Eye fundus image registration method and system based on two-stage parametric deformation modeling

Abstract

The invention provides an eye fundus image registration method and system based on two-stage parametric deformation modeling, and relates to the technical field of eye fundus image processing, and the method comprises the steps: obtaining a target image and a floating image, and extracting G channels of the target image and the floating image; based on the G channel of the target image and the floating image, a coarse-fine two-stage parameterization method is adopted for registration, and a registration image is obtained; wherein in the coarse registration stage, the mass centers of an optic disc and a fovea centralis and a blood vessel bifurcation point are detected, affine transformation is rapidly estimated through three pairs of feature points, and an initialized deformation field is obtained; in the fine registration stage, a multi-scale iterative optimization strategy is adopted to carry out parametric modeling on a gray scale transformation relation between an initialized deformation field and a target image on each scale, and a complete parametric deformation field is speculated through iterative parametric fitting; and performing up-sampling processing on the parameterized deformation field subjected to fitting speculation, and deforming an originally input floating image by using the up-sampled parameterized deformation field to obtain a final registration image.

Description

Technical Field

[0001] The present disclosure relates to the technical field of fundus image processing, and in particular to a fundus image registration method and system based on two-stage parameterized deformation modeling. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] Fundus diseases usually have no obvious clinical symptoms in the early stages and are easily overlooked, so professional fundus examination equipment is needed for accurate diagnosis. Fundus images, as the key basis for ophthalmic diagnosis, contain a wealth of medical information, including key pathological indicators such as vascular distribution characteristics, bleeding point locations, pigment deposition areas, and cyst morphology. This information provides an important reference for the diagnosis of a variety of ophthalmic diseases. However, fundus images need to be registered. Fundus image registration technology, as an efficient processing strategy, can not only assist doctors in comprehensively and quickly analyzing lesions and assist in improving diagnosis and treatment effects, but is also the core technology of the automatic auxiliary diagnosis system for fundus diseases based on artificial intelligence.

[0004] The challenges faced by fundus image registration mainly include poor image quality, modality differences, eye movement and projection distortion, pathological changes, large displacement and deformation, high computational complexity, lack of accurate evaluation criteria, multi-scale features, and interference from illumination and shadows, etc. These challenges make fundus image registration a complex and challenging task.

[0005] Some existing fundus registration schemes mostly use feature-based methods, intensity-based methods, and deep learning-based methods. However, the above existing registration methods still have the following problems: 1) Feature-based methods have poor robustness in low-quality images, small overlapping areas, or texture-missing scenes; 2) Intensity-based methods are sensitive to non-uniform illumination and random intensity variations, and have high computational complexity; 3) Existing evaluation methods rely on subjective judgment or manual labeling, and objective quantitative indicators are not robust; 4) Elastic registration methods (such as optical flow method) are inaccurate in estimating large displacements and rely on the assumption of constant grayscale, which has certain limitations for large-scale deformation registration. Summary of the invention

[0006] In order to solve the above problems, the present invention proposes a fundus image registration method and system based on two-stage parametric deformation modeling, and proposes a parametric modeling method for adaptively adjusting the complexity of the deformation field, and performs coarse-fine two-stage parametric deformation modeling registration. In the first stage, affine transformation is calculated by extracting three pairs of feature points, and in the second stage, multi-scale LAP iteration is performed, and local all-pass filter (LAP) is combined with polynomial displacement field model iterative optimization, which systematically solves the core defects of existing retinal image registration technology in terms of robustness, efficiency, accuracy and applicability, and provides an efficient and reliable automated tool for clinical pathological analysis.

[0007] According to some embodiments, the present disclosure adopts the following technical solutions: The fundus image registration method based on two-stage parametric deformation modeling includes: Obtain the target image and the floating image, downsample them, and extract the G channels of the target image and the floating image respectively; Based on the extracted target image and the G channel of the floating image, a coarse-fine two-stage parameterization method is used for registration to obtain the final registered image. In the coarse registration stage, the centroid of the optic disc and fovea and the vascular bifurcation point are detected, and the affine transformation is quickly estimated using three pairs of feature points to obtain the initial deformation field. In the fine registration stage, a multi-scale iterative optimization strategy is used to construct an inverted pyramid model. Based on the inverted pyramid model, the grayscale transformation between the initial deformation field and the target image is parameterized at each scale. The deformation field and grayscale transformation are iteratively calculated, and the complete parameterized deformation field is inferred by parametric fitting. The fitted and inferred parameterized deformation field is upsampled, and the original input floating image is deformed using the upsampled parameterized deformation field to obtain the final registered image.

[0008] According to some embodiments, the present disclosure adopts the following technical solutions: Fundus image registration system based on two-stage parametric deformation modeling, including: An image acquisition module is used to acquire a target image and a floating image, and downsample them to extract the G channels of the target image and the floating image respectively; A registration module is used to perform registration based on the extracted target image and the G channel of the floating image using a coarse-fine two-stage parameterization method to obtain the final registered image; In the coarse registration stage, the centroid of the optic disc and fovea and the vascular bifurcation point are detected, and the affine transformation is quickly estimated using three pairs of feature points to obtain the initial deformation field. In the fine registration stage, a multi-scale iterative optimization strategy is used to construct an inverted pyramid model. Based on the inverted pyramid model, the grayscale transformation between the initial deformation field and the target image is parameterized at each scale. The deformation field and grayscale transformation are iteratively calculated, and the complete parameterized deformation field is inferred by parametric fitting. The fitted and inferred parameterized deformation field is upsampled, and the original input floating image is deformed using the upsampled parameterized deformation field to obtain the final registered image.

[0009] According to some embodiments, the present disclosure adopts the following technical solutions: A computer program product comprises a computer program, wherein when the computer program is executed by a processor, the fundus image registration method based on two-stage parameterized deformation modeling is implemented.

[0010] According to some embodiments, the present disclosure adopts the following technical solutions: A non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the fundus image registration method based on two-stage parameterized deformation modeling is implemented.

[0011] According to some embodiments, the present disclosure adopts the following technical solutions: An electronic device comprises: a processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device implements the fundus image registration method based on two-stage parametric deformation modeling.

[0012] Compared with the prior art, the present invention has the following beneficial effects: The disclosed fundus image registration method based on two-stage parametric deformation modeling proposes a coarse-fine two-stage registration method. In the first stage, the centroid of the optic disc and fovea and the vascular bifurcation point are detected, and three pairs of feature points are used to quickly realize affine transformation initialization; in the second stage, multi-scale LAP iteration is performed, and iterative optimization is performed using a local all-pass filter (LAP) combined with a polynomial displacement field model. In order to solve the problem of grayscale differences between images, a quadratic polynomial is used to fit the illumination changes. In order to balance the registration efficiency and accuracy during the iteration process, a method for adaptively adjusting the number of iterations is proposed, which can improve the reliable measurement of alignment quality and has an indicator with low computational complexity.

[0013] The fundus image registration method based on two-stage parametric deformation modeling disclosed in the present invention proposes a parametric modeling method for adaptively adjusting the complexity of the deformation field, in view of the fact that the fixed 3D eyeball model cannot flexibly respond to morphological changes and has the problem of projection distortion. Experimental results show that on the FIRE and HRF datasets, the registration quality evaluation index SalC reaches more than 60%, the calculation time is shortened to seconds, and it is robust to images of different qualities, providing efficient and reliable technical support for clinical diabetic retinopathy analysis, image fusion and other applications.

[0014] The fundus image registration method based on two-stage parametric deformation modeling disclosed in the present invention, under the ideal situation that there is no grayscale difference between the target image and the floating image, the two are associated by a grayscale consistency equation, and a grayscale transformation function is introduced to describe the grayscale transformation between the target image and the floating image, so as to solve the problem that in the actual application of fundus image registration, the target image and the floating image are often difficult to meet the grayscale consistency due to the complexity of fundus lesions and the differences in shooting conditions.

[0015] The fundus image registration method based on two-stage parametric deformation modeling disclosed in the present invention solves the problem of local misregistration through explicit constraints, eliminates preprocessing and post-processing steps, improves algorithm efficiency, and makes it more suitable for fundus image registration scenarios with grayscale changes, pathological changes and large deformations. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure. The illustrative embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute an improper limitation on the present disclosure.

[0017] Figure 1 is the deformation field obtained by the LAP algorithm of the embodiment of the present disclosure; Figure 2 The deformation field after fitting with a second-order polynomial according to an embodiment of the present disclosure; Figure 3 A comparison diagram of the results of the method proposed by the present disclosure and other methods in the large-scale deformation subset of the FIRE dataset in an embodiment of the present disclosure; in, Figure 3 (a) in the figure is the target image. Figure 3 (b) in the figure is a floating image. Figure 3 (c) is the result of fusion of the target image and the floating image, where the target image is marked in orange and the floating image is marked in blue; Figure 3 (d) in the figure is the result of fusion of the registered image and the target image by the REMPE method; Figure 3 (e) in the figure is the result of fusion of the registered image and the target image by the Harris-PIIFD method; Figure 3(f) in the figure is the result of fusion of the registered image and the target image of the SURF-PIIFD-RPM method; Figure 3 (g) is the result of fusion of the registered image and the target image of the disclosed method; Figure 4 It is a flowchart of a fundus image registration method based on two-stage parameterized deformation modeling according to an embodiment of the present disclosure; Figure 5 A method for adaptively adjusting the number of iterations according to an embodiment of the present disclosure; Figure 6 It is a schematic diagram of the inverted pyramid model structure of an embodiment of the present disclosure. DETAILED DESCRIPTION

[0018] The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.

[0019] It should be noted that the following detailed descriptions are all illustrative and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.

[0020] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0021] Example 1 In one embodiment of the present disclosure, a fundus image registration method based on two-stage parameterized deformation modeling is provided, and the method steps include: Step 1: Obtain the target image and the floating image, downsample them, and extract the G channel of the target image and the floating image respectively; Step 2: Based on the extracted target image and the G channel of the floating image, a coarse-fine two-stage parameterization method is used for registration to obtain the final registered image; in the coarse registration stage, the centroid of the optic disc and the fovea and the vascular bifurcation point are detected, and the affine transformation is quickly estimated using three pairs of feature points to obtain the initial deformation field; in the fine registration stage, an inverted pyramid model is constructed using a multi-scale iterative optimization strategy, and the grayscale transformation between the initial deformation field and the target image is parameterized at each scale based on the inverted pyramid model, and the deformation field and grayscale transformation are iteratively calculated, and the complete parameterized deformation field is inferred by parameterized fitting; The fitted and inferred parameterized deformation field is upsampled, and the original input floating image is deformed using the upsampled parameterized deformation field to obtain the final registered image.

[0022] As an embodiment, the present invention discloses a fundus image registration method based on two-stage parametric deformation modeling. Due to the difference in shooting angles, many fundus images have significant deformations, and the overlapping area is small, which significantly increases the difficulty of registration. In order to achieve fast deformation field estimation with pixel-level accuracy in the case of large-scale deformation, the present invention proposes a two-stage image registration method of coarse registration-fine registration, such as Figure 4 The specific implementation process is as follows: Step 1: Get the target image and the floating image, downsample them, and extract the G channel of the target image and the floating image respectively; Specifically, with the development of fundus imaging technology, the image size usually exceeds 2000×2000 pixels, but in some cases the image contrast is low and the noise is high. In order to reduce the computational cost, the present disclosure downsamples the image to a size within 1000×1000 pixels.

[0023] Since the G channel of the fundus image has the strongest contrast, it can clearly show the distribution of blood vessels and background differences, while avoiding the problem of overbrightness of the R channel and low brightness and high noise of the B channel. Therefore, in the present disclosure, the G channel of the target image and the floating image are respectively used. Among them, the G channel is one of the three channels of the RGB color image.

[0024] Step 2: Based on the extracted target image and the G channel of the floating image, a coarse-fine two-stage parameterization method is used for registration to obtain the final registered image; Specifically, in the coarse registration stage, by detecting the centroids of the optic disc and fovea and the vascular bifurcation points, the affine transformation is quickly estimated using three pairs of feature points to obtain the initial deformation field; In the fine registration stage, a multi-scale iterative optimization strategy is adopted to construct an inverted pyramid model, and the deformation field and the grayscale transformation between the images to be registered (floating image and target image) are parameterized and modeled at each scale, and the deformation field and grayscale transformation are calculated iteratively. Figure 4 The blue dashed box in the middle describes a single ( j The process of iteration is different from the current mainstream method based on image resolution from coarse to fine. The present disclosure adopts a strategy that does not change the image resolution, such as Figure 6 As shown in the figure, by using the inverted pyramid model, the window size of the all-pass filter in the LAP algorithm is dynamically adjusted to achieve gradual optimization from large windows to small windows, thereby completing the progressive process from rough estimation to precise estimation of spatial deformation. It can quickly capture rough alignment, gradually refine to high-resolution layers, gradually correct small displacements, and improve registration accuracy and efficiency.

[0025] First, in the coarse registration stage, the centroids of the optic disc and fovea as well as the vascular bifurcation points are detected, and the affine transformation is quickly estimated using three pairs of feature points to obtain the initial deformation field. Specifically, in the coarse registration stage, the centroids of the optic disc and fovea are detected by a bandpass filter, the vascular bifurcation points are obtained by SIFT feature extraction, and feature screening is performed by the RANSAC method. Finally, a pair of points are randomly selected from the centroids of the optic disc and fovea, and two pairs of points are randomly selected from the vascular bifurcation points as the final feature points to estimate the affine transformation.

[0026] In the fine registration stage, the present disclosure imposes explicit constraints on the deformation field based on the LAP (Local All-Pass) algorithm, uses a linear parameterized model to represent the deformation field, and adaptively adjusts the complexity of the deformation field according to the image characteristics. The LAP algorithm regards the image displacement in the spatial domain as a phase change in the frequency domain through an all-pass filter, assumes that the local deformation field changes slowly and is approximately a constant displacement, and efficiently estimates the dense elastic deformation field through a linear system. The LAP algorithm has the advantages of fast speed and high accuracy when meeting the grayscale consistency assumption, but its limitations are that it is sensitive to grayscale differences, relies on preprocessing and postprocessing steps, and performs poorly under large displacement conditions, which may lead to excessive complexity of the deformation field and changes in the topological structure. In response to these problems, the present disclosure solves the problem of local misregistration through explicit constraints, eliminates preprocessing and postprocessing steps, improves the efficiency of the algorithm, and makes it more suitable for fundus image registration scenarios with grayscale changes, pathological changes, and large deformations.

[0027] Furthermore, the geometric deformation between images It can be expressed as a linear combination of a series of basis functions as follows: (1) in, are known basis functions, is the number of basis functions, are the coefficients of the basis functions. Therefore, calculating the two-dimensional deformation field is equivalent to solving Coefficient The type and number of basis functions determine the complexity of the deformation field. Polynomial functions are used to represent the spatial transformation between two fundus images. For example, as the order of the polynomial increases, the number of model parameters increases accordingly, and the deformation field that can be represented becomes more complex. Specifically, the constant term (0th-order polynomial) can describe the translation transformation between images, the first-order polynomial can characterize the affine transformation, the second-order polynomial can approximately simulate the transmission transformation, and the higher-order polynomial can characterize more complex deformations. However, in the fundus image registration task, the higher the model order (that is, the more parameters), the better. Too high an order will not only significantly increase the amount of calculation, but may also cause the local registration to be too fine, thereby changing the topological structure characteristics of the fundus image, such as the tortuosity of blood vessels, deformation of anatomical structures such as the optic disc and macula, but may also cause excessive distortion of non-overlapping areas, changing their anatomical structure, which in turn has a significant impact on clinical diagnosis.

[0028] The present disclosure adopts a polynomial model, but the proposed framework has wide applicability and supports a variety of "elastic" models based on global representation (such as higher-order polynomials, Fourier series) or local representation (such as B-spline basis, radial basis function), thereby providing flexible solutions for different application scenarios.

[0029] These coefficients are obtained by minimizing the parameterized model The dense deformation field obtained by LAP algorithm In the area The difference within is determined by: (2) Accurate deformation field estimation depends on reliable fitting regions , and the accurate fitting area depends on the accurate deformation field. To solve this interdependence problem, the present disclosure proposes to dynamically update the fitting area during the iteration process. In the initial iteration, the estimated fitting area usually deviates from the actual overlapping area, but these deviations will gradually decrease as the number of iterations increases.

[0030] The present disclosure only needs to utilize the accurate deformation of a partial area and infer the complete parameterized deformation field through parameterized fitting. This fitting strategy has the same effect as the repair and mean filtering of the deformation field in the LAP algorithm.

[0031] Traditional iterative registration methods use a strategy of fixed number of iterations, which can easily lead to insufficient registration accuracy or waste of computing resources. To solve this problem, the present disclosure proposes a method for adaptively adjusting the number of iterations. The core of this method is to find an indicator that can reliably measure the alignment quality and has low computational complexity. Figure 5As shown, the present invention extracts the reverse green channel from the stable vascular structure in the fundus image, innovatively uses Fourier-Argand filter and Gaussian filter to filter the reverse green channel image, and extracts the response value (filter difference) of the vascular area as a significant feature for calculating SalC, thereby improving the robustness of feature expression. In the coarse-to-fine registration framework, in addition to the preset maximum number of iterations at each scale, the number of iterations is also dynamically controlled by SalC: when SalC starts to decrease, the iteration stops and the previous result is returned, thereby balancing accuracy and efficiency.

[0032] In order to ensure the real-time performance of fundus image registration, the present disclosure uses a low-complexity fast shifted linear interpolation method to interpolate the floating image in each iteration process. This method can not only ensure computational efficiency, but also effectively maintain registration accuracy.

[0033] In the target image and floating images In an ideal situation where there is no grayscale difference between them, the two are related by the grayscale consistency equation, (3) in, is a two-dimensional deformation field. However, in the actual application of fundus image registration, due to the complexity of fundus lesions and the differences in shooting conditions, it is often difficult for the target image and the floating image to meet the assumption of grayscale consistency. In order to solve this problem, the present disclosure introduces the grayscale transformation relationship function To describe the target image With floating images Grayscale transformation between: (4) This function It may involve a combination of multiple complex factors such as convolution, non-uniform illumination changes, nonlinear pixel-level transformations, etc. Therefore, in order to accurately solve the deformation field , it is necessary to estimate the grayscale mapping function at the same time , which is a key step to achieve high-precision fundus image registration.

[0034] Similar to the parameterized model of the deformation field, the grayscale relationship between images can be represented by a linear combination of a series of basis functions: (5) Basis function It means that by minimizing the least square difference between the grayscale of the target image and the floating image in the field of view, it can be calculated Coefficient , thereby estimating the grayscale relationship between images.

[0035] For global grayscale changes, the following parameterized model can be used: (6) in, Represents the grayscale relationship between the target image and the floating image. In previous work, it has been verified that global illumination changes can be effectively represented by second-order polynomials. For local non-uniform grayscale changes, a linear combination of basis functions such as higher-order polynomial functions, uniform B-spline basis, radial basis functions, or Fourier series can be used to represent them. The selection of these basis functions can more flexibly adapt to the complexity of local grayscale changes, thereby improving the accuracy and robustness of grayscale relationship modeling.

[0036] Aiming at the nonlinear grayscale transformation between the images to be registered caused by fundus diseases and artifacts, this paper proposes a nonlinear grayscale transformation model based on polynomial expansion to construct the target image With floating images The nonlinear grayscale transformation between them is combined with local filtering to reduce the impact of local grayscale differences. The specific model is as follows: (7) in, is the model coefficient, is the model order. By introducing the polynomial expansion form, the model can better describe the nonlinear grayscale transformation between images, thereby effectively alleviating the impact of local grayscale differences on registration accuracy.

[0037] After completing the two stages of coarse registration and fine registration, the deformation field Afterwards, the deformation field is upsampled, and then the original input floating image is deformed using the upsampled deformation field to obtain the final registered image.

[0038] Simulation experiment Figure 1 , Figure 2 The comparison shows the dense elastic deformation field estimated by the LAP algorithm and the optimized deformation field after second-order polynomial fitting, where: Figure 1 and Figure 2 The different colors of the compass diagram in the upper right corner represent different directions of the displacement field, and the arrows also indicate the directions. It can be clearly seen that the polynomial fitting strategy for the dense elastic deformation field effectively replaces the post-processing step of the deformation field in LAP, and can correct the local misalignment problem of the LAP algorithm. Since no additional regularization terms are introduced in the objective function (Formula 2), its solution process is simplified to the calculation problem of a set of linear equations, which not only greatly reduces the computational complexity, but also ensures the uniqueness and optimality of the solution.

[0039] Figure 3 The comparison of the results of the method proposed in the present disclosure and other methods in the large-scale deformation subset of the FIRE dataset is shown. The image after registration and the target image are fused and annotated in orange and blue respectively. In addition, the control points are annotated in the figure. The closer the distance between the control points, the higher the registration accuracy. As shown in Table 1, the registration results of different methods in the FIRE dataset are compared.

[0040] Table 1 Comparison of registration results of different methods on the FIRE dataset

[0041] Example 2 In one embodiment of the present disclosure, a fundus image registration system based on two-stage parametric deformation modeling is provided, comprising: An image acquisition module is used to acquire a target image and a floating image, and downsample them to extract the G channels of the target image and the floating image respectively; A registration module is used to perform registration based on the extracted target image and the G channel of the floating image using a coarse-fine two-stage parameterization method to obtain the final registered image; In the coarse registration stage, the centroid of the optic disc and fovea and the vascular bifurcation point are detected, and the affine transformation is quickly estimated using three pairs of feature points to obtain the initial deformation field. In the fine registration stage, a multi-scale iterative optimization strategy is used to construct an inverted pyramid model. Based on the inverted pyramid model, the grayscale transformation between the initial deformation field and the target image is parameterized at each scale. The deformation field and grayscale transformation are iteratively calculated, and the complete parameterized deformation field is inferred by parametric fitting. The fitted and inferred parameterized deformation field is upsampled, and the original input floating image is deformed using the upsampled parameterized deformation field to obtain the final registered image.

[0042] Example 3 In one embodiment of the present disclosure, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the fundus image registration method based on two-stage parameterized deformation modeling is implemented.

[0043] Example 4 In one embodiment of the present disclosure, a non-transitory computer-readable storage medium is provided, wherein the non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the fundus image registration method based on two-stage parametric deformation modeling is implemented.

[0044] Example 5 In one embodiment of the present disclosure, an electronic device is provided, comprising: a processor, a memory, and a computer program; wherein the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes the fundus image registration method based on two-stage parametric deformation modeling.

[0045] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0046] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0047] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A fundus image registration method based on two-stage parametric deformation modeling, characterized in that: include: Obtain the target image and the floating image, downsample them, and extract the G channels of the target image and the floating image respectively; Based on the extracted target image and the G channel of the floating image, a coarse-fine two-stage parameterization method is used for registration to obtain the final registered image. In the coarse registration stage, the centroid of the optic disc and fovea and the vascular bifurcation point are detected, and the affine transformation is quickly estimated using three pairs of feature points to obtain the initial deformation field. In the fine registration stage, a multi-scale iterative optimization strategy is used to construct an inverted pyramid model. Based on the inverted pyramid model, the grayscale transformation between the initial deformation field and the target image is parameterized at each scale. The deformation field and grayscale transformation are iteratively calculated, and the complete parameterized deformation field is inferred by parametric fitting. The fitted and inferred parameterized deformation field is upsampled, and the original input floating image is deformed using the upsampled parameterized deformation field to obtain the final registered image.

2. The fundus image registration method based on two-stage parametric deformation modeling as claimed in claim 1, characterized in that: The target image and floating image are acquired and downsampled to a size of 1000×1000 pixels or less. The G channel of the fundus image has the strongest contrast and can clearly show the vascular distribution and background differences, while avoiding excessive brightness of the R channel and low brightness and high noise of the B channel. The G channels of the target image and floating image are extracted respectively.

3. The fundus image registration method based on two-stage parametric deformation modeling as claimed in claim 1, characterized in that: The coarse-fine two-stage parameterization method explicitly constrains the deformation field in both stages, constructs an objective function without additional constraints, transforms the image registration problem into a linear optimization problem, obtains a closed-form solution, and adopts a strategy that does not change the image resolution, that is, by dynamically adjusting the window size of the all-pass filter in the multi-scale iterative optimization algorithm, a gradual optimization from large windows to small windows is achieved, thereby completing the progressive process from coarse estimation to fine estimation of spatial deformation.

4. The fundus image registration method based on two-stage parametric deformation modeling as claimed in claim 1, characterized in that: In the fine registration stage, based on the multi-scale iterative optimization algorithm, explicit constraints are imposed on the deformation field, a linear parameterized model is used to represent the deformation field, and the complexity of the deformation field is adaptively adjusted according to the image characteristics. The multi-scale iterative optimization algorithm uses an all-pass filter to regard the image displacement in the spatial domain as a phase change in the frequency domain. It is assumed that the local deformation field changes slowly and is approximately a constant displacement, and the dense elastic deformation field is efficiently estimated through a linear system.

5. The fundus image registration method based on two-stage parametric deformation modeling as claimed in claim 1, characterized in that: In the ideal case where there is no grayscale difference between the target image and the floating image, the target image and the floating image are related by the grayscale consistency equation, and a grayscale transformation function is introduced to describe the target image. With floating images Grayscale transformation between: Among them, the function It involves a combination of complex factors such as convolution, non-uniform illumination changes, or non-linear pixel-level transformations. is a two-dimensional deformation field.

6. The fundus image registration method based on two-stage parametric deformation modeling as claimed in claim 5, characterized in that: Aiming at the nonlinear grayscale transformation between target images caused by fundus diseases and artifacts, a nonlinear grayscale mapping model based on polynomial expansion is constructed. With floating images The nonlinear grayscale mapping relationship between them is combined with local filtering to reduce the impact of local grayscale differences. The specific model is as follows: in, is the model coefficient, is the model order.

7. A fundus image registration system based on two-stage parametric deformation modeling, characterized in that: include: An image acquisition module is used to acquire a target image and a floating image, and downsample them to extract the G channels of the target image and the floating image respectively; A registration module is used to perform registration based on the extracted target image and the G channel of the floating image using a coarse-fine two-stage parameterization method to obtain the final registered image; In the coarse registration stage, the centroid of the optic disc and fovea and the vascular bifurcation point are detected, and the affine transformation is quickly estimated using three pairs of feature points to obtain the initial deformation field. In the fine registration stage, a multi-scale iterative optimization strategy is used to construct an inverted pyramid model. Based on the inverted pyramid model, the grayscale transformation relationship between the initial deformation field and the target image is parameterized at each scale. The deformation field and grayscale transformation are iteratively calculated, and the complete parameterized deformation field is inferred by parametric fitting. The fitted and inferred parameterized deformation field is upsampled, and the original input floating image is deformed using the upsampled parameterized deformation field to obtain the final registered image.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the fundus image registration method based on two-stage parameterized deformation modeling described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by the processor, the fundus image registration method based on two-stage parametric deformation modeling as described in any one of claims 1-6 is implemented.

10. An electronic device, characterized in that: include: A processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes the fundus image registration method based on two-stage parametric deformation modeling as described in any one of claims 1-6.

Images