CN112967236A

CN112967236A - Image registration method and device, computer equipment and storage medium

Info

Publication number: CN112967236A
Application number: CN202110189069.3A
Authority: CN
Inventors: 高菲菲; 薛忠; 董昢; 曹晓欢
Original assignee: Shanghai United Imaging Intelligent Healthcare Co Ltd
Current assignee: Lianying Intelligent Medical Technology Beijing Co ltd; Shanghai United Imaging Intelligent Healthcare Co Ltd
Priority date: 2018-12-29
Filing date: 2018-12-29
Publication date: 2021-06-15
Anticipated expiration: 2038-12-29
Also published as: CN109754396A; CN109754396B; CN112967236B

Abstract

The application relates to an image registration method, an image registration device, a computer device and a storage medium. The method comprises the following steps: acquiring a reference image and a floating image to be registered; acquiring a reference anatomy mark point set to be registered corresponding to a reference image and a floating anatomy mark point set to be registered corresponding to a floating image; determining the intersection of the mark points according to the matching result of the names of the mark points in the reference anatomical mark point set to be registered and the mark points in the floating anatomical mark point set to be registered; respectively determining an initial reference anatomical mark point set and an initial floating anatomical mark point set from the reference anatomical mark point set to be registered and the floating anatomical mark point set to be registered according to the mark point intersection; and performing at least one stage of image registration on the reference image and the floating image according to the initial reference anatomical marker point set, the initial floating anatomical marker point set and the anatomical marker point-based registration model. The method greatly improves the application range of image registration.

Description

Image registration method and device, computer equipment and storage medium

The patent application of the invention is a divisional application of Chinese patent application with application date of 2018, 12 and 29, application number of 201811637721.8 and name of image registration method, device, computer equipment and storage medium.

Technical Field

The present application relates to the field of image processing technologies, and in particular, to an image registration method and apparatus, a computer device, and a storage medium.

Background

The image registration can realize matching and overlaying of two or more images acquired at different times and under different imaging devices or under different conditions, for example, images such as a Computed Tomography (CT) image and a Positron Emission Tomography (PET) image can be matched and overlaid to display information of the CT image and information of the PET image participating in registration on the same image, so that a good auxiliary effect is provided for clinical medical diagnosis, and the method is a key technology in the field of image processing.

In the conventional technology, if a Region Of Interest (ROI) is an irregular Region, the irregular Region in an image to be registered is extracted, and registration is performed based on the irregular Region; and if the ROI is a key point, extracting the key point in the image to be registered, and registering based on the key point.

However, in the conventional technology, when image registration is performed, registration of an image to be registered can only be performed based on single semantic information, such as an irregular region or a key point, and the like, so that the conventional registration method has a low application range.

Disclosure of Invention

Based on this, it is necessary to provide an image registration method, an apparatus, a computer device, and a storage medium, for solving the problem that the conventional technology can only register the registered image based on single semantic information such as irregular regions or key points, which results in a low application range of the conventional registration method.

In a first aspect, an embodiment of the present application provides an image registration method, which may include:

acquiring a reference image and a floating image to be registered;

extracting semantic information from the reference image and the floating image to obtain a marked reference image and a marked floating image which comprise the semantic information;

determining target image registration models respectively corresponding to the marked reference image and the marked floating image from preset image registration models according to the semantic information;

and carrying out image registration on the reference image and the floating image according to the semantic information and the target image registration model.

In one embodiment, the semantic information includes: at least one of a segmentation region and anatomical landmark points of the floating image, and at least one of a segmentation region and anatomical landmark points of the reference image; the preset image registration model comprises a segmentation-based image registration model and an anatomical marker point-based registration model.

In one embodiment, when the target image registration model is the anatomical landmark point-based registration model, the image registering the reference image and the floating image according to the semantic information and the target image registration model includes:

acquiring a reference anatomy mark point set to be registered of the marked reference image and a floating anatomy mark point set to be registered of the marked floating image;

and carrying out image registration on the reference image and the floating image according to the reference anatomical marking point set to be registered, the floating anatomical marking point set to be registered and the registration model based on the anatomical marking points.

In one embodiment, the image registration of the reference image and the floating image according to the reference anatomical marker point set to be registered, the floating anatomical marker point set to be registered, and the anatomical marker point-based registration model includes:

determining marker point intersection according to the matching result of the names of the marker points in the reference anatomical marker point set to be registered and the marker points in the floating anatomical marker point set to be registered;

according to the mark point intersection, respectively determining an initial reference anatomical mark point set and an initial floating anatomical mark point set from the reference anatomical mark point set to be registered and the floating anatomical mark point set to be registered;

and carrying out image registration on the reference image and the floating image according to the initial reference anatomical mark point set, the initial floating anatomical mark point set and the anatomical mark point-based registration model.

In one embodiment, when the target image registration model is the segmentation-based image registration model, the image registering the reference image and the floating image according to the semantic information and the target image registration model includes:

acquiring a segmentation reference image corresponding to the marking reference image and a segmentation floating image corresponding to the floating image;

and performing image registration on the reference image and the floating image according to the segmentation reference image, the segmentation floating image and the segmentation-based image registration model.

In one embodiment, the method further comprises:

acquiring a registration result after image registration is carried out on the reference image and the floating image;

and carrying out image integration on the registration result according to the registration result and a preset image integration model.

In one embodiment, after the image registering the reference image and the floating image, the method further comprises:

acquiring the target transformation matrix;

determining a similarity metric value between the downsampled reference image and a transformed floating image corresponding to the downsampled floating image according to the target transformation matrix, the downsampled reference image obtained by downsampling the reference image and the downsampled floating image obtained by downsampling the floating image;

performing at least one operation of translation operation, rotation operation, tilting operation and scaling operation on the target transformation matrix, and extracting initial parameters corresponding to the target transformation matrix;

and determining a target parameter according to the similarity metric value, the initial parameter and a preset gradient descent method.

In a second aspect, an embodiment of the present application provides an image registration apparatus, which may include:

the first acquisition module is used for acquiring a reference image and a floating image to be registered;

the first extraction module is used for extracting semantic information from the reference image and the floating image to obtain a marked reference image and a marked floating image which comprise the semantic information;

the first determining module is used for determining target image registration models respectively corresponding to the mark reference image and the mark floating image from preset image registration models according to the semantic information;

and the registration module is used for carrying out image registration on the reference image and the floating image according to the semantic information and the target image registration model.

In a third aspect, an embodiment of the present application provides a computer device, where the computer device includes a memory and a processor, the memory stores a computer program, and the processor executes the computer program, and performs the following steps:

acquiring a reference image and a floating image to be registered;

In a fourth aspect, an embodiment of the present application provides a readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the following steps:

acquiring a reference image and a floating image to be registered;

In the image registration method, the image registration apparatus, the computer device and the readable storage medium provided by this embodiment, the computer device may acquire a reference image and a floating image to be registered; extracting semantic information from the reference image and the floating image to obtain a marked reference image and a marked floating image which comprise the semantic information; determining target image registration models respectively corresponding to the marked reference image and the marked floating image from preset image registration models according to semantic information; and finally, carrying out image registration on the marked reference image and the marked floating image according to the semantic information and the target image registration model. In the embodiment, the computer device can extract the semantic information of the reference image and the floating image firstly, so that the reference image and the floating image are registered by adopting different target image registration models according to different semantic information to complete the registration of the reference image and the floating image comprising various semantic information, the limitation that the reference image and the floating image can only be registered based on single semantic information in the prior art is solved, and the application range of image registration is greatly improved.

Drawings

FIG. 1 is a schematic diagram of an internal structure of a computer device according to an embodiment;

FIG. 2 is a flowchart illustrating an image registration method according to an embodiment;

fig. 3 is a schematic flowchart of an image registration method according to another embodiment;

FIG. 4 is a flowchart illustrating an image registration method according to yet another embodiment;

FIG. 5 is a flowchart illustrating an image registration method according to yet another embodiment;

FIG. 6 is a flowchart illustrating an image registration method according to yet another embodiment;

FIG. 7 is a schematic structural diagram of an image registration apparatus according to an embodiment;

fig. 8 is a schematic structural diagram of an image registration apparatus according to another embodiment;

fig. 9 is a schematic structural diagram of an image registration apparatus according to yet another embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

The image registration method provided by the embodiment of the application can be applied to computer equipment shown in fig. 1. The computer device comprises a processor and a memory connected by a system bus, wherein a computer program is stored in the memory, and the steps of the method embodiments described below can be executed when the processor executes the computer program. Optionally, the computer device may further comprise a network interface, a display screen and an input device. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a nonvolatile storage medium storing an operating system and a computer program, and an internal memory. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. Optionally, the computer device may be a server, a Personal Computer (PC), a personal digital assistant (pda), other terminal devices such as a tablet computer (PAD), a mobile phone, and the like, a cloud end or a remote server, and the specific form of the computer device is not limited in the embodiment of the present application.

It should be noted that, in the image registration method provided in the embodiment of the present application, an implementation subject may be an image registration apparatus, and the image registration apparatus may be implemented as part or all of a computer device by software, hardware, or a combination of software and hardware. In the following method embodiments, the execution subject is a computer device as an example.

Fig. 2 is a flowchart illustrating an image registration method according to an embodiment. The embodiment relates to a realization process that computer equipment determines an image registration model according to semantic information extracted from a reference image and a floating image and performs image registration on the reference image and the floating image. As shown in fig. 2, the method may include:

s202, acquiring a reference image and a floating image to be registered.

Specifically, the reference image and the floating image may be images of the same modality or images of different modalities, for example, both the reference image and the floating image may be CT images, or one may be a CT image and the other is a PET image. Alternatively, the computer device may register the two or more images obtained, such as with one of the images as a reference image and the other as a floating image, and map the floating image to the reference image to achieve alignment of the reference image with the floating image under the anatomical structure. Optionally, the reference image and the floating image may be images of the same individual, or images of different individuals, or images including the same anatomical structure, or images including partially the same anatomical structure, and the source of the reference image and the floating image is not limited in this embodiment. Optionally, the reference image and the floating image may be two-dimensional images or three-dimensional images, which is not specifically limited in this embodiment.

S204, extracting semantic information from the reference image and the floating image to obtain a marked reference image and a marked floating image which comprise the semantic information.

Specifically, after the computer device obtains the input reference image and the floating image, semantic information in the reference image and the floating image can be extracted according to a preset trained neural network model, for example, if a region corresponding to a lung is detected, the computer device can segment the region corresponding to the lung, so as to extract the semantic information corresponding to the lung: if the bone is detected, marking the position corresponding to the bone by using a marking point, so as to mention semantic information corresponding to the bone: anatomical landmarks. After the computer equipment utilizes the preset neural network model to extract the language information of the reference image and the floating image, a marked reference image and a marked floating image containing the extracted semantic information can be obtained.

S206, according to the semantic information, determining target image registration models respectively corresponding to the marked reference image and the marked floating image from preset image registration models.

Specifically, the image registration model is a model for registering the marked reference image and the marked floating image obtained after extracting the semantic information, such as a corresponding algorithm model of a surface matching algorithm, a mutual information method, a standard orthogonalization matrix method, a least square method, and the like. For the marked reference image and the marked floating image containing different semantic information, the computer device can register the marked reference image and the marked floating image with different registration models, namely the marked reference image and the marked floating image comprising the segmentation area and the marked reference image and the marked floating image comprising the anatomical marked points can correspond to different image registration models.

Optionally, the semantic information includes: at least one of a segmented region and anatomical landmark points of the floating image, and at least one of a segmented region and anatomical landmark points of the reference image. The semantic information may be anatomical marker points in the reference image and the floating image, or may be segmentation areas in the reference image and the floating image. Furthermore, the anatomical mark points can be geometric mark points, such as a gray extreme value or a linear structure intersection point, or can be clearly visible and accurately positionable anatomical mark points in anatomical form, such as key mark points or characteristic points of human tissues, organs or lesions; the segmented region may be a curve or a curved surface corresponding to the reference image and the floating image, such as a lung, a liver, or an irregular region.

Optionally, the preset image registration model may include a segmentation-based image registration model and an anatomical marker-based registration model. Further, the segmented image registration model is an image registration model capable of performing image registration on the marked reference image and the marked floating image including the segmented region, such as an algorithm model corresponding to a surface matching algorithm, a mutual information method, a gray mean square error method and the like; the anatomical marker point-based registration model is a registration model capable of performing image registration on a marked reference image and a marked floating image which comprise the anatomical marker points, such as an algorithm model corresponding to methods such as a singular value decomposition algorithm, an iterative closest point method and a standard orthogonalization matrix method.

And S208, carrying out image registration on the reference image and the floating image according to the semantic information and the target image registration model.

Specifically, according to different semantic information, the computer device may select a corresponding target image registration model to perform image registration on the reference image and the floating image. Optionally, one reference image or one floating image may include the segmented region and the anatomical point at the same time, at this time, the computer device may register the anatomical points in the reference image and the floating image by using the target image registration model corresponding to the anatomical point, and then register the segmented regions in the reference image and the floating image by using the target image registration model corresponding to the segmented region; the segmentation regions in the reference image and the floating image may be registered by using the target image registration model corresponding to the segmentation regions, and then the anatomical points in the reference image and the floating image may be registered by using the target image registration model corresponding to the anatomical points, or the anatomical points in the reference image and the floating image may be registered by using the target image registration model corresponding to the anatomical points, and the segmentation regions in the reference image and the floating image may be registered by using the target image registration model corresponding to the segmentation regions, which is not limited in this embodiment.

Optionally, the computer Device may further introduce a Graphics Processing Unit (GPU) Processing partial operation supporting a parallel computing Architecture (CUDA) to further speed up the above-mentioned registration algorithm for registering the reference image and the floating image, while ensuring that the CPU therein is continuously used for the correlation operation Processing of image registration.

In the image registration method provided by the embodiment, the computer device can acquire the reference image and the floating image to be registered; extracting semantic information from the reference image and the floating image to obtain a marked reference image and a marked floating image which comprise the semantic information; determining target image registration models respectively corresponding to the marked reference image and the marked floating image from preset image registration models according to semantic information; and finally, carrying out image registration on the marked reference image and the marked floating image according to the semantic information and the target image registration model. In the embodiment, the computer device can extract the semantic information of the reference image and the floating image firstly, so that the reference image and the floating image are registered by adopting different target image registration models according to different semantic information to complete the registration of the reference image and the floating image comprising various semantic information, the limitation that the reference image and the floating image can only be registered based on single semantic information in the prior art is solved, and the application range of image registration is greatly improved.

Fig. 3 is a flowchart illustrating an image registration method according to another embodiment. The embodiment relates to a process that when the target image registration model is the registration model based on the anatomical marker points, the computer device registers the reference image and the floating image according to the registration model based on the anatomical marker points and the semantic information. On the basis of the foregoing embodiment, optionally, the foregoing S208 may include:

s302, acquiring a reference anatomy mark point set to be registered of the marked reference image and a floating anatomy mark point set to be registered of the marked floating image.

Specifically, the reference anatomical marker point set to be registered and the floating anatomical marker point set to be registered are sets of coordinate information of each anatomical marker point. Alternatively, the anatomical landmark points may be manually pre-marked.

S304, carrying out image registration on the reference image and the floating image according to the reference anatomical mark point set to be registered, the floating anatomical mark point set to be registered and the registration model based on the anatomical mark points.

Specifically, the anatomical landmark point-based registration model may be any one of algorithm models corresponding to methods such as a singular value decomposition algorithm, an iterative closest point algorithm, and a standard orthogonalization matrix method. The computer equipment can perform image registration on the reference image and the floating image according to the acquired reference anatomical marker point set to be registered, the acquired floating anatomical marker point set to be registered and a preset registration model based on the anatomical marker points.

Optionally, the S304 specifically includes: determining marker point intersection according to the matching result of the names of the marker points in the reference anatomical marker point set to be registered and the marker points in the floating anatomical marker point set to be registered; according to the mark point intersection, respectively determining an initial reference anatomical mark point set and an initial floating anatomical mark point set from the reference anatomical mark point set to be registered and the floating anatomical mark point set to be registered; and carrying out image registration on the reference image and the floating image according to the initial reference anatomical mark point set, the initial floating anatomical mark point set and the registration model based on the anatomical mark points.

Each anatomical marker point has a unique name, and the anatomical marker points with the same name in the reference anatomical marker point set to be registered and the floating anatomical marker point set to be registered form the marker point intersection of the reference anatomical marker point set and the floating anatomical marker point set to be registered. Optionally, the computer device may also use, as the intersection of the two anatomical landmark points, the anatomical landmark points with the same number as the anatomical landmark points in the reference anatomical landmark point set to be registered and the floating anatomical landmark point set to be registered. After determining the intersection of the marker points, the computer device may use a point set corresponding to the intersection of the marker points in the to-be-registered reference anatomical marker point set as an initial reference anatomical marker point set, and select a point set corresponding to the intersection of the marker points in the to-be-registered floating anatomical marker point set as an initial floating anatomical marker point set, so that the initial reference anatomical marker point set and the initial floating anatomical marker point set may be input into a preset registration model based on the anatomical marker points, thereby achieving alignment of the reference image and the floating image in the same anatomical structure.

In the step S304, the computer device may perform image registration on the reference image and the floating image according to an initial reference anatomical landmark set and an initial floating anatomical landmark set selected from the reference anatomical landmark set and the floating anatomical landmark set to be registered, and by using a registration model based on the anatomical landmark points. Optionally, the process of performing image registration on the reference image and the floating image by using the registration model based on the anatomical marker points to perform image registration on the reference image and the floating image may be divided into three stages of registration processes, each stage may obtain a corresponding registration result, and the three stages of registration processes are as follows:

the registration process of the first stage may be seen in S3042 to S3046:

s3042, determining a first registration result according to the initial reference anatomical marker point set and the initial floating anatomical marker point set and the anatomical marker point-based registration model; the first registration result includes a first registration result point set and a first transformation matrix.

Specifically, after the computer device inputs the initial reference anatomical landmark point set and the initial floating anatomical landmark point set into a preset registration model based on the anatomical landmark points, a first registration result point set and a first transformation matrix after spatial transformation is performed on the floating anatomical landmark point set to be registered can be obtained. The first registration result point set and the first transformation matrix constitute a first registration result.

S3044, determining a first floating anatomical marker point set corresponding to the first spatial distance in a preset ratio according to the first spatial distance set and the preset ratio; wherein, the first spatial distance set records first spatial distances between the reference anatomical marker point set to be registered and each corresponding marker point in the first registration result point set.

In particular, after obtaining the first set of registration result points, the computer device may obtain the first set of registration result points according to the formula D1 | | | P_f1–P_re1||₂Calculating a first spatial distance D1 between the reference anatomical marker point set to be registered and each corresponding marker point in the first registration result point set, wherein P_f1For a set of points, P, of the set of reference anatomical landmark points to be registered, constituted by the landmark points corresponding to the set of first registration results_re1Is a first set of registration result points. Alternatively, the above-mentioned preset ratio may be (0, 1) set as required]Any value within. Optionally, a first floating anatomical marker point set corresponding to a first spatial distance in a preset ratio may be directly selected, or each distance in the first spatial distance may be sorted in an ascending order, and then the first floating anatomical marker point set corresponding to the first spatial distance in the preset ratio is selected, because the smaller the first spatial distance between the reference anatomical marker point set to be registered and each corresponding marker point in the first registration result point set is, the higher the accuracy of the registration result is, and therefore, after sorting in an ascending order, each distance in the first spatial distance is selected, then the first floating anatomical marker point set corresponding to the first spatial distance in the preset ratio is selectedThe point set is marked by science, and the registration accuracy can be improved. The first floating anatomical landmark point set is a point set corresponding to a first spatial distance in a preset ratio selected from the floating anatomical landmark point set to be registered.

S3046, when the number of marker points in the first set of floating anatomical marker points is less than the preset number threshold, taking the first transformation matrix as a target transformation matrix.

Specifically, the target transformation matrix is a matrix for image registration of the marked reference image and the marked floating image, and the computer device can use the target transformation matrix to realize registration of the marked reference image and the marked floating image. Optionally, the computer device may compare the number of the marker points in the first floating anatomical marker point set with a preset number threshold, and determine whether to use the first transformation matrix as the target transformation matrix according to the comparison result. Optionally, the preset number threshold may be 5. When the number of the marker points in the first floating anatomical marker point set is smaller than the preset number threshold, the first transformation matrix is used as the target transformation matrix, and the process continues to step S30422.

When the number of the marker points in the first floating anatomical marker point set is not less than the preset number threshold, a second stage of registration process is required.

The registration process of the second stage may be seen in S3048 to S30416:

s3048, obtaining a first reference anatomical landmark point set corresponding to the first floating anatomical landmark point set in the reference anatomical landmark point set to be registered.

In this step, the first reference anatomical marker set is a point set formed by marker points corresponding to marker points of which the names or numbers of the marker points in the reference anatomical marker set to be registered are the same as those of the markers in the first floating anatomical marker set.

S30410, determining a second transformation matrix according to the first set of reference anatomical marker points, the first set of floating anatomical marker points, and the anatomical marker points-based registration model.

Specifically, as with the method for determining the first transformation matrix described above, the computer device may input the first set of reference anatomical landmark points and the first set of floating anatomical landmark points into a preset anatomical landmark point-based registration model, thereby obtaining a second transformation matrix.

S30412, determining a second registration result point set according to the second transformation matrix and the floating anatomical mark point set to be registered.

In this step, the computer device may perform spatial transformation on the floating anatomical landmark set to be registered by using the second transformation matrix according to the obtained product of the second transformation matrix and the floating anatomical landmark set to be registered, and obtain a second registration result point set by combining interpolation methods such as neighbor interpolation, bilinear interpolation, trilinear interpolation, and the like.

S30414, determining a second floating anatomical landmark set corresponding to a second spatial distance smaller than a preset distance threshold according to the second spatial distance set and the preset distance threshold; and recording second spatial distances between the reference anatomical mark point set to be registered and each corresponding mark point in the second registration result point set in the second spatial distance set.

In this step, after obtaining the second registration result point set, the computer device may obtain the second registration result point set according to the formula D2 | | | P_f–P_re2||₂Calculating a second spatial distance D2 between the reference anatomical marker point set to be registered and each corresponding marker point in the second registration result point set, wherein P_f2A point set P corresponding to each mark point in the reference anatomical mark point set to be registered and the second registration result point set_re2Is a second set of registration result points. Optionally, the preset distance threshold may be set as needed, for example, the distance threshold may be determined according to an actual distance between the reference anatomical marker point set to be registered and each corresponding marker point in the second registration result point set, which is acceptable to the user. The second floating anatomical landmark point set is a point set corresponding to a second spatial distance within a preset distance threshold value selected from the floating anatomical landmark point set to be registered.

S30416, when the number of marker points in the second set of floating anatomical marker points is less than the preset threshold number, taking the second transformation matrix as the target transformation matrix.

In this step, the computer device may compare the number of marker points in the second floating anatomical marker point set with a preset number threshold, and determine whether to use the second transformation matrix as the target transformation matrix according to the comparison result. When the number of the marker points in the second floating anatomical marker point set is smaller than the preset number threshold, the second transformation matrix is used as the target transformation matrix, and the process continues to step S30422.

When the number of marker points in the second set of floating anatomical marker points is not less than the preset number threshold, a third stage of registration process is required.

The registration process of the third stage may be seen in S30418 to S30420:

s30418, obtaining a second reference anatomical landmark point set corresponding to the second floating anatomical landmark point set in the reference anatomical landmark point set to be registered.

In this step, the second reference anatomical marker point set is a point set corresponding to a marker point selected from the reference anatomical marker point set to be registered, the marker point having the same name or number as the marker point in the second floating anatomical marker point set.

S30420, determining a third transformation matrix according to the second reference anatomical landmark point set, the second floating anatomical landmark point set, and the anatomical landmark point-based registration model, and using the third transformation matrix as the target transformation matrix.

In this step, as in the method for determining the first transformation matrix and the second transformation matrix, the computer device may input the second reference anatomical landmark set and the second floating anatomical landmark set into a preset registration model based on the anatomical landmark points, so as to obtain a third transformation matrix, and after obtaining the third transformation matrix, the computer device may directly use the third transformation matrix as a target transformation matrix.

S30422, performing image registration on the reference image and the floating image according to the target transformation matrix.

Specifically, the computer device may map the marked floating image to the marked reference image space according to a product of a matrix formed by the coordinate position of each pixel point of the floating image and the target transformation matrix, and by combining an interpolation method such as neighbor interpolation, bilinear interpolation, trilinear interpolation, or the like, to achieve alignment of the marked reference image and the marked floating image under the anatomical structure, thereby completing image registration of the marked reference image and the marked floating image.

Optionally, the preset ratio and the preset distance threshold may be adjusted as follows: adding noise to each mark point in a reference image and a floating image to be registered, registering the reference image and the floating image to be registered by using the three-stage registration mode to obtain a new target transformation matrix, registering the reference image and the floating image by using the new target transformation matrix, calculating the similarity metric value between the registered reference image and the floating image by using a preset similarity metric model according to the obtained registration result, comparing the similarity metric value with a preset similarity metric threshold value, if the similarity metric value is smaller than the preset similarity metric threshold value, adjusting at least one of the preset ratio and the preset distance threshold value until the finally obtained similarity is larger than the preset similarity metric threshold value, and adjusting the preset ratio and the preset distance threshold value to be proper values, and then the registration accuracy of the image registered by the algorithm model of the adjusted preset ratio and the preset threshold value can be higher. It should be noted that the mean, variance, and number of the above-mentioned added noises may be randomly set.

In the image registration method provided by this embodiment, a computer device may obtain a to-be-registered reference anatomical marker point set for marking a reference image and a to-be-registered floating anatomical marker point set for marking a floating image; according to the reference anatomy mark point set to be registered, the floating anatomy mark point set to be registered and the registration model based on the anatomy mark points, the image registration is carried out on the mark reference image and the mark floating image in three stages, each stage carries out the image registration by using mark points in a certain condition such as a preset ratio or mark points in a preset distance threshold value instead of carrying out the image registration by using all the mark points, the calculation amount is greatly reduced, and the registration speed is improved; in addition, the marker sets in each stage are different, so that the influence of false detection on part of anatomical markers, which may affect the registration accuracy, can be reduced, and the markers in each stage are the markers which are determined by screening according to a preset ratio or a preset distance threshold and can improve the registration accuracy.

When the target image registration model is a segmentation-based image registration model, the computer device may perform image registration on the marker reference image and the marker floating image by using an image registration method provided by another embodiment shown in fig. 4. The embodiment relates to a realization process of image registration of the mark reference image and the mark floating image by computer equipment according to the extracted segmentation region and the corresponding segmentation-based image registration model. On the basis of the foregoing embodiment, optionally, another optional implementation manner of the foregoing S208 may include:

s402, acquiring a segmentation reference image corresponding to the marking reference image and a segmentation floating image corresponding to the floating image.

Specifically, the reference image and the floating image may be corresponding images obtained by extracting semantic information from the reference image and the floating image to be registered according to the preset trained neural network model. Optionally, the computer device may perform any region segmentation on the reference image to be registered and the floating image to obtain a segmented reference image and a segmented floating image by using the preset trained neural network model.

S404, carrying out image registration on the reference image and the floating image according to the segmentation reference image, the segmentation floating image and the segmentation-based image registration model.

Specifically, the image registration model based on the segmentation may be any one of algorithm models corresponding to registration methods such as a surface matching algorithm, a mutual information method, a gray mean square error method, and the like. The computer equipment can determine a target segmentation transformation matrix according to the obtained segmentation reference image, the segmentation floating image and the image registration model based on the segmentation, so that the floating image to be registered is mapped to the space coordinate of the reference image according to the target segmentation transformation matrix to complete the registration of the reference image and the floating image.

In the image registration method provided by this embodiment, the computer device may obtain the segmented reference image corresponding to the marked reference image and the segmented floating image corresponding to the floating image; and performing image registration on the reference image and the floating image according to the segmented reference image, the segmented floating image and the segmented image registration model. In this embodiment, the computer device may perform image registration on the reference image and the floating image directly by using a preset image registration model based on the segmentation according to the segmented reference image and the segmented floating image obtained after semantic information extraction, and the implementation manner is relatively simple.

Fig. 5 is an image registration method according to yet another embodiment. The embodiment relates to a process that computer equipment performs image integration on a registration result obtained by registering a reference image and a floating image according to the embodiment by using a preset image integration model. On the basis of the foregoing embodiment, optionally, the foregoing method may further include:

and S502, acquiring a registration result after image registration is carried out on the reference image and the floating image.

In this step, the registration result is a registered reference image and floating image obtained by performing image registration on the reference image and floating image.

S504, according to the registration result and a preset image integration model, image integration is carried out on the registration result.

In this step, the preset image integration model may be any one of methods such as a trilinear interpolation method and a B-spline interpolation method. The image integration may be performed by combining two or more registered images from different imaging devices or acquired at different times, and using an algorithm to organically combine the images. The computer device can integrate the reference image and the floating image in the registration result by using a preset image integration model to obtain a distorted image in which the floating image and the reference image are integrated together under the reference image space.

In the image registration method provided by the embodiment, the computer device can obtain the registration result after the image registration is performed on the reference image and the floating image; and integrating the images of the registration result according to the registration result and a preset image integration model to integrate the reference image and the floating image into one image, thereby complementarily and organically combining the advantages of the images to obtain a new image with richer information content, and better assisting a doctor to judge the condition of the patient by using the integrated image.

Fig. 6 is a flowchart illustrating an image registration method according to yet another embodiment. The embodiment relates to an implementation process in which a computer device adjusts a similarity metric value by using a gradient descent method according to a target matrix obtained by the above embodiment and an image obtained by down-sampling a reference image and a floating image to determine a target parameter. On the basis of the foregoing embodiment, optionally, the foregoing method may further include:

s602, obtaining the target transformation matrix.

And S604, determining a similarity metric value between the down-sampling reference image and the converted floating image corresponding to the down-sampling floating image according to the target transformation matrix, the down-sampling reference image obtained by performing down-sampling operation on the reference image and the down-sampling floating image obtained by performing down-sampling operation on the floating image.

Specifically, the computer device may perform downsampling on the reference image and the floating image to obtain a downsampled reference image and a downsampled floating image, optionally perform downsampling on the reference image and the floating image once to obtain a downsampled reference image and a downsampled floating image, perform spatial transformation on the downsampled floating image by using the target transformation matrix to obtain a transformed floating image, and determine a similarity metric between the transformed floating image and the downsampled reference image by using a preset calculation model of the similarity metric, such as an algorithm model corresponding to a mutual information method, a gray scale mean square error method, and the like.

S606, at least one operation of translation operation, rotation operation, tilting operation and scaling operation is carried out on the target transformation matrix, and initial parameters corresponding to the target transformation matrix are extracted.

Specifically, if the reference image and the floating image are three-dimensional images, the corresponding target transformation matrix may be a 4 × 4 matrix, and the computer device may perform a translation operation, a rotation operation, a tilt operation, and a scaling operation on the target transformation matrix, decompose the target transformation matrix into four 4 × 4 matrices, such as a translation matrix, a rotation matrix, a tilt matrix, and a scaling matrix, and further obtain initial parameters corresponding to 12 target transformation matrices according to a translation distance, a rotation angle, a tilt angle, a scaling ratio, and the like of the four 4 × 4 matrices in a three-dimensional coordinate system. Similarly, if the reference image and the floating image are two-dimensional images, the computer device may obtain initial parameters corresponding to the 8 object transformation matrices.

S608, determining a target parameter according to the similarity metric value, the initial parameter and a preset gradient descent method.

Specifically, the computer device may adjust the initial parameter according to a preset gradient descent method, so that the similarity metric value is optimal, and the adjusted parameter corresponding to the optimal similarity metric value is used as the target parameter. Optionally, the computer device may determine a final transformation matrix corresponding to the target parameter according to the target parameter, and register the reference image and the floating image by using the final transformation matrix.

Optionally, the computer device may also perform a plurality of downsampling operations on the reference image and the floating image, for example, perform downsampling three times and obtain corresponding downsampled reference image and downsampled floating image respectively. Further, the downsampled reference image may include a first downsampled reference image corresponding to a first downsampling, a second downsampled reference image corresponding to a second downsampling, and a third downsampled reference image corresponding to a third downsampling, and similarly, the downsampled floating image may include a first downsampled floating image corresponding to the first downsampling, a second downsampled floating image corresponding to the second downsampling, and a third downsampled floating image corresponding to the third downsampling. At this time, the target parameter may be determined using the following method: the first step is as follows: the computer device can perform spatial transformation on the third downsampled floating image by using the target transformation matrix, so that the third downsampled floating image is mapped to a spatial coordinate system corresponding to the third downsampled reference image to obtain a transformed third floating image, and a first similarity metric value between the transformed third floating image and the third downsampled reference image is determined by using a preset similarity metric value calculation model; the second step is that: the computer device can adjust the initial parameters by using a preset gradient descent method to enable the first similarity metric value to be optimal, determine a new target transformation matrix according to the parameters corresponding to the optimal first similarity metric value, continuously perform the first step and the second step on the second downsampling floating image and the downsampling reference image by using the new target transformation matrix until the first step and the second step are performed on the initial reference image and the floating image, and use the finally obtained parameters corresponding to the optimal similarity metric value as target parameters, so that the computer device can determine a final transformation matrix corresponding to the target parameters according to the target parameters and register the reference image and the floating image by using the final transformation matrix.

Optionally, the computer device may first perform image integration on the registration result obtained after the image registration is performed on the reference image and the floating image by using the image integration method corresponding to the embodiment shown in fig. 5, then optimize the integration result obtained in the embodiment shown in fig. 5 by using the registration result obtained by performing registration on the reference image and the floating image by using the final transformation matrix provided in this embodiment, may also perform image optimization on the registration result obtained after the image registration is performed on the reference image and the floating image by using the image optimization method provided in this embodiment, and then perform image integration on the registration result obtained by performing registration on the reference image and the floating image by using the final transformation matrix provided in this embodiment by using the image integration method corresponding to the embodiment shown in fig. 5, which is not limited in this embodiment.

In the image registration method provided by this embodiment, the computer device may obtain the target transformation matrix, and determine a similarity metric between the downsampled reference image and the transformed floating image corresponding to the downsampled floating image according to the target transformation matrix, the downsampled reference image obtained by downsampling the reference image, and the downsampled floating image obtained by downsampling the floating image; performing at least one operation of translation operation, rotation operation, tilting operation and scaling operation on the target transformation matrix, and extracting initial parameters corresponding to the target transformation matrix; and then determining a target parameter according to the similarity metric value, the initial parameter and a preset gradient descent method, wherein the target parameter is a parameter corresponding to the optimal similarity metric value, so that a final transformation matrix determined according to the target parameter is also better, the precision of registering the floating image and the reference image is higher by using the final transformation matrix, and the precision of image registration is further improved.

The following describes the process of the image registration method according to the embodiment of the present application by way of a simple example. See in particular the following steps:

s702, the computer equipment acquires a reference image and a floating image to be registered.

S704, extracting semantic information from the reference image and the floating image by computer equipment to obtain a marked reference image and a marked floating image which comprise the semantic information; the semantic information includes: at least one of a segmentation region and anatomical landmark points of the floating image, and at least one of a segmentation region and anatomical landmark points of the reference image.

S706, determining target image registration models respectively corresponding to the marked reference image and the marked floating image from preset image registration models by computer equipment according to the semantic information; the preset image registration model comprises a segmentation-based image registration model and an anatomical marker point-based registration model.

And S708, judging whether the target image registration model is the registration model based on the anatomical mark points or not by the computer equipment, if so, continuing to execute S710, and if not, executing S740.

S710, the computer equipment acquires a to-be-registered reference anatomical marking point set of the marked reference image and a to-be-registered floating anatomical marking point set of the marked floating image.

And S712, the computer device determines the marker point intersections of the marker points in the reference anatomical marker point set to be registered and the floating anatomical marker point set to be registered, the marker point intersections of the marker points in the reference anatomical marker point set to be registered and the floating anatomical marker point set to be registered are the same in name, selects the marker point intersections in the reference anatomical marker point set to be registered as an initial reference anatomical marker point set, and selects the marker point intersections in the floating anatomical marker point set to be registered as an initial floating anatomical marker point set.

S714, the computer device determines the first registration result according to the initial reference anatomical marker point set, the initial floating anatomical marker point set and the anatomical marker point-based registration model.

S716, determining a first floating anatomical marker point set corresponding to a first spatial distance in a preset ratio by the computer equipment according to the first spatial distance set and the preset ratio; wherein, the first spatial distance set records first spatial distances between the reference anatomical marker point set to be registered and each corresponding marker point in the first registration result point set.

S718, the computer device determines whether the number of the marker points in the first floating anatomical marker point set is less than the preset number threshold, if so, continues to execute S720, and if not, executes S722.

S720, the computer equipment takes the first transformation matrix as a target transformation matrix.

And S722, the computer equipment acquires a first reference anatomical mark point set corresponding to the first floating anatomical mark point set in the reference anatomical mark point set to be registered.

S724, the computer device determines a second transformation matrix according to the first reference anatomical marker point set, the first floating anatomical marker point set and the anatomical marker point-based registration model.

And S726, the computer equipment determines a second registration result point set according to the second transformation matrix and the floating anatomical mark point set to be registered.

S728, the computer device determines a second floating anatomical landmark set corresponding to a second spatial distance smaller than a preset distance threshold according to the second spatial distance set and the preset distance threshold; and recording second spatial distances between the reference anatomical mark point set to be registered and each corresponding mark point in the second registration result point set in the second spatial distance set.

And S730, judging whether the number of the marker points in the second floating anatomical marker point set is less than the preset threshold number by the computer equipment, if so, continuing to execute S732, and if not, executing S734.

S732, the computer device takes the second transformation matrix as the target transformation matrix.

S734, the computer device obtains a second reference anatomical landmark point set corresponding to the second floating anatomical landmark point set in the reference anatomical landmark point set to be registered.

S736, the computer device determines a third transformation matrix according to the second reference anatomical marker point set, the second floating anatomical marker point set and the anatomical marker point-based registration model, and takes the third transformation matrix as the target transformation matrix.

S738, carrying out image registration on the reference image and the floating image by the computer equipment according to the target transformation matrix; after the execution of S738, the execution continues with S744.

And S740, the computer equipment acquires a segmentation reference image corresponding to the marking reference image and a segmentation floating image corresponding to the floating image.

And S742, the computer equipment carries out image registration on the reference image and the floating image according to the segmentation reference image, the segmentation floating image and the segmentation-based image registration model.

S744, the computer device obtains a registration result after image registration is carried out on the reference image and the floating image.

And S746, the computer equipment carries out image integration on the registration result according to the registration result and a preset image integration model.

S748, the computer device obtains the target transformation matrix.

And S750, determining a similarity metric value between the down-sampling reference image and the converted floating image corresponding to the down-sampling floating image by the computer equipment according to the target transformation matrix, the down-sampling reference image obtained by performing down-sampling operation on the reference image and the down-sampling floating image obtained by performing down-sampling operation on the floating image.

S752, the computer device performs at least one of a translation operation, a rotation operation, a tilt operation, and a zoom operation on the target transformation matrix, and extracts an initial parameter corresponding to the target transformation matrix.

And S754, determining a target parameter by the computer equipment according to the similarity metric value, the initial parameter and a preset gradient descent method.

The working principle and technical effect of the image registration method provided by this embodiment are as described in the above embodiments, and are not described herein again.

It should be understood that, although the steps in the flowcharts of fig. 2 to 6 are shown in sequence as indicated by the arrows, the steps are not necessarily performed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-6 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performing the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

Fig. 7 is a schematic structural diagram of an image registration apparatus according to an embodiment. As shown in fig. 7, the apparatus may include a first acquisition module 702, a first extraction module 704, a first determination module 706, and a registration module 708.

Specifically, the first obtaining module 702 is configured to obtain a reference image and a floating image to be registered;

a first extraction module 704, configured to extract semantic information from the reference image and the floating image to obtain a labeled reference image and a labeled floating image that include the semantic information;

a first determining module 706, configured to determine, according to the semantic information, target image registration models corresponding to the marker reference image and the marker floating image respectively from preset image registration models;

a registration module 708, configured to perform image registration on the reference image and the floating image according to the semantic information and the target image registration model.

Optionally, the semantic information includes: at least one of a segmentation region and anatomical landmark points of the floating image, and at least one of a segmentation region and anatomical landmark points of the reference image; the preset image registration model comprises a segmentation-based image registration model and an anatomical marker point-based registration model.

The image registration apparatus provided in this embodiment may implement the above method embodiments, and the implementation principle and the technical effect are similar, which are not described herein again.

In another embodiment, on the basis of the embodiment shown in fig. 7, when the target image registration model is the registration model based on the anatomical marker points, the registration module 708 may optionally include a first obtaining unit and a first registration unit.

Specifically, the first obtaining unit is configured to obtain a to-be-registered reference anatomical marker point set of the marker reference image and a to-be-registered floating anatomical marker point set of the marker floating image;

and the first registration unit is used for carrying out image registration on the reference image and the floating image according to the reference anatomical mark point set to be registered, the floating anatomical mark point set to be registered and the registration model based on the anatomical mark points.

In an image registration apparatus provided in a further embodiment, on the basis of the above embodiment, optionally, the first registration unit may include a first determination subunit, a second determination subunit, and a registration subunit.

Specifically, the first determining subunit is configured to determine a marker intersection according to a matching result of the names of the markers in the reference anatomical marker set to be registered and the floating anatomical marker set to be registered;

a second determining subunit, configured to determine, according to the intersection of the marker points, an initial reference anatomical marker point set and an initial floating anatomical marker point set from the reference anatomical marker point set to be registered and the floating anatomical marker point set to be registered, respectively;

and the registration subunit is used for carrying out image registration on the reference image and the floating image according to the initial reference anatomical marker point set, the initial floating anatomical marker point set and the anatomical marker point-based registration model.

In an image registration apparatus structure provided in yet another embodiment, on the basis of the above embodiment, optionally, the registration module 708 may further include a second obtaining unit and a second registration unit.

A second acquiring unit, configured to acquire a divided reference image corresponding to the marked reference image and a divided floating image corresponding to the floating image;

and the second registration unit is used for carrying out image registration on the reference image and the floating image according to the segmentation reference image, the segmentation floating image and the segmentation-based image registration model.

Fig. 8 is a schematic structural diagram of an image registration apparatus according to yet another embodiment. On the basis of the above embodiment, optionally, the apparatus may further include a second obtaining module 710 and an integrating module 712.

A second obtaining module 710, configured to obtain a registration result after performing image registration on the reference image and the floating image;

and an integrating module 712, configured to perform image integration on the registration result according to the registration result and a preset image integration model.

Fig. 9 is a schematic structural diagram of an image registration apparatus according to yet another embodiment. On the basis of the foregoing embodiment, optionally, the apparatus may further include a third obtaining module 714, a second determining module 716, a second extracting module 718, and a third determining module 720.

A third obtaining module 714, configured to obtain the target transformation matrix.

A second determining module 716, configured to determine a similarity metric between the downsampled reference image and the transformed floating image corresponding to the downsampled floating image according to the target transformation matrix, the downsampled reference image obtained by downsampling the reference image, and the downsampled floating image obtained by downsampling the floating image;

a second extracting module 718, configured to perform at least one of a translation operation, a rotation operation, a tilt operation, and a scaling operation on the target transformation matrix, and extract an initial parameter corresponding to the target transformation matrix;

and a third determining module 720, configured to determine a target parameter according to the similarity metric, the initial parameter, and a preset gradient descent method.

In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:

acquiring a reference image and a floating image to be registered;

The implementation principle and technical effect of the computer device provided by the above embodiment are similar to those of the above method embodiment, and are not described herein again.

In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:

acquiring a reference image and a floating image to be registered;

The implementation principle and technical effect of the computer-readable storage medium provided by the above embodiments are similar to those of the above method embodiments, and are not described herein again.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A method of registration of images, the method comprising:

acquiring a reference image and a floating image to be registered;

acquiring a reference anatomical mark point set to be registered corresponding to the reference image and a floating anatomical mark point set to be registered corresponding to the floating image;

and performing at least one stage of image registration on the reference image and the floating image according to the initial reference anatomical marker point set, the initial floating anatomical marker point set and a registration model based on anatomical marker points.

2. The method of claim 1, wherein said at least one stage of image registration of said reference image and said floating image from said initial set of reference anatomical marker points, said initial set of floating anatomical marker points, and an anatomical marker points-based registration model, comprises:

carrying out image registration according to the initial reference anatomical mark point set, the initial floating anatomical mark point set and the anatomical mark point-based registration model to obtain a first registration result; the first registration result comprises a first registration result point set and a first transformation matrix;

determining a first spatial distance set according to the reference anatomical mark point set to be registered and the first registration result point set; wherein, the first spatial distance set records first spatial distances between the reference anatomical marker point set to be registered and each corresponding marker point in the first registration result point set;

determining a first floating anatomical marking point set corresponding to a first spatial distance in a preset ratio according to the first spatial distance set and the preset ratio;

if the number of the mark points in the first floating anatomical mark point set is smaller than a preset number threshold, taking the first transformation matrix as a target transformation matrix;

and carrying out image registration on the reference image and the floating image according to the target transformation matrix.

3. The method of claim 2, wherein determining a first set of floating anatomical landmark points for a first spatial distance within a preset ratio based on the first set of spatial distances and the preset ratio comprises:

sorting each of the first spatial distances in the first set of spatial distances;

and selecting first spatial distances in the preset ratio according to the sequence from small to large, and determining a first floating anatomical marking point set corresponding to the selected first spatial distances.

4. The method of claim 2, wherein after said determining a first set of floating anatomical landmark points for which a first spatial distance within the preset ratio corresponds, the method further comprises:

if the number of the marking points in the first floating anatomical marking point set is not less than a preset number threshold, acquiring a first reference anatomical marking point set corresponding to the first floating anatomical marking point set in the reference anatomical marking point set to be registered;

determining a second transformation matrix from the first set of reference anatomical landmark points, the first set of floating anatomical landmark points, and the anatomical landmark point-based registration model;

determining a second registration result point set and a second spatial distance set according to the second transformation matrix and the floating anatomical mark point set to be registered; second spatial distances between the reference anatomical mark point set to be registered and each corresponding mark point in the second registration result point set are recorded in the second spatial distance set;

determining a second floating anatomical marking point set corresponding to a second spatial distance smaller than a preset distance threshold according to the second spatial distance set and the preset distance threshold;

and if the number of the marker points in the second floating anatomical marker point set is less than the preset number threshold, taking the second transformation matrix as the target transformation matrix.

5. The method of claim 4, wherein after said determining a second set of floating anatomical landmark points for a second spatial distance less than the preset distance threshold, the method further comprises:

if the number of the marking points in the second floating anatomical marking point set is not less than the preset number threshold, acquiring a second reference anatomical marking point set corresponding to the second floating anatomical marking point set in the reference anatomical marking point set to be registered;

and determining a third transformation matrix according to the second reference anatomical marker point set, the second floating anatomical marker point set and the anatomical marker point-based registration model, and taking the third transformation matrix as the target transformation matrix.

6. The method according to any one of claims 2-5, wherein after said image registering said reference image and said floating image according to said target transformation matrix, said method further comprises:

acquiring the target transformation matrix;

performing down-sampling operation on the reference image according to the target transformation matrix to obtain a down-sampling reference image, performing down-sampling operation on the floating image to obtain a down-sampling floating image, and determining a similarity metric value between the down-sampling reference image and the converted floating image corresponding to the down-sampling floating image;

7. The method of claim 6, wherein after said obtaining the target transformation matrix, the method further comprises:

carrying out down-sampling operation on the reference image for multiple times to obtain a plurality of down-sampling reference images, and carrying out down-sampling operation on the floating image for multiple times to obtain a plurality of down-sampling floating images;

carrying out spatial transformation on the ith down-sampling floating image by using the target transformation matrix to obtain an ith transformed floating image; the ith downsampled floating image is the last downsampled floating image;

determining an ith similarity metric between the transformed ith floating image and an ith downsampled reference image; the ith down-sampling reference picture is the last down-sampling reference picture;

adjusting the initial parameters by using the preset gradient descent method to enable the ith similarity metric value to be optimal, and determining a new target transformation matrix according to the parameters corresponding to the optimal ith similarity metric value;

and calculating the i-1 th similarity metric value between the i-1 th floating image after the i-1 th transformation and the i-1 th down-sampling reference image by using the new target transformation moment until the similarity metric value between the initial reference image and the floating image is calculated, and taking the parameter corresponding to the optimal similarity metric value as the target parameter.

8. An apparatus for registration of images, the apparatus comprising:

the image acquisition module is used for acquiring a reference image and a floating image to be registered;

a point set obtaining module, configured to obtain a reference anatomical marker point set to be registered corresponding to the reference image and a floating anatomical marker point set to be registered corresponding to the floating image;

the intersection determining module is used for determining the intersection of the mark points according to the matching result of the names of the mark points in the reference anatomical mark point set to be registered and the floating anatomical mark point set to be registered;

a marker point set determining module, configured to determine an initial reference anatomical marker point set and an initial floating anatomical marker point set from the reference anatomical marker point set to be registered and the floating anatomical marker point set to be registered, respectively, according to the marker point intersection;

and the registration module is used for carrying out image registration of at least one stage on the reference image and the floating image according to the initial reference anatomical marker point set, the initial floating anatomical marker point set and a registration model based on anatomical marker points.

9. A computer arrangement comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method according to any one of claims 1-7 when executing the computer program.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.