CN115690207A - An automatic positioning method and device based on head clinical images - Google Patents

Abstract

The present invention relates to the technical field of image processing, in particular to an automatic positioning method and device based on head clinical images, which performs coordinate transformation on the head anatomical structure of CVH according to the spatial direction mapping relationship and the spatial size mapping relationship, and obtains the CVH registration head Head anatomy structure; CVH registration head anatomy structure and the position identification and organ name of each organ region in the CVH registration head anatomy structure form a new CVH anatomy knowledge map. Multi-planar reconstruction of clinical images of the patient's head is performed to obtain the anatomical structure of the patient's head; there is a real-time mapping relationship between the anatomical structure of the patient's head and the anatomical structure of the CVH-registered head, so that the anatomical structure of the patient's head and the CVH-registered head There is a one-to-one correspondence between the position identification of each organ region in the anatomical structure and the name of the organ. Select the region of interest from the anatomical knowledge map, and automatically locate the corresponding position of the region of interest in the anatomical structure of the patient's head, reducing the difficulty of identifying the clinical image of the patient's head.

Description

An automatic positioning method and device based on head clinical images

technical field

The invention relates to the technical field of image processing, in particular to an automatic positioning method and device based on head clinical images.

Background technique

Clinical medical tomographic images such as CT and MRI play an important role in the diagnosis and treatment of diseases. However, due to the abstraction of imaging methods and the complexity of human anatomy, clinical medical images often require the professional knowledge and experience of senior doctors, especially radiologists, to read them correctly. In particular, head clinical tomographic images with complex structures are too abstract and difficult to accurately identify for junior medical students, interdisciplinary researchers, medical engineers and ordinary patients.

For interdisciplinary professionals, young physicians, and medical students, it is difficult for non-experienced medical professionals to effectively locate and identify some complex lesion areas and subtle anatomical structures in medical images. For patients and their families, in the process of disease diagnosis and treatment, the awareness of active participation of patients and their families continues to increase. Medical units also hand over image data with browsers to individuals through disk burning and other methods. However, the existing browsers still require professional doctors It is possible to help patients establish an intuitive understanding of the lesions with the help of guidance, and patients and their families cannot understand the various organ regions in the images by themselves.

Therefore, in the era of intelligent information that emphasizes universality, how to quickly locate the region of interest in clinical head clinical images and present them to non-senior imaging professionals to reduce the difficulty of identifying clinical head clinical images has unique value .

Contents of the invention

Aiming at the deficiencies in the prior art, the present invention proposes an automatic positioning method and device based on head clinical images, so as to reduce the recognition difficulty of patient head clinical images and improve universality.

In a first aspect, the present invention provides an automatic positioning method based on head clinical images.

In the first practicable manner, an automatic positioning method based on head clinical images includes:

Obtain clinical images of the patient's head;

Register the clinical image of the patient's head with the CVH to obtain the spatial direction mapping relationship and spatial size mapping relationship;

Carry out coordinate transformation on the head anatomical structure of CVH according to the spatial direction mapping relationship and spatial size mapping relationship to obtain the CVH registration head anatomical structure; The new CVH anatomical knowledge map is composed of location identification and organ names;

Perform multi-planar reconstruction of the clinical image of the patient's head to obtain the anatomical structure of the patient's head; there is a real-time mapping relationship between the anatomical structure of the patient's head and the head anatomical structure of the CVH registration;

Select the region of interest from the anatomical knowledge map, and automatically locate the corresponding position of the region of interest in the anatomical structure of the patient's head.

Combining with the first achievable way, in the second achievable way, the clinical image of the patient's head and the CVH are registered to obtain the spatial direction mapping relationship and the spatial size mapping relationship, including:

Perform the first registration of the patient's head clinical image and CVH, and obtain the spatial direction mapping relationship and the first registration result;

According to the first registration result, the clinical image of the patient's head and the CVH are registered for the second time to obtain the spatial dimension mapping relationship.

In combination with the second achievable way, in the third achievable way, the first registration is performed on the clinical image of the patient's head and the CVH, and the spatial direction mapping relationship and the first registration result are obtained, including:

Perform three-dimensional contour reconstruction on the clinical image of the patient's head to obtain the three-dimensional contour structure of the patient;

Acquiring the first feature point according to the patient's three-dimensional contour structure;

Obtain the second feature point of the CVH three-dimensional contour structure;

Register the first feature point and the second feature point to obtain the spatial direction mapping relationship and the first registration result; the first registration result is the initial registration structure of the patient's 3D contour, and the initial registration structure of the patient's 3D contour and the CVH 3D The spatial orientation of the contour structure is consistent.

In combination with the third achievable manner, in the fourth achievable manner, the first feature point is obtained according to the patient's three-dimensional contour structure, including:

The eye-nose triangle is selected from the patient's three-dimensional contour structure, and the first feature point of the eye-nose triangle is generated.

In combination with the third achievable way, in the fifth achievable way, a second registration is performed on the clinical image of the patient's head and the CVH according to the first registration result to obtain the spatial dimension mapping relationship, including:

Obtaining a first circular contour according to the patient's three-dimensional contour structure;

Obtain the second annular contour of the CVH three-dimensional contour structure;

In the initial registration structure of the patient's three-dimensional contour, registration is performed according to the first circular contour and the second circular contour to obtain a spatial dimension mapping relationship.

In combination with the third possible way, in the sixth possible way, coordinate transformation is performed on the head anatomical structure of the CVH according to the spatial direction mapping relationship and the spatial size mapping relationship to obtain the CVH registration head anatomical structure, including:

Obtain a homogeneous transformation matrix according to the spatial direction mapping relationship and the spatial dimension mapping relationship; the homogeneous transformation matrix is used to characterize the coordinate transformation relationship between the patient's three-dimensional contour structure and the head anatomical structure of the CVH;

According to the homogeneous transformation matrix, the coordinate transformation of the head anatomical structure of CVH is carried out, and the head anatomical structure of CVH registration is obtained.

Combined with the sixth possible way, in the seventh possible way, the homogeneous transformation matrix is obtained by the following formula:

其中，matrix1为空间方向映射关系的变换矩阵，matrix2_i为空间尺寸映射关系的变换矩阵，matirx为齐次变换矩阵，i和n为正整数。

Among them, matrix1 is the transformation matrix of the spatial direction mapping relationship, matrix2 _i is the transformation matrix of the spatial dimension mapping relationship, matirx is the homogeneous transformation matrix, and i and n are positive integers.

Combined with the first possible way, the eighth way includes:

After automatically locating the corresponding position of the region of interest in the anatomical structure of the patient's head, the corresponding position area is marked with color, and the plane to which the corresponding position area belongs is displayed.

In a first aspect, the present invention provides an automatic head clinical image positioning device.

In a ninth practicable manner, an automatic head clinical image positioning device includes:

an acquisition module configured to acquire clinical images of the patient's head;

The registration module is configured to register the clinical image of the patient's head and the CVH to obtain a spatial direction mapping relationship and a spatial dimension mapping relationship;

The coordinate transformation module is configured to perform coordinate transformation on the head anatomical structure of the CVH according to the spatial direction mapping relationship and the spatial dimension mapping relationship to obtain the CVH registration head anatomy structure; the CVH registration head anatomy structure and the CVH registration head anatomy structure The location identification and organ names of each organ region in the anatomical structure form a new CVH anatomical knowledge map;

The multi-planar reconstruction module is configured to perform multi-planar reconstruction on the clinical image of the patient's head to obtain the anatomical structure of the patient's head; there is a real-time mapping relationship between the patient's head anatomical structure and the CVH registration head anatomical structure;

The automatic positioning module is configured to select the region of interest from the anatomical knowledge map, and automatically locate the corresponding position of the region of interest in the anatomical structure of the patient's head.

As can be seen from the above technical solutions, the beneficial technical effects of the present invention are as follows:

Carry out coordinate transformation on the head anatomical structure of CVH according to the spatial direction mapping relationship and spatial size mapping relationship to obtain the CVH registration head anatomical structure; The new CVH anatomical knowledge map is composed of position identification and organ names, and then multi-planar reconstruction is performed on the clinical images of the patient's head to obtain the anatomical structure of the patient's head; there is a real-time mapping between the patient's head anatomical structure and the CVH registration head anatomical structure Therefore, there is a one-to-one correspondence between the anatomical structure of the patient's head and the position identification and organ name of each organ region in the CVH-registered anatomical structure of the head. Select the region of interest from the anatomical knowledge map, and automatically locate the corresponding position of the region of interest in the anatomical structure of the patient's head. In this way, professional doctors in non-senior fields can also identify the various organ regions of the clinical images of the patient's head, which reduces the difficulty of identifying clinical images of the patient's head, and is beneficial to interdisciplinary professionals, young doctors, medical students, patients and their families. Understand the lesions in the clinical imaging of the patient's head to improve universality.

Description of drawings

In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the accompanying drawings that need to be used in the description of the specific embodiments or the prior art. Throughout the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, elements or parts are not necessarily drawn in actual scale.

FIG. 1 is a schematic diagram of an automatic positioning method based on head clinical images provided by an embodiment of the present invention;

Fig. 2 is a schematic structural diagram of a method for automatic head clinical image positioning provided by an embodiment of the present invention.

Detailed ways

Embodiments of the technical solutions of the present invention will be described in detail below in conjunction with the accompanying drawings. The following examples are only used to illustrate the technical solutions of the present invention more clearly, and therefore are only examples, rather than limiting the protection scope of the present invention.

It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application shall have the usual meanings understood by those skilled in the art to which the present invention belongs.

As shown in Figure 1, the present embodiment provides an automatic positioning method based on head clinical images, including:

Step S01, obtaining clinical images of the patient's head;

Step S02, registering the clinical image of the patient's head with the CVH to obtain the spatial direction mapping relationship and the spatial size mapping relationship;

Step S03, perform coordinate transformation on the head anatomical structure of CVH according to the spatial direction mapping relationship and the spatial size mapping relationship to obtain the CVH registration head anatomy structure; each of the CVH registration head anatomy structure and the CVH registration head anatomy structure The new CVH anatomical knowledge map is composed of the location identification and organ names of organ regions;

Step S04: Perform multi-planar reconstruction of the clinical image of the patient's head to obtain the anatomical structure of the patient's head; the CVH-registered head anatomical structure and the position identification and organ name of each organ region in the CVH-registered head anatomical structure form a new CVH Anatomical knowledge map;

Step S05, selecting the region of interest from the anatomical knowledge map, and automatically locating the corresponding position of the region of interest in the anatomical structure of the patient's head.

Carry out coordinate transformation on the head anatomical structure of CVH according to the spatial direction mapping relationship and spatial size mapping relationship to obtain the CVH registration head anatomical structure; The new CVH anatomical knowledge map is composed of position identification and organ names, and then multi-planar reconstruction is performed on the clinical images of the patient's head to obtain the anatomical structure of the patient's head; there is a real-time mapping between the patient's head anatomical structure and the CVH registration head anatomical structure Therefore, there is a one-to-one correspondence between the anatomical structure of the patient's head and the position identification and organ name of each organ region in the CVH-registered anatomical structure of the head. Select the region of interest from the anatomical knowledge map, and automatically locate the corresponding position of the region of interest in the patient's head anatomy according to the real-time mapping relationship between the patient's head anatomy and the CVH registration head anatomy. In this way, professional doctors in non-senior fields can also identify the various organ regions of the clinical images of the patient's head, which reduces the difficulty of identifying clinical images of the patient's head, and is beneficial to interdisciplinary professionals, young doctors, medical students, patients and their families. Understand the lesions in the clinical imaging of the patient's head to improve universality.

Optionally, a clinical image of the head of the patient is obtained by CT (Computed Tomography, computerized tomography) or MRI (Nuclear Magnetic Resonance Imaging, nuclear magnetic resonance imaging).

In some embodiments, CVH (Chinese Visible Human) includes a cross-sectional anatomical data set of Chinese visualized human body and corresponding CT and MRI data sets, and the data set includes clinical images and anatomical structures of various parts of the body. CVH includes an anatomical knowledge map, which includes the anatomical structure of each part, as well as relevant knowledge such as the position identification and organ name of each organ area in the structure. The three-dimensional contour structure of the head in CVH was obtained by reconstructing the three-dimensional contour of the head clinical images of CVH.

Optionally, the CVH registration head anatomical structure replaces the CVH head anatomical structure in the old CVH anatomical knowledge graph to form a new CVH anatomical knowledge graph. The anatomical structure of the patient's head is mapped one by one with the anatomical structure of the CVH registration head, and the position identification and organ name of each organ region exist in the CVH registration head anatomy structure. Therefore, after corresponding the regions in the patient's head anatomical structure to the regions in the CVH registered head anatomical structure, the position identification and organ name of each organ region in the CVH registered head anatomical structure are also consistent with the patient's head anatomical structure Corresponding to each region in the head anatomical structure, according to the corresponding relationship between the position identification and organ name of each organ region in the head anatomical structure of the CVH registration and each region in the patient's head anatomical structure, automatic positioning is realized. For example, there is a mapping relationship between A and B. If A is selected, B corresponds to A, and the location identifier and organ name related to B also corresponds to A.

Optionally, register the clinical image of the patient's head with the CVH to obtain the spatial direction mapping relationship and the spatial dimension mapping relationship, including: performing the first registration on the patient's head clinical image and the CVH to obtain the spatial direction mapping relationship and The first registration result; according to the first registration result, a second registration is performed on the clinical image of the patient's head and the CVH to obtain the spatial dimension mapping relationship.

Optionally, the iterative closest point algorithm (Iterative Closest Point, ICP) is used to perform the first registration of the clinical image of the patient's head and the CVH, and the transformation matrix of the spatial direction mapping relationship and the first registration result are obtained; the ICP algorithm is used according to As a result of the first registration, a second registration is performed on the clinical image of the patient's head and the CVH to obtain the transformation matrix of the spatial size mapping relationship. In some embodiments, in ordinary heterogeneous registration, due to certain morphological differences in the heads of different human bodies, protruding parts such as noses and ears and similar gradient parts such as forehead and occiput are likely to cause errors in the registration process. "Best" matching, that is, the local extremum problem, so the deviation is still relatively large, and the registration result is very unsatisfactory. In this application, the result of the first registration is used as the basis for the second registration, and the second progressive registration is performed, which avoids the erroneous " The "best" matching situation effectively reduces the deviation and improves the registration degree between the clinical image of the patient's head and the CVH, thereby improving the accuracy of automatic positioning.

Optionally, the first registration is performed on the clinical image of the patient's head and the CVH, and the spatial direction mapping relationship and the first registration result are obtained, including: performing three-dimensional contour reconstruction on the clinical image of the patient's head to obtain the three-dimensional contour structure of the patient; Obtain the first feature point according to the patient's three-dimensional contour structure; obtain the second feature point of the CVH three-dimensional contour structure; register the first feature point and the second feature point to obtain the spatial direction mapping relationship and the first registration result; the first The registration result is the first registration of the three-dimensional contour structure of the CVH, and the three-dimensional contour structure of the patient is consistent with the spatial direction of the first registration of the three-dimensional contour structure of the CVH.

Optionally, the first feature point and the second feature point are registered through an iterative closest point algorithm (Iterative Closest Point, ICP). In some embodiments, the iterative closest point algorithm is a method based on the quaternion method, which implements registration of surfaces by iteratively calculating the sum of squared residuals of corresponding points between surfaces. For example, there are two sets of coordinate points P={Pi,i=0,1,2,...,k} (red block) and U={Ui,i=0,1,2, ...,n} (blue block). Set the point set P as the target point set, and U as the source point set. Assuming that P and U can roughly correspond in space (usually k≥n), then a new point set can be obtained by continuously rotating and translating the point set U U', so that the distance between the homologous points of the point set U' and P is the smallest (make U' and P overlap as much as possible). U' can be obtained by the rigid body geometric transformation formula U'=RU+T. Among them, R represents the three-dimensional rotation matrix of the transformed point set U, and T represents the translation vector of the transformed point set U. The core of this process is to use the least root mean square method to iteratively calculate the residual square sum of the corresponding points between the point sets U' and P by continuously correcting R and T, and find the root mean square minimum error between U' and Q , if the error is less than the preset limit value, the iteration ends, and the optimal solution for registration is obtained.

Optionally, acquiring the first feature points according to the patient's three-dimensional contour structure includes: selecting eye-nose triangles respectively from the patient's three-dimensional contour structure, and generating the first feature points of the eye-nose triangles. In some embodiments, the eye-nose triangle area is obvious and has large differences, and is used as facial feature points for registration, which avoids the local extremum problem of ordinary heterogeneous registration and makes registration more accurate.

Optionally, a second registration is performed on the clinical image of the patient's head and the CVH according to the first registration result to obtain a spatial dimension mapping relationship, including: obtaining the first circular contour according to the patient's three-dimensional contour structure; obtaining the CVH three-dimensional contour structure The second circular contour of the patient; on the basis of the three-dimensional contour structure of the patient and the three-dimensional contour structure of the CVH initial registration, the registration is performed according to the first circular contour and the second circular contour to obtain a spatial dimension mapping relationship.

In some embodiments, obtaining the annular contour includes utilizing the FO-DICOM library of C++ open source to read the original data, and converting it into an 8-bit bitmap format, and then utilizing the Ostu adaptive threshold segmentation method to perform binarization; utilizing the open operation Fill the small intracranial holes, and use the FloodFill algorithm to fill the entire intracranial area; finally, use the 3*3 corrosion template to perform forward and backward subtraction operations to extract the outline of one frame of data, and continue to process the next frame of data until the batch processing is completed. Finally, the extraction of circular contours is realized.

In some embodiments, the spatial direction mapping relationship is determined for the first registration. After the first registration, the spatial direction transformation between the patient's three-dimensional contour structure and the CVH three-dimensional contour structure is consistent, and the CVH three-dimensional contour structure of the patient's three-dimensional contour structure After the spatial directions between the contour structures are consistent, the circular feature contours with the same data are selected manually from the patient's three-dimensional contour structure and the CVH three-dimensional contour structure, and the spatial size registration is further carried out according to the circular feature contours to adapt to different The difference in size of the human head makes the CVH three-dimensional contour structure close to the patient's three-dimensional contour structure after two registrations, which is conducive to real-time mapping between the two. At the same time, the risk of local extremum is reduced by manually selecting circular feature contours with the same data from the three-dimensional contour structure of the patient and the three-dimensional contour structure of the CVH.

Optionally, coordinate transformation is performed on the head anatomical structure of the CVH according to the spatial direction mapping relationship and the spatial size mapping relationship to obtain the CVH registration head anatomical structure, including: obtaining a homogeneous transformation according to the spatial direction mapping relationship and the spatial size mapping relationship Matrix; the homogeneous transformation matrix is used to characterize the coordinate transformation relationship between the patient's three-dimensional contour structure and the head anatomy of CVH; according to the homogeneous transformation matrix, the coordinate transformation of the head anatomy of CVH is carried out to obtain the CVH registration head anatomy structure.

Optionally, the CVH head anatomical structure is transformed through the spatial direction mapping relationship and the spatial size mapping relationship, so that the head anatomical structure in the CVH is closer to the three-dimensional contour structure of the patient's head, consistent with its spatial direction and spatial size. There is a one-to-one mapping relationship between the transformed CVH registration head anatomical structure and the data of the patient's head three-dimensional contour structure. Then, the three-dimensional outline structure of the patient's head is reconstructed in multiple planes, and the same scene is superimposed to obtain the final multi-plane anatomical structure of the patient's head. There is a one-to-one correspondence between the anatomical structure of the patient's head and the anatomical structure of the head registered by CVH; and there is a one-to-one correspondence between the anatomical structure of the head of CVH and the anatomical knowledge map. There is a mapping relationship in knowledge graphs.

In some embodiments, the clinical image of the patient's head (such as patient image data such as CT/MRI) is used as the target image, and the CVH data set is used as a floating image to register and move to the target image to obtain a homogeneous transformation matrix; CVH general anatomical structure segmentation map The layer changes synchronously with the movement of the floating image, and is registered with the clinical image of the patient's head. After 3D reconstruction, real-time mapping at any viewing angle is obtained to achieve the purpose of lesion/organ navigation. The registration results allow rapid evaluation and correction.

Optionally, the homogeneous transformation matrix is obtained by the following formula:

其中，matrix1为空间方向映射关系的变换矩阵，matrix2_i为空间尺寸映射关系的变换矩阵，matirx为齐次变换矩阵，i和n为正整数，n代表递进配准的次数。

Among them, matrix1 is the transformation matrix of the spatial direction mapping relationship, matrix2 _i is the transformation matrix of the spatial dimension mapping relationship, matirx is the homogeneous transformation matrix, i and n are positive integers, and n represents the number of progressive registrations.

Optionally, during the multi-plane reconstruction of the clinical image of the patient's head, the multi-plane reconstruction is performed with reference to the multi-plane reconstruction method in the prior art to obtain the patient's anatomical structure in multiple planes.

Optionally, the automatic positioning method based on head clinical images includes: after automatically positioning the corresponding position of the region of interest in the anatomical structure of the patient's head, marking the corresponding position area with a color, and displaying the plane to which the corresponding position area belongs.

Optionally, this embodiment provides an automatic positioning method based on head clinical images, including: acquiring the patient's head clinical images; performing a second decrement on the patient's head clinical images and CVH (Chinese Visible Human, Chinese Visual Human) Registration, to obtain CVH registration head anatomical structure; multi-planar reconstruction of clinical images of the patient's head to obtain the patient's head anatomical structure; real-time mapping between the patient's head anatomical structure and the anatomical knowledge map in CVH; from anatomical knowledge Select the region of interest in the atlas, and automatically locate the corresponding position of the region of interest in the anatomical structure of the patient's head.

Results As shown in FIG. 2 , the present invention provides an automatic head clinical image positioning device, including: an acquisition module 101 , a registration module 102 , a coordinate transformation module 103 , a multi-plane reconstruction module 104 , and an automatic positioning module 105 . The acquisition module 101 is configured to acquire the clinical image of the patient's head; the registration module 102 is configured to register the clinical image of the patient's head with the CVH, and obtain the spatial direction mapping relationship and the spatial dimension mapping relationship; the coordinate transformation module 103 is configured to Carry out coordinate transformation on the head anatomical structure of CVH according to the spatial direction mapping relationship and spatial size mapping relationship to obtain the CVH registration head anatomical structure; The new CVH anatomical knowledge map is composed of the position identification and the organ name; the multi-plane reconstruction module 104 is configured to perform multi-plane reconstruction on the clinical image of the patient's head to obtain the anatomical structure of the patient's head; the anatomical structure of the patient's head is registered with the CVH head There is a real-time mapping relationship between the anatomical structures; the automatic positioning module 105 is configured to select the region of interest from the anatomical knowledge map, and automatically locate the corresponding position of the region of interest in the anatomical structure of the patient's head.

In some embodiments, the method for automatic head clinical image positioning includes: step S21, the user selects clinical image data, and displays the three-dimensional contour structure model of the CVH standard atlas and the automatic three-dimensional contour structure model of the clinical image data selected by the user. Step S22, respectively adjust the two contour models to the angle of view that is convenient for selecting the "eye-nose triangle area" (no need to set the same size and orientation), select the "eye-nose triangle area" by drawing a rectangle, and separate Blue and red feature points are generated on the CVH (left) and CT/MRI models. S23. Perform the first registration, obtain the matrix equation of the spatial direction mapping relationship, and display the rough registration result; wherein the blue CVH feature points will move closer to the red CT/MRI feature points. S24. Automatically extract the circular contour and perform the second registration, and obtain the final homogeneous matrix equation to guide the CVH anatomical knowledge data to move closer to the CT/MRI data, visually present the registration result by superimposing the same scene, and Shows the overlay effect of the two sets of images.

Carry out multiplanar reconstruction (MPR) of CT/MRI, display the results of multiplanar reconstruction, and list the gross organs of CVH for screening and selection. For example, select "Brainstem", that is, the brainstem area will be marked in yellow in the MPR reconstruction area of CT/MRI, which is convenient for users to read.

In some embodiments, the head region of the clinical image set (CT and MRI) corresponds to the rigid structure of the general anatomical structure of the CVH, and the three-dimensional reconstruction and feature point picking of the rigid features of the external contour are completed. Then, the second progressive registration method is adopted, that is, the rigid transformation of the facial features of the "eye-nose triangle" is used to determine the spatial orientation for the first time, and the rough registration is completed; further, the circular contour feature is used to perform similarity registration to determine the spatial size, Solve the local extremum and deformation problems that are easy to occur in heterogeneous registration, and summarize the standardized operation process including the secondary registration method. Finally, the head labor images are resampled and spatially transformed to obtain the anatomical structure of the patient's head. According to the registration results, the coordinate field transformation is carried out on the volume data field formed by the general anatomical structure layer set of CVH to realize clinical The purpose of anatomical structure mapping of image data is to design and optimize the interactive UI interface of MPR multi-planar reconstruction to achieve a productized application form.

In some embodiments, the existing research on the registration of anatomical structure knowledge maps and medical images has certain requirements for the integrity and quality of the image structure, which is different from the clinical image scanning norms. In order to prevent excessive medical treatment and control medical costs, real clinical data are often partial scans rather than thin-slice high-field scans required by scientific research. It is difficult to meet the "complete structure" and quality requirements required by current mainstream registration algorithms, and solutions often require The cost of huge computing consumption. However, the automatic positioning method based on clinical images of the head provided by this application has a small amount of calculation, low requirements for image scanning, and has the characteristics of light weight.

In some embodiments, the automatic positioning method based on head clinical images provided by this application can map the anatomical knowledge map in CVH to private CT/MRI data in real time without the need for professional explanations and special 3D organ printing annotations It can visually present the current organ area and locate the area of interest in the way of intelligent navigation, realize the purpose of "explaining" with unmanned intelligent assistance, and has the characteristics of human-computer interaction, serving the active participation needs of patients and their families in disease diagnosis and treatment as well as The work needs of junior medical students, interdisciplinary researchers, and medical engineering personnel jointly support the general direction of "patient-centered".

In some embodiments, the open source components of the present application include: Visual Studio 2022 community edition development platform, core research and development platform;) VTK (Visualization Toolkit), for visual reconstruction of medical tomographic images; Operation; FO-DICOM, used for reading and writing of medical DICOM images.

Optionally, an automatic head clinical image-based automatic positioning method further includes: acquiring test data, performing time-consuming tests and functional tests on the head clinical image-based automatic positioning method according to the test data, and obtaining test results.

In some embodiments, Table 3 is an example table of test data. In Table 3, data 1, 2, and 3 have different positions, different head scan intervals, and come from three patients with very different face shapes. Therefore, the scope of application of the method in this paper can be evaluated to a certain extent, and each set of data is repeated twice. , a more objective repeatability evaluation can be obtained. Different from the current mainstream registration algorithm that requires a "complete structure" of image data and reflects the diverse characteristics of clinical real image scanning parts, three sets of representative tomographic image data sets were screened in this test. Among them, data 1 is a CT image, data 2 to 3 are MRI images with poor spatial resolution, the spatial resolutions are 512*512*121, 288*384*18, 160*126*160, and the pixel pitches are respectively 0.43*0.43*0.70, 0.625*0.625*3.15, 1.625*1.625*1.625. Considering that to prevent over-medication and control medical costs, clinical image scanning emphasizes the specificity of the lesion area and performs local imaging. The scanning interval of data 1 only covers the forehead to the nose, and the range of data 2 and 3 is further compressed to From the eyebrow arch to the nose wing, you can better test the robustness of this project scheme in registration.

Table 1 Example table of test data

In some embodiments, the time-consuming test of the head clinical image-based automatic positioning method based on the test data includes: performing three-dimensional reconstruction on the same scene with the CT/MRI and the CVH dataset after registration transformation, and according to the two to evaluate the accuracy of registration. Since CVH is a 24-bit true-color bitmap data set, which is completely different from CT and MRI imaging characteristics, the coloring range and transparency scheme of the two sets of volume rendering can be clearly separated. During the test, some typical anatomical structures were selected to be mapped onto the CT/MRI images, and the registration accuracy was evaluated according to the proportion of the covered parts. Typical anatomical structures include brainstem and optic nerve, the positional relationship between the two in the head is more representative, and the shape and size are moderate.

In some embodiments, Table 2 is the test results. As shown in Table 2, on i5-4210, 2.60GH in *2, 8GB memory, Windows 7 ordinary personal computer configuration, a total of five repeatability tests were carried out, all of which were running normally, and the method did not take more than 40 seconds , faster.

Table 2

In some embodiments, the functional test of the automatic positioning method based on head clinical images is carried out according to the test data, and the images of data 1, 2, and 3 are respectively registered using the method of the present application to obtain three registration results. In the same-scene overlay of the unregistered clinical image and the original spatial arrangement of the CVH head, the spatial difference between the two can be clearly seen. Observation screenshots of two fixed viewing angles after registration. Through observations from different viewing angles, it is subjectively believed that the registration results are very ideal. This method is also a reference basis for ordinary operators to perform more than three corrected registrations. On this basis, five-year-qualified radiologists were invited to evaluate the observation results on a five-point scale, and after performing three registrations on the three groups of data, they obtained nine registration results. The distribution of the scores is shown in Table 3 below:

table 3

As shown in Table 3, among the 9 registrations, 7 registrations were successful once according to the process, and the effect was satisfactory, and 2 registrations required another similar registration correction to be satisfactory, and the purpose of registration was not basically achieved and the following situations. And from the perspective of professional radiologists, it is generally believed that the registration results meet the general needs of the public proposed by this project.

In some embodiments, the function test of the head clinical image-based automatic positioning method is performed according to the test data, including: selecting the brainstem and the optic nerve for calibration measurement, and performing the function test according to the measurement results. Data 1, 2, and 3 were registered twice respectively, and the tomograms of the same position in one of the registrations were randomly selected for calibration. In some embodiments, one of the registered tomograms at the same position is arbitrarily extracted for calibration, and multiple frames of brainstem maps and optic nerve maps of the complete anatomical structure are obtained, as well as brainstem maps of the anatomical structures cut with the sampling plane diagram and optic nerve map. On this basis, five-year-qualified radiologists were invited to evaluate the observation results on a five-point scale, and each of the three groups of data was registered three times to obtain a score of nine registration effects. The distribution of the scores is shown in Table 4 below:

Table 4

The mapping accuracy of the brainstem and optic nerve in the table shows that 3 and 5 of the 9 registrations have a matching degree of A, and 6 and 4 times of matching are respectively B. For heterogeneous data, It is impossible to achieve a perfect fit under rigid and similar registration, but most of the matching has achieved the purpose of this project to quickly read clinical tomographic images for patients with common diseases, interdisciplinary researchers, medical engineers, and junior medical students. And it has great advantages in light weight and processing speed. And from the perspective of professional radiologists, it is generally believed that the registration results meet the general needs of the public proposed by this project.

In some embodiments, the present application proposes an automatic positioning method based on head clinical images, which combines the tomographic image data (such as CT, MRI) of the head region from any clinical patient with the Chinese Visual Human Dataset (CVH) Carry out rapid regional similarity registration and spatial mapping, and mark the specific anatomical names of each region of clinical image data according to CVH's existing general anatomical structure knowledge atlas, so that the original abstract clinical tomographic images can be compared with CVH high-definition atlases, and Through a variety of retrieval methods, any anatomical region can be intuitively displayed in the form of multi-planar reconstruction (MPR) and volume reconstruction (VR), which reflects certain technological innovations. This method has the characteristics of simple operation, light calculation, and green installation-free. It can intuitively present the interested area of the current image to the user in the form of intelligent navigation, realize the purpose of "explaining" with unmanned intelligent assistance, and serve patients and their families. The active participation needs of the students, as well as the work needs of junior medical students, interdisciplinary researchers, and medical engineering personnel, reflect a certain degree of application innovation. According to the test results, it can be seen that this method can realize the function of heterogeneous multimodal registration, can meet the functional goals proposed in this paper, and has good adaptability and repeatability to clinical image data, and is suitable for the general public Requirements; the processing time on ordinary personal computers is less than 40 seconds, which has a very good lightweight advantage; and can run on any Windows desktop system. NET Framework4. Quantify the development requirements of the intelligent positioning system.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it still The technical solutions described in the foregoing embodiments may be modified, or some or all of the technical features may be equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention. , which should be included within the scope of the claims and description of the present invention.

Claims (9)

1. An automatic positioning method based on head clinical images, characterized in that, comprising: Obtain clinical images of the patient's head; Registering the clinical image of the patient's head with the CVH to obtain a spatial direction mapping relationship and a spatial dimension mapping relationship; Perform coordinate transformation on the head anatomical structure of the CVH according to the spatial direction mapping relationship and the spatial size mapping relationship to obtain the CVH registration head anatomy structure; the CVH registration head anatomy structure and the CVH registration head anatomy The position identification and organ name of each organ region in the structure constitute a new CVH anatomical knowledge map; performing multi-plane reconstruction on the patient's head clinical image to obtain the patient's head anatomy; there is a real-time mapping relationship between the patient's head anatomy and the CVH registration head anatomy; The region of interest is selected from the anatomical knowledge map, and the corresponding position of the region of interest in the anatomical structure of the patient's head is automatically located. 2. The method according to claim 1, characterized in that, registering the patient's head clinical image and CVH to obtain a spatial direction mapping relationship and a spatial dimension mapping relationship, comprising: Performing the first registration on the clinical image of the patient's head and the CVH to obtain the spatial direction mapping relationship and the first registration result; A second registration is performed on the clinical image of the head of the patient and the CVH according to the first registration result to obtain a spatial dimension mapping relationship. 3. The method according to claim 2, wherein the first registration is performed on the clinical image of the patient's head and the CVH, and the spatial direction mapping relationship and the first registration result are obtained, including: Perform three-dimensional contour reconstruction on the clinical image of the patient's head to obtain the three-dimensional contour structure of the patient; Acquiring first feature points according to the patient's three-dimensional contour structure; Obtain the second feature point of the CVH three-dimensional contour structure; Registering the first feature point and the second feature point to obtain a spatial direction mapping relationship and a first registration result; the first registration result is a CVH initial registration three-dimensional contour structure, and the patient's three-dimensional The contour structure is consistent with the spatial direction of the three-dimensional contour structure in the first CVH registration. 4. The method according to claim 3, wherein obtaining the first feature point according to the patient's three-dimensional contour structure comprises: The eye-nose triangle is selected from the patient's three-dimensional contour structure, and the first feature point of the eye-nose triangle is generated. 5. The method according to claim 3, wherein, according to the first registration result, the clinical image of the patient's head and the CVH are registered for the second time to obtain a spatial dimension mapping relationship, including: Acquiring a first circular contour according to the patient's three-dimensional contour structure; Obtain the second annular contour of the CVH three-dimensional contour structure; In the three-dimensional contour structure of the patient and the three-dimensional contour structure for the first registration of the CVH, registration is performed according to the first circular contour and the second circular contour to obtain a spatial dimension mapping relationship. 6. The method according to claim 3, wherein, according to the spatial direction mapping relationship and the spatial size mapping relationship, the head anatomical structure of the CVH is coordinate transformed to obtain the CVH registration head anatomical structure, including : Acquiring a homogeneous transformation matrix according to the spatial direction mapping relationship and the spatial size mapping relationship; the homogeneous transformation matrix is used to characterize the coordinate transformation relationship between the patient's three-dimensional contour structure and the head anatomical structure of CVH; Coordinate transformation is performed on the head anatomical structure of the CVH according to the homogeneous transformation matrix to obtain the CVH registered head anatomical structure. 7. The method according to claim 6, wherein the homogeneous transformation matrix is obtained by the following formula:

Among them, matrix1 is the transformation matrix of the spatial direction mapping relationship, matrix2 _i is the transformation matrix of the spatial dimension mapping relationship, matirx is the homogeneous transformation matrix, and i and n are positive integers. 8. The method of claim 1, comprising: After automatically locating the corresponding position of the region of interest in the anatomical structure of the patient's head, the region of the corresponding position is marked with a color, and the plane to which the region of the corresponding position belongs is displayed. 9. A head clinical image automatic positioning device, characterized in that it comprises: an acquisition module configured to acquire clinical images of the patient's head; The registration module is configured to register the clinical image of the patient's head with the CVH to obtain a spatial direction mapping relationship and a spatial size mapping relationship; The coordinate transformation module is configured to perform coordinate transformation on the head anatomical structure of the CVH according to the spatial direction mapping relationship and the spatial dimension mapping relationship to obtain the CVH registration head anatomy structure; the CVH registration head anatomy structure and the CVH registration The location identification and organ names of each organ region in the head anatomy form a new CVH anatomical knowledge map; The multi-plane reconstruction module is configured to perform multi-plane reconstruction on the patient's head clinical image to obtain the patient's head anatomy; there is a real-time mapping relationship between the patient's head anatomy and the CVH registration head anatomy; The automatic positioning module is configured to select the region of interest from the anatomical knowledge map, and automatically locate the corresponding position of the region of interest in the anatomical structure of the patient's head.

Images