodiment of this specification, which specifically includes the following steps:

[0090] Step 401: Mesh the reference image to obtain multiple finite element meshes of the reference image.

[0091] Step 402: Calculate the brightness correction value and contrast correction value of each node in each finite element mesh, and perform intensity correction on the reference image to reduce the intensity difference between the reference image and the deformed image.

[0092] In this specification, the contrast correction value of the reference image is calculated using the following formula:

[0093]

[0094]

[0095] Where b(x) represents the overall brightness correction field of the reference image, and c(x) represents the overall contrast correction field of the reference image; θ k (x) represents the finite element shape function corresponding to the k-th node, b k c represents the brightness correction value of the k-th finite element node. k This represents the contrast correction value for the k-th finite element node.

[0096] In the embodiments of this specification, the reference image contains multiple nodes. The brightness correction values ​​of the finite element nodes corresponding to all nodes are multiplied by the finite element shape function to obtain the overall brightness correction field of the reference image; the contrast correction values ​​of the finite element nodes corresponding to all nodes in the reference image are multiplied by the finite element shape function to obtain the overall contrast correction field of the reference image. The overall contrast correction value and the overall brightness correction value of the reference image constitute the intensity correction value of the reference image.

[0097] In the embodiments of this specification, by adjusting the brightness and contrast of the image, intensity distortion caused by various factors is effectively reduced or eliminated, solving the problem caused by the coupling between image intensity changes and motion, and greatly improving the stability of the registration algorithm.

[0098] Figure 5The diagram shown is a flowchart of an iterative motion correction and intensity correction method according to an embodiment of this specification, which specifically includes the following steps:

[0099] Step 501: Reduce the weight value of the mechanical regularization term, and perform motion correction on the deformed image based on the reduced weight value to obtain the updated displacement field.

[0100] This step, based on steps 102 and 103, adjusts the weight value of the mechanical regularization term. The adjusted weight value of the mechanical regularization term is less than the first value. Using the same method as in step 102, the reduced weight is applied to correct the first reference corrected image and the deformed image. The displacement field registered between the deformed image and the first reference corrected image is obtained; this displacement field is considered an update to the displacement field of the deformed image relative to the reference image in step 102.

[0101] Step 502: Based on the updated displacement field, perform intensity correction on the first correction reference image to obtain the second correction reference image.

[0102] In this embodiment, the updated displacement field {u1} obtained in step 501 is used as a motion "seed" to fix the displacement field {u1}. The image intensity of the first correction reference image is corrected to obtain the corrected reference image.

[0103] Step 503: Calculate the value of the image residual function and determine whether the value of the image residual function converges to a preset value. Based on the currently updated displacement field {u1} and rectangular stiffness matrix [K], calculate the mechanical regularization term and the value of the image residual function, and further determine whether the value of the residual function is less than a preset value.

[0104] Step 504: If yes, confirm that the deformed image has been corrected and record the weight value of the mechanical regularization term.

[0105] Step 505: If not, repeat the steps of reducing the weight value of the mechanical regularization term and performing intensity correction on the second correction reference image until the value of the image residual function converges to the preset value.

[0106] In this step, if the value of the residual function is not less than the preset value, it indicates that the registration between the deformed image and the reference image is incomplete. The weight of the mechanical regularization term is further reduced, and the displacement fields of the deformed image and the second reference image are updated based on the new weight values. Based on this displacement field, the intensity of the second corrected reference image is corrected to obtain the third corrected reference image. This process is repeated continuously, adjusting the weight of the mechanical regularization term, correcting the deformed image based on the reference image according to the weights to update the displacement field, and performing intensity correction on the reference image based on the updated displacement field.

[0107] Figure 6The diagram shown is a flowchart of a method for determining a reference image for each deformed image according to an embodiment of this specification, which specifically includes the following steps:

[0108] Step 601: Calculate the structural similarity between each image in the magnetic resonance image sequence and the first frame image. In the embodiments of this specification, the deformed image sequence obtained by CEST MRI has significant changes in brightness and motion across multiple frames of deformed images. Therefore, it is not possible to determine the same reference image for all deformed images.

[0109] Therefore, this step calculates the structural similarity (SSIM, denoted as γ) between each image in the CEST MRI magnetic resonance image sequence and the first frame image in the image sequence.

[0110] Step 602: Combine multiple images with structural similarity higher than a first preset threshold into a first image sequence group.

[0111] In this step, the images are grouped into different sequence groups based on the calculated structural similarity and the value of a first preset threshold. Specifically, if the first preset threshold is set to 0.5, then based on the first preset threshold γ... cut The magnetic resonance image sequence is divided into two groups based on a first preset threshold of 0.5. Images with a deformation similarity greater than 0.5 are grouped into the first image sequence group. In an embodiment of this specification, for example, a CEST MRI sequence contains 60 deformed images. The deformed images from frame 2 to frame 10 show relatively small changes. If the structural similarity between these 10 images and the first frame is calculated to be greater than 0.5, then these 10 images constitute the first image sequence.

[0112] Step 603: Multiple images with structural similarity below a first preset threshold are grouped into a second image sequence group. If the first preset threshold is set to 0.5, deformed images with a first preset threshold less than 0.5 are grouped into the second image sequence group. The low structural similarity between the deformed images in the second image sequence group and the first frame image in the image sequence indicates that the differences in the deformed images in the second image sequence group are relatively drastic and significant. For example, in a CEST MRI sequence with 60 deformed images, the changes in the deformed images from frame 20 to frame 39 are drastic. The calculated structural similarity between these 20 frames and the first frame image is less than 0.5; therefore, these 20 frames constitute the second image sequence.

[0113] Step 604: The first frame image in the first image sequence group is used as the reference image for each deformed image in the first image sequence group. If γ>0.5, the first frame image I0 in the magnetic resonance image sequence is directly used as the reference image, and correlation analysis can be performed directly between it and the reference image, which is called direct digital image correlation (D-DIC) analysis.

[0114] Step 605: The previous frame of each deformed image in the second image sequence group is used as the reference image for that deformed image. If γ < 0.5, incremental digital image correlation (I-DIC) analysis is performed, and the previous image I... n-1 with I n Correlation analysis was performed to overcome the drastic intensity variations between deformed images.

[0115] The adaptive reference image selection strategy described in this specification can automatically select the most suitable reference image based on the structural similarity of the image sequence. Furthermore, the incremental calculation method employed effectively solves the problem of difficulty in matching the reference image and the image to be registered when there are significant intensity differences.

[0116] Figure 7 The diagram shown is a flowchart of a method for determining the true displacement of a deformed image relative to a reference image according to an embodiment of this specification, which specifically includes the following steps:

[0117] Step 701: Determine the actual displacement of the last frame image relative to its reference image in the first image sequence group, and record this displacement as the first displacement. Assume the last frame image is the m-th frame image in the first image sequence group, and record the first displacement as...

[0118] In the embodiments described in this specification, for example, a CEST MRI sequence contains 60 deformed images, wherein the deformed images from frame 1 to frame 20 constitute the first image sequence. The displacement of the 20th deformed image relative to the reference image is denoted as...

[0119] Step 702: Use the forward accumulation method to sequentially determine the true displacement of all deformed images in the second image group that are located before the deformed image relative to the reference image, and record the true displacement as the second displacement.

[0120] In the embodiments described in this specification, for example, the CEST MRI sequence contains 60 deformed images, of which frames 20 to 39 constitute the second image sequence. Therefore, it is necessary to determine the displacements of frames 20, 21, 22... to 38 relative to their respective reference images before the 39th deformed image. For example, the displacement of frame 21 relative to frame 20 is denoted as... The displacement of the 38th frame relative to the 37th frame is denoted as . Using the forward accumulation method, the displacements of all deformed images preceding the deformed image relative to the reference image are summed sequentially. This displacement is denoted as the second displacement, thus obtaining the displacement field.

[0121] Step 703: Add the first displacement and the second displacement to obtain the true displacement of each deformed image in the second image sequence group relative to the reference image.

[0122] Add the first displacement from step 702 to the second displacement from step 702 to obtain the true displacement of a deformed image relative to the reference image in a certain frame of the second image sequence group. Based on the example above, the displacement field... With displacement field By adding them together, the displacement field can be obtained. This indicates the actual displacement of the 38th frame image in the second image sequence group relative to the first frame image (reference image) in the CEST MRI sequence.

[0123] This manual's application is not limited to CEST MRI; it is equally applicable to images from Dynamic Contrast-Enhanced (DCE) MRI and PET-CT, which exhibit significant brightness or contrast changes during diagnosis. RI-DIC provides a rational physical interpretation through an integrated physical model, ensuring accurate motion correction. In DCE-MRI applications, this method effectively corrects for contrast agent-induced brightness variations over time, enhancing the reliability of imaging data, which is particularly important for lesion detection and treatment assessment. Similarly, in multimodal PET-CT imaging, RI-DIC ensures precise alignment between different imaging modalities, improving the matching between anatomical and functional images.

[0124] like Figure 8 The diagram shown is a schematic representation of an image sequence motion correction device based on intensity-corrected optical flow registration according to an embodiment of this specification. The basic structure of the device is illustrated in this diagram. The functional units and modules can be implemented in software, or using general-purpose chips or specific chips. The device specifically includes:

[0125] The reference image determination unit 801 is used to determine a reference image for each deformed image in the deformed image sequence, wherein the deformed image sequence includes multiple frames of deformed images caused by the movement of organs and tissues;

[0126] The motion correction unit 802 is used to perform motion correction on the deformed image by adjusting the weights of the mechanical regularization terms of the reference image and the deformed image, so as to obtain the displacement field of the deformed image relative to the reference image. The mechanical regularization terms are constructed based on the principle of mechanical equilibrium and the motion of organs and tissues.

[0127] The strength correction unit 803 is used to perform strength correction on the reference image based on the displacement field, using the finite element shape function and the strength correction value of the finite element node, to obtain a first corrected reference image;

[0128] The iterative adjustment unit 804 is used to iteratively adjust the weight of the mechanical regularization term, perform motion correction on the deformed image to optimize the displacement field, perform intensity correction on the first correction reference image based on the optimized displacement field, and repeat the process alternately until the image residual function constrained by the mechanical regularization term converges to a preset value, determine that the deformed image has been corrected, and record the optimal solution of the weight value of the mechanical regularization term of the deformed image.

[0129] The true displacement determination unit 805 is used to determine the true displacement of each deformed image relative to the reference image based on the optimal solution of the weight values.

[0130] This specification aims to introduce mechanical regularization terms, using biomechanical models as additional constraints to provide physical interpretability for the deformation of biological tissues and organs. It introduces intensity correction techniques, interleaving motion and image intensity corrections to address the effects of their coupling and improve registration accuracy. An adaptive incremental calculation method is used to optimize the reference image selection strategy, further improving image correlation and motion correction accuracy.

[0131] Figure 9 The diagram illustrates motion correction of a CEST MRI image sequence based on intensity-corrected optical flow registration, as described in this specification. The figure shows the differences between nine frames of images. Based on the structural similarity between each image and the first frame, the multiple frames are divided into two groups: the D-DIC group and the I-DIC group. Motion correction is performed on the D-DIC and I-DIC group image sequences respectively. After calculating the motion field of the deformed image relative to the reference image, the reference image is further corrected.

[0132] like Figure 10The diagram illustrates a computer device 1002 provided in an embodiment of this specification. The computer device 1002 may include one or more processors 1004, such as one or more central processing units (CPUs), each of which may implement one or more hardware threads. The computer device 1002 may also include any memory 1006 for storing information of any kind, such as code, settings, data, etc. Non-limitingly, for example, the memory 1006 may include any type of RAM, any type of ROM, flash memory, hard disk, optical disk, etc. More generally, any memory can use any technology to store information. Furthermore, any memory may provide volatile or non-volatile retention of information. Furthermore, any memory may represent a fixed or removable component of the computer device 1002. In one case, when the processor 1004 executes associated instructions stored in any memory or combination of memories, the computer device 1002 may perform any operation of the associated instructions. The computer device 1002 also includes one or more drive mechanisms 1008 for interacting with any memory, such as hard disk drive mechanisms, optical disk drive mechanisms, etc.

[0133] Computer device 1002 may also include an input / output module 1010 (I / O) for receiving various inputs (via input device 1012) and providing various outputs (via output device 1014). A specific output mechanism may include a presentation device 1016 and an associated graphical user interface (GUI) 1018. In other embodiments, the input / output module 1010 (I / O), input device 1012, and output device 1014 may be omitted, and the device may function solely as a computer device within a network. Computer device 1002 may also include one or more network interfaces 1020 for exchanging data with other devices via one or more communication links 1022. One or more communication buses 1024 couple the components described above together.

[0134] The communication link 1022 can be implemented in any way, such as via a local area network, a wide area network (e.g., the Internet), a point-to-point connection, or any combination thereof. The communication link 1022 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc., governed by any protocol or combination of protocols.

[0135] Corresponding to Figures 1 to 7 In addition to the methods described above, embodiments of this specification also provide a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the methods described above.

[0136] This specification also provides computer-readable instructions, wherein when a processor executes the instructions, the program therein causes the processor to perform the following... Figures 1 to 7 The method shown.

[0137] It should be understood that in the various embodiments of this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this specification.

[0138] It should also be understood that, in the embodiments of this specification, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this specification generally indicates that the preceding and following related objects have an "or" relationship.

[0139] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this specification can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this specification.

[0140] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0141] In the several embodiments provided in this specification, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the couplings or direct couplings or communication connections shown or discussed may be indirect couplings or communication connections through some interfaces, devices, or units, or they may be electrical, mechanical, or other forms of connection.

[0142] The units described as separate components may or may not be physically separate. The components shown as units 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 units can be selected to achieve the purpose of the embodiments described in this specification, depending on actual needs.

[0143] Furthermore, the functional units in the various embodiments of this specification 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 as a software functional unit.

[0144] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this specification, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this specification. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0145] This specification uses specific embodiments to illustrate the principles and implementation methods of this specification. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this specification. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this specification. Therefore, the content of this specification should not be construed as a limitation of this specification.

Claims

1. A motion correction method for image sequences based on intensity-corrected optical flow registration, characterized in that, The method includes: Determine a reference image for each deformed image in the deformed image sequence, wherein the deformed image sequence comprises multiple frames of deformed images caused by organ and tissue movement; By adjusting the weights of the mechanical regularization terms of the reference image and the deformed image, motion correction is performed on the deformed image to obtain the displacement field of the deformed image relative to the reference image. The mechanical regularization terms are based on the principle of mechanical equilibrium and are constructed accordingly. Based on the displacement field, the reference image is intensity-corrected using the finite element shape function and the intensity correction values ​​of the finite element nodes to obtain a first corrected reference image. The intensity correction of the reference image using the finite element shape function and the intensity correction values ​​of the finite element nodes includes: The reference image is meshed to obtain multiple finite element meshes of the reference image; Calculate the brightness correction value and contrast correction value of each node in each finite element mesh, and perform intensity correction on the reference image to reduce the intensity difference between the reference image and the deformed image; ;in, Represents the overall brightness correction field of the reference image. Indicates the overall contrast correction field; This represents the finite element shape function corresponding to the k-th finite element node. This represents the brightness correction value for the k-th finite element node. This represents the contrast correction value for the k-th finite element node. This represents the corrected reference image; The weights of the mechanical regularization term are iteratively adjusted to perform motion correction on the deformed image, thereby optimizing the displacement field. Based on the optimized displacement field, intensity correction is performed on the first correction reference image. This process is repeated alternately until the image residual function constrained by the mechanical regularization term converges to a preset value, indicating that the deformed image has been corrected. The optimal solution of the weight values ​​of the mechanical regularization term for the deformed image is recorded. The image residual function constrained by the mechanical regularization term is determined by the following formula: ;in, Represents the image residual function. This represents the weight of the applied mechanical regularization term, and its value is related to... Proportional Indicates the regularization length. This represents the motion residual between the reference image and the deformed image; The mechanical regularization term represents the deformed image. Represents the image residual function constrained by mechanical regularization terms; ; Coordinates representing spatial location Represents the pixel values ​​in the reference image. Represents the pixel values ​​in the deformed image. Indicates deformed image Each pixel in the image is moved to the reference image. The displacement field at the corresponding position in the middle. Indicates through displacement field The corrected deformed image; Based on the optimal solution of the weight values, the true displacement of each deformed image relative to the reference image is determined.

2. The image sequence motion correction method based on intensity-corrected optical flow registration according to claim 1, characterized in that, Before adjusting the weights of the mechanical regularization term for the deformed image, the mechanical regularization term for the image is determined as follows: The reference image is meshed to obtain multiple finite element meshes of the reference image; Based on the principles of organ and tissue deformation and the mechanical equilibrium of nodes within arbitrary finite element meshes, mechanical regularization terms are constructed for the reference image and the deformed image, as shown in the following formula: ;in, A rectangular stiffness matrix representing the finite element mesh and material parameters. Indicates the required nodal displacement. The mechanical regularization term of the deformed image is represented by t, which represents the transpose.

3. The image sequence motion correction method based on intensity-corrected optical flow registration according to claim 1, characterized in that, Motion correction of the deformed image and its reference image is performed by adjusting the weights of the mechanical regularization term of the deformed image, including: The initial weight of the mechanical regularization term is determined to be the first value; Under the weighted condition of the first numerical mechanical regularization term, the deformed image is subjected to initial motion correction to obtain the displacement field of the initial motion correction.

4. The method according to claim 1, characterized in that, Iteratively adjusting the weights of the mechanical regularization term, performing motion correction on the deformed image to optimize the displacement field, and performing intensity correction on the first correction reference image based on the optimized displacement field, repeating this process alternately until the image residual function constrained by the mechanical regularization term converges to a preset value, including: The weight of the mechanical regularization term is reduced, and motion correction is performed on the deformed image based on the reduced weight value to obtain the updated displacement field. Based on the updated displacement field, intensity correction is performed on the first correction reference image to obtain the second correction reference image; Calculate the value of the image residual function and determine whether the value of the image residual function converges to a preset value; If so, confirm that the deformed image has been corrected and record the weight value of the mechanical regularization term; If not, repeat the steps of reducing the weight of the mechanical regularization term and performing intensity correction on the second correction reference image until the value of the image residual function converges to the preset value.

5. The method according to claim 1, characterized in that, The reference image for each deformed image in the deformed image sequence includes: Calculate the structural similarity between each image in the magnetic resonance image sequence and the first frame image; Multiple images with a structural similarity higher than a first preset threshold are grouped into a first image sequence group; Multiple images with structural similarity below a first preset threshold are grouped into a second image sequence group; The first frame image in the first image sequence group is used as the reference image for each deformed image in the first image sequence group; In the second image sequence group, the previous frame of each deformed image is used as the reference image for the deformed image.

6. The method according to claim 5, characterized in that, When the deformed image belongs to the second image sequence group, determining the true displacement of each deformed image relative to the reference image includes: Determine the true displacement of the last frame image relative to its reference image in the first image sequence group, and record the true displacement as the first displacement. The forward accumulation method is used to determine the true displacement of all deformed images in the second image group that are located before the deformed image relative to the reference image, and the true displacement is recorded as the second displacement. The first displacement and the second displacement are added together to obtain the true displacement of each deformed image in the second image sequence group relative to the reference image.

7. An image sequence motion correction device based on intensity-corrected optical flow registration, characterized in that, The device includes: a reference image determination unit, used to determine a reference image for each deformed image in the deformed image sequence, wherein the deformed image sequence includes multiple frames of deformed images caused by organ and tissue movement; The motion correction unit is used to perform motion correction on the deformed image by adjusting the weights of the mechanical regularization terms of the reference image and the deformed image, so as to obtain the displacement field of the deformed image relative to the reference image. The mechanical regularization terms are constructed based on the principle of mechanical equilibrium and the motion of organs and tissues. An intensity correction unit is used to perform intensity correction on the reference image based on the displacement field, using a finite element shape function and intensity correction values ​​of finite element nodes, to obtain a first corrected reference image. The intensity correction of the reference image using the finite element shape function and intensity correction values ​​of finite element nodes includes: The reference image is meshed to obtain multiple finite element meshes of the reference image; Calculate the brightness correction value and contrast correction value of each node in each finite element mesh, and perform intensity correction on the reference image to reduce the intensity difference between the reference image and the deformed image; ;in, Represents the overall brightness correction field of the reference image. Indicates the overall contrast correction field; This represents the finite element shape function corresponding to the k-th finite element node. This represents the brightness correction value for the k-th finite element node. This represents the contrast correction value for the k-th finite element node. This represents the corrected reference image; An iterative adjustment unit is used to iteratively adjust the weights of the mechanical regularization term. The deformed image undergoes motion correction to optimize the displacement field. Based on the optimized displacement field, an intensity correction is performed on the first correction reference image. This process is repeated alternately until the image residual function constrained by the mechanical regularization term converges to a preset value, indicating that the deformed image has been corrected. The optimal solution of the weights of the mechanical regularization term in the deformed image is recorded. The image residual function constrained by the mechanical regularization term is determined by the following formula: ;in, Represents the image residual function. This represents the weight of the applied mechanical regularization term, and its value is related to... Proportional Indicates the regularization length. This represents the motion residual between the reference image and the deformed image; The mechanical regularization term represents the deformed image. Represents the image residual function constrained by mechanical regularization terms; ; Coordinates representing spatial location Represents the pixel values ​​in the reference image. Represents the pixel values ​​in the deformed image. Indicates deformed image Each pixel in the image is moved to the reference image. The displacement field at the corresponding position in the middle. Indicates through displacement field The corrected deformed image; The true displacement determination unit is used to determine the true displacement of each deformed image relative to the reference image based on the optimal solution of the weight values.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 6.

Images