Atlas assisted drawing method, computer readable storage medium and electronic device

Abstract

The present invention provides a map-assisted drawing method, a computer-readable storage medium and an electronic device, which includes steps S1, constructing an axial correspondence relationship of a brain slice; step S2, normalizing old map data; step S3, extracting regional features; step S4, nonlinear registration; and step S5, integrating old map information, so as to achieve accurate matching of brain region positions and boundary information on the old map to the image of a new map to be drawn.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a graph-assisted drawing method, a computer-readable storage medium and an electronic device. Background Art

[0002] Brain maps are often called Baidu maps in the fields of neuroscience and medicine. They are the basis for conducting work in different directions, such as neuronal projection positioning, surgical operations, precise drug delivery, and brain functional circuit research. The history of brain maps is very long. As early as the 19th century, when neuroscience was born, pioneers in the field of neuroanatomy began to observe the morphology and distribution characteristics of neurons on brain slices through microscopes, and based on this, they divided independent brain regions on white paper.

[0003] To this day, although computer-assisted drawing technology has replaced traditional paper and pen, the basic strategy used by neuroanatomists to draw brain maps is still essentially the same as it was more than a hundred years ago: after collecting brain slices, high-magnification photographs of the brain slices are taken through a microscope, then transferred to a computer and opened using vector drawing software represented by Adobe Illustrator or Corel Draw. Tools such as brushes are used within the software to identify and draw the boundaries of different brain regions.

[0004] Generation after generation of neuroanatomists continue to draw brain maps of the same species. This is not repetitive work, because with the advancement of imaging technology, people can continue to observe clearer image texture details such as neuron shape and size on brain slices; on the other hand, sample labeling technology is also rapidly improving, which can more specifically mark cells involved in specific functional circuits and expressing specific functions, allowing people to divide the brain areas on the old version of the map into more detailed sub-divisions.

[0005] In the above process, atlas reuse is a technical step that is easily overlooked but very critical. The so-called atlas reuse refers to the process of atlas drawing, referring to other published whole-brain atlases or specific brain area atlases of the same species, and comparing the similarities and differences in cell architecture characteristics between the brain slice images (hereinafter referred to as the images at hand) prepared for drawing the atlas and the images at the same spatial position of the old version of the atlas, identifying each brain region on the image at hand and drawing its boundaries. In atlas reuse, the brain area distribution and boundary information on the old version of the atlas are very important references, which can be used to quickly determine the specific location and boundary lines of each brain region on the image at hand. Compared with judging from scratch directly based on experience, it can greatly improve work efficiency. Therefore, even experts who have decades of neuroanatomy training will refer to the existing work results of their predecessors through atlas reuse in the process of drawing the atlas.

[0006] As a critical step in the process of graph drawing, graph reuse can be divided into two strategies.

[0007] The first is human eye comparison. First, a neuroanatomy expert selects an image in the old version of the atlas that is closest to the image at hand in the anatomical space. Then, he finds several characteristic structures that are easily recognizable by the human eye on the image of the old version of the atlas. By comparing the two images, he finds the location of these characteristic structures on the image at hand. Finally, starting from these characteristic structures, according to the similarity of the spatial distribution relationship of brain regions in the brains of different individuals of the same species, he infers the distribution of brain regions around these characteristic structures on the image at hand, and then draws the boundaries of all brain regions on the entire image.

[0008] However, the images used to draw the atlas usually contain a variety of rich cell architectural details such as cell morphology, cell distribution and cell density. It is inevitable that the human eye will miss some of the key information that can be used to identify brain regions. On the other hand, in the absence of experience, it is difficult for the atlas drawer to form a complete map of the spatial distribution and adjacent relationships of various brain regions in the whole brain. There is a lack of sensitivity to the location of brain regions in space. If the cell architectural details are used as the only criterion for judgment, it may even lead to the misidentification of brain regions. The most typical example of the latter is the cerebellum. The cerebellum appears as a coronal on the standard coronal plane, and each "petal" is a lobe. Each lobe performs a specific function, but there is almost no difference in cell architecture between them, and their spatial positions are adjacent. If you browse the coronal planes containing the cerebellum area one by one in the head and tail direction, you can find that each lobe will appear and disappear quickly, and the surrounding other lobes will gradually fill the area vacated by the disappearance of the lobe. Because the cerebral lobes appear sequentially in the coronal plane along the head-to-tail direction, once a lobe is misidentified, all subsequent lobes will be misidentified, resulting in catastrophic consequences for anatomical spatial positioning.

[0009] The second is overlay matching. After selecting an image from the old version of the atlas that is closest to the position of the image at hand, the image with the brain area boundaries drawn in the old version of the atlas is overlaid on the image at hand. Before the popularization of computer technology in the last century, neuroanatomists made slides of the old atlas and the images at hand, and then projected them onto blank paper on the desktop with a slide projector. With the help of a slide projector, the brain area boundary information of the old atlas can be overlaid on the newly collected image at hand by the user, thereby providing an effective reference for atlas drawing; after the 1980s, more scientists tried to use vector drawing software to simplify the above process, which greatly accelerated the efficiency of atlas reuse.

[0010] Compared with the previous strategy, the biggest improvement of the superposition matching strategy is that it can provide a comprehensive and real-time reference for the location of brain regions, so that the cartographer no longer has to worry about missing any information in the human eye comparison. However, this strategy still cannot avoid the occurrence of misidentification of brain regions. The reason is that there may be a large difference between the sample preparation technology of the images used in the old version of the atlas and the sample preparation technology used by the cartographer to obtain the images at hand, resulting in different shrinkage rates and irregular deformation states of brain tissue, and ultimately leading to large differences in image size, brain region shape and spatial distribution. Simply superimposing the two images together does not eliminate the impact of this difference. Therefore, the brain region boundaries of the old version of the atlas are directly applied to the new image without modification, and the resulting brain region boundaries are often wrong.

[0011] In summary, whether it is human eye comparison or superposition matching, the current atlas reuse strategy of neuroanatomy experts has its inherent defects, resulting in low accuracy and reducing the reference value of the old atlas. Therefore, a new atlas reuse method must be developed to improve accuracy. Summary of the invention

[0012] Based on the aforementioned defects of the prior art, the present invention provides a map-assisted drawing method with high accuracy, a computer-readable storage medium and an electronic device.

[0013] In order to achieve the above object, the present invention provides a method for auxiliary drawing of a map, comprising:

[0014] Step S1, constructing the axial correspondence of brain slices: constructing the correspondence according to the landmark anatomical structures on the old atlas and the new atlas to be drawn;

[0015] Step S2, normalizing the old version of the atlas data: converting the old version of the atlas file into a reusable .xml data format;

[0016] Step S3, regional feature extraction: marking the landmark anatomical structures on the old atlas and the new atlas to be drawn as regional features;

[0017] Step S4, nonlinear registration: based on the regional features and the corresponding relationship between the old map and the new map to be drawn, the old map is nonlinearly registered to the new map to be drawn to obtain the deformation field;

[0018] Step S5, integrating old map information: based on the correspondence between the old map and the new map to be drawn, the old map in .xml data format and the deformation field, the old map is superimposed on the new map to be drawn.

[0019] In one embodiment, the step S1, constructing the axial correspondence of the brain slices, comprises:

[0020] Acquire the first image sequence of the old atlas; acquire the first image sequence of the new atlas to be drawn; select landmark anatomical structures on the first image sequence of the old atlas and the first image sequence of the new atlas to be drawn, and for each image in the first image sequence of the new atlas to be drawn, find an image in the first image sequence of the old atlas that is closest to the image in terms of brain area distribution according to the morphology of the landmark anatomical structures and their changes in the anterior-posterior direction, thereby constructing a mapping function.

[0021] In one embodiment, the landmark anatomical structures include the hippocampus, cerebellum, and anterior commissure.

[0022] In one embodiment, the step S2, normalizing the old version of the atlas data, comprises:

[0023] The first image sequence of the old version of the atlas is converted into the second image sequence of the old version of the atlas in .svg format using vector graphics software. The .svg format file stores the boundary information of each brain region in the form of ascii code. The boundary of each brain region is independently represented by a pair of <path>< / path> The path label is used to record the brain region boundary. In the path label, the brain region boundary is divided into multiple line segments and curve segments. Different segments are recorded with different letters as the beginning and followed by the coordinate values ​​of several points. Different letters indicate different numbers of control points and nodes.

[0024] Converting a .svg format file to a .xml format file includes: traversing each image in the second image sequence of the old version of the atlas; for each image, traversing all its path labels; for each path label, first creating two empty queue data structures, respectively recorded as the first queue and the second queue, and then using letters as markers, using a string segmentation function to divide it into a number of line segments and curve segments, and then traversing each segment in turn, storing the coordinate value of each point of the segment at the end of the first queue, and then storing it at the end of the second queue with different characters according to whether the point is a control point or a node; performing the above operations until After traversing all the line segments and curve segments under the path label, the first queue now records the coordinate values ​​of all points on the boundary of the brain region, and the second queue records whether the attributes of these points are control points or nodes; after traversing all the path labels of each image, create an .xml format file, in which a brain region label is created for each brain region to store its path information, and in this label, the information of the first and second queues obtained previously is placed; after traversing all the images and performing the above operations, the third image sequence of the old version of the atlas in the .xml format file corresponding to the first image sequence of the old version of the atlas is formed.

[0025] In one embodiment, the different letters represent different numbers of control points and nodes, including: C represents a cubic Bezier curve, including two control points and one node; S represents a simplified Bezier curve, including one control point and one node; L represents a straight line segment, including only one node.

[0026] In one embodiment, the step S3, regional feature extraction, includes:

[0027] Using vector graphics software to convert each image in the second image sequence of the old version of the atlas into a bitmap, and then marking the landmark anatomical structure selected in step S1, while leaving the remaining structures unmarked, to form a marked fourth image sequence of the old version of the atlas;

[0028] The boundaries of the landmark anatomical structures are marked on each image of the first image sequence of the new atlas to be drawn as regional features used in subsequent registration calculations, thereby forming a second image sequence of the new atlas to be drawn with marked regional features.

[0029] In one embodiment, the step S4, nonlinear registration, comprises:

[0030] Based on the regional features extracted in step S3, for each image in the second image sequence of the new atlas to be drawn, the corresponding image in the fourth image sequence of the old atlas is selected according to the mapping function obtained in step S1, and the fourth image sequence of the old atlas is aligned to the second image sequence of the new atlas to be drawn using a nonlinear registration algorithm based on the differential homeomorphic large deformation mapping model to obtain the deformation field.

[0031] In one embodiment, the step S5, integrating old version atlas information, includes:

[0032] Traverse all images of the first image sequence of the new atlas to be drawn, and for any image therein, select the image corresponding to it in the third image sequence of the old atlas through the mapping function obtained in step S1, traverse all control points and nodes in the image, use the deformation field obtained in step S4 to obtain the coordinates of the new positions to which these control points and nodes need to be moved after nonlinear alignment, use the coordinate values ​​of these new positions to replace the original coordinate values ​​in the first queue of the corresponding control points and nodes in the third image sequence of the old atlas, and then save them as new images to form the fifth image sequence of the old atlas;

[0033] Traverse each image in the fifth image sequence of the old version of the atlas, for each image, traverse all its brain region labels, and create an empty .svg format file for storing all brain region information contained in the image; for each brain region label, sequentially read the second queue information of each point, create line segments or curve segments according to the distribution rules of control points and nodes, start with different letters, and record the coordinate values ​​of each point; finally, concatenate the strings corresponding to different segments in sequence, so as to put the entire brain region boundary information into a string, and put the string into a <path>< / path> path label; perform the above operation on each brain region label in turn, thereby obtaining a .svg format file containing all brain region boundary information of each image; traverse each image, perform the above operation, and form the sixth image sequence of the old version of the atlas of the .svg format file;

[0034] Traverse each image in the first image sequence of the new map to be drawn, use vector graphics drawing software to import the image in the first image sequence of the new map to be drawn as the background layer, and then use the mapping function obtained in step S1 to select the corresponding image in the sixth image sequence of the old version of the map, import it as a new layer, superimpose it on the background layer, and match the brain area boundary information of the old version of the map to the image of the new map to be drawn.

[0035] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and the computer program executes the above-mentioned atlas-assisted drawing method when read and executed by a processor.

[0036] The present invention also provides an electronic device, comprising: a processor, a memory and a communication bus, wherein the processor is connected to the memory via the communication bus for communication to execute the above-mentioned map-assisted drawing method.

[0037] The atlas-assisted drawing method, computer-readable storage medium and electronic device described in the present invention, through step S1, constructing the axial correspondence relationship of brain slices; step S2, normalizing the old atlas data; step S3, extracting regional features; step S4, nonlinear alignment; step S5, integrating the old atlas information, can achieve accurate matching of the brain area position and boundary information on the old atlas to the image of the new atlas to be drawn. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The drawings described herein are only for explanation purposes and are not intended to limit the scope of the present invention in any way. In addition, the shapes and proportional dimensions of the components in the drawings are only for illustration purposes and are used to help understand the present invention. They do not specifically limit the shapes and proportional dimensions of the components of the present invention. Under the guidance of the present invention, those skilled in the art can select various possible shapes and proportional dimensions to implement the present invention according to specific circumstances. In the drawings:

[0039] Figure 1 A schematic flow chart of a method for assisting atlas drawing provided in the first embodiment of the present invention. DETAILED DESCRIPTION

[0040] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0041] See also Figure 1 As shown, the first embodiment of the present invention provides a method for auxiliary drawing of a map, comprising:

[0042] Step S1, constructing the axial correspondence of brain slices: constructing the correspondence according to the landmark anatomical structures on the old atlas and the new atlas to be drawn;

[0043] Step S2, normalizing the old version of the atlas data: converting the old version of the atlas file into a reusable .xml data format;

[0044] Step S3, regional feature extraction: marking the landmark anatomical structures on the old atlas and the new atlas to be drawn as regional features;

[0045] Step S4, nonlinear registration: based on the regional features and the corresponding relationship between the old map and the new map to be drawn, the old map is nonlinearly registered to the new map to be drawn to obtain the deformation field;

[0046] Step S5, integrating old map information: based on the correspondence between the old map and the new map to be drawn, the old map in .xml data format and the deformation field, the old map is superimposed on the new map to be drawn.

[0047] In one embodiment, the step S1, constructing the axial correspondence of the brain slices, comprises:

[0048] Acquire the first image sequence of the old atlas; acquire the first image sequence of the new atlas to be drawn; select landmark anatomical structures on the first image sequence of the old atlas and the first image sequence of the new atlas to be drawn, and for each image in the first image sequence of the new atlas to be drawn, find an image in the first image sequence of the old atlas that is closest to the image in terms of brain area distribution according to the morphology of the landmark anatomical structures and their changes in the anterior-posterior direction, thereby constructing a mapping function.

[0049] Among them, the landmark anatomical structures include the hippocampus, cerebellum, and anterior commissure.

[0050] In a specific embodiment, the first image sequence of the old version of the atlas is A1={ai | i = 1, 2, 3, ...N}, where N is the total number of images contained in the old version of the atlas; the first image sequence of the new atlas to be drawn is B1={bi | i = 1, 2, 3, ...M}, where M is the total number of images contained in the first image sequence of the new atlas to be drawn. Wherein N can be greater than M, or less than M, or equal to M, and the present invention is not particularly limited to this. Select landmark anatomical structures on A1 and B1, such as the hippocampus, cerebellum, anterior commissure, etc., and according to their morphology and changes in the anterior-posterior direction, for each image bi in B1, find an image aj in A1 that is closest to it in brain area distribution, thereby constituting a standard mapping function f: f (bi)=aj, where i = 1, 2, 3...M, j∈{1, 2, 3...N}.

[0051] In one embodiment, the step S2, normalizing the old version of the atlas data, comprises:

[0052] The first image sequence of the old version of the atlas is converted into the second image sequence of the old version of the atlas in .svg format using vector graphics software. The .svg format file stores the boundary information of each brain region in the form of ascii code. The boundary of each brain region is independently represented by a pair of <path>< / path> The path label is used to record the brain region boundary. In the path label, the brain region boundary is divided into multiple line segments and curve segments. Different segments are recorded with different letters as the beginning and followed by the coordinate values ​​of several points. Different letters indicate different numbers of control points and nodes.

[0053] Converting a .svg format file to a .xml format file includes: traversing each image in the second image sequence of the old version of the atlas; for each image, traversing all its path labels; for each path label, first creating two empty queue data structures, respectively recorded as the first queue and the second queue, and then using letters as markers, using a string segmentation function to divide it into a number of line segments and curve segments, and then traversing each segment in turn, storing the coordinate value of each point of the segment at the end of the first queue, and then storing it at the end of the second queue with different characters according to whether the point is a control point or a node; performing the above operations until After traversing all the line segments and curve segments under the path label, the first queue now records the coordinate values ​​of all points on the boundary of the brain region, and the second queue records whether the attributes of these points are control points or nodes; after traversing all the path labels of each image, create an .xml format file, in which a brain region label is created for each brain region to store its path information, and in this label, the information of the first and second queues obtained previously is placed; after traversing all the images and performing the above operations, the third image sequence of the old version of the atlas in the .xml format file corresponding to the first image sequence of the old version of the atlas is formed.

[0054] The different letters represent different numbers of control points and nodes, including: C represents a cubic Bezier curve, including two control points and one node; S represents an abbreviated Bezier curve, including one control point and one node; L represents a straight line segment, including only one node.

[0055] Because with the popularization of computer-aided drawing technology, neuroanatomists now usually use vector drawing software such as Adobe Illustrator to draw atlases, and then directly publish the vector source file in .ai format, or make it into an e-book in .pdf format, publish it as an attachment to a paper or publish it as a separate book. Therefore, the first image sequence of the old version of the atlas may be a vector source file in .ai format or a .pdf format file. In one embodiment, if the old version of the atlas is a vector source file in .ai format, it is first converted into a .svg format file using vector drawing software (such as Adobe Illustrator); if it is in .pdf format, it is first opened page by page using vector drawing software, and then each page is converted into a .svg format file, so as to obtain the second image sequence A2 of the old version of the atlas composed of .svg format files.

[0056] Because the .svg file in A2 saves the boundary information of each brain region in the form of human-readable ASCII code; specifically, the boundary of each brain region is independently represented by a pair of <path>< / path>Labels, that is, path labels, are used to record. Therefore, each image in A2 is then traversed, and for each current image, all its path labels are traversed. For each path label, that is, the boundary of the corresponding brain region, two empty queue data structures are first created, denoted as P and T, and then letters (such as C, S, L) are used as markers to divide it into several line segments and curve segments using the string segmentation function, and then each segment is traversed in turn, and the coordinate information of each point in the segment is stored at the end of queue P, and then the characters "n" and "c" are used as values ​​according to whether the point is a node or a control point, respectively, and stored at the end of queue T. The above steps are performed until all line segments and curve segments under the path label are traversed. At this time, queue P records the coordinate values ​​of all points on the boundary of the brain region, and queue T records whether the attributes of these points are nodes or control points. After traversing all path labels of the current image, a .xml format file is created. In the .xml file, a brain region label is created for each brain region to store its path information, and the information of queue P and queue T obtained previously is placed in the label. After traversing all images and performing the above operations, the .xml format files corresponding to each image of the old atlas can be obtained. At this time, the .xml files constitute the third image sequence A3 of the old atlas.

[0057] In one embodiment, the step S3, regional feature extraction, includes:

[0058] Using vector graphics software to convert each image in the second image sequence of the old version of the atlas into a bitmap, and then marking the landmark anatomical structure selected in step S1, while leaving the remaining structures unmarked, to form a marked fourth image sequence of the old version of the atlas;

[0059] The boundaries of the landmark anatomical structures are marked on each image of the first image sequence of the new atlas to be drawn as regional features used in subsequent registration calculations, thereby forming a second image sequence of the new atlas to be drawn with marked regional features.

[0060] In a specific embodiment, the bitmap may be in .png format or .tif format. The fourth image sequence of the old version of the atlas is A4, and the second image sequence of the new atlas to be drawn is B2.

[0061] In one embodiment, the step S4, nonlinear registration, comprises:

[0062] Based on the regional features extracted in step S3, for each image in the second image sequence of the new atlas to be drawn, the corresponding image in the fourth image sequence of the old atlas is selected according to the mapping function obtained in step S1, and the fourth image sequence of the old atlas is aligned to the second image sequence of the new atlas to be drawn using a nonlinear registration algorithm based on the differential homeomorphic large deformation mapping model to obtain the deformation field.

[0063] Among them, the nonlinear registration calculation based on the differential homeomorphism large deformation model can well simulate the irregular and anisotropic spatial deformation of biological brain tissue samples and establish accurate spatial correspondence between different brain tissues. The deformation field obtained is W={wi | i = 1, 2, 3, …N}.

[0064] In one embodiment, the step S5, integrating old version atlas information, includes:

[0065] Traverse all images of the first image sequence of the new atlas to be drawn, and for any image therein, select the image corresponding to it in the third image sequence of the old atlas through the mapping function obtained in step S1, traverse all control points and nodes in the image, use the deformation field obtained in step S4 to obtain the coordinates of the new positions to which these control points and nodes need to be moved after nonlinear alignment, use the coordinate values ​​of these new positions to replace the original coordinate values ​​in the first queue of the corresponding control points and nodes in the third image sequence of the old atlas, and then save them as new images to form the fifth image sequence of the old atlas;

[0066] Traverse each image in the fifth image sequence of the old version of the atlas, for each image, traverse all its brain region labels, and create an empty .svg format file for storing all brain region information contained in the image; for each brain region label, sequentially read the second queue information of each point, create line segments or curve segments according to the distribution rules of control points and nodes, start with different letters, and record the coordinate values ​​of each point; finally, concatenate the strings corresponding to different segments in sequence, so as to put the entire brain region boundary information into a string, and put the string into a <path>< / path> path label; perform the above operation on each brain region label in turn, thereby obtaining a .svg format file containing all brain region boundary information of each image; traverse each image, perform the above operation, and form the sixth image sequence of the old version of the atlas of the .svg format file;

[0067] Traverse each image in the first image sequence of the new map to be drawn, use vector graphics drawing software to import the image in the first image sequence of the new map to be drawn as the background layer, and then use the mapping function obtained in step S1 to select the corresponding image in the sixth image sequence of the old version of the map, import it as a new layer, superimpose it on the background layer, and match the brain area boundary information of the old version of the map to the image of the new map to be drawn.

[0068] In a specific embodiment, all images in B1 are traversed, and for any image bi, the corresponding image aj in A3 is selected by the mapping function f obtained in step S1. All nodes and control points in aj are traversed, and the coordinates of the new positions to which these nodes and control points need to be moved after nonlinear alignment are obtained using the deformation field wi obtained in step S4. These new coordinates are used to replace the original coordinates of the corresponding nodes or control points in aj, and then saved as a new image. The fifth image sequence of the old version of the atlas composed of the new image is recorded as A5.

[0069] Traverse each image in A5. For the current image, traverse all its brain region labels and create an empty .svg format file to store all brain region information contained in the image. For each brain region label, read the point attribute queue information sequentially, and create line segments or curve segments according to the distribution rules of nodes and control points: if a node appears after two control points, it is regarded as a cubic Bezier curve segment, and a string starting with the letter "C" is used to record the coordinate information of the three points; if a control point is followed by a node, it is regarded as an abbreviated Bezier curve segment, and a string starting with the letter "S" is used to record the coordinate information of the two points, and so on. Finally, the strings corresponding to different segments are connected in sequence, so that the entire brain region boundary information is put into a string. Put the string into a path label, that is, <path>< / path> The above operations are performed on each brain region label in turn, thereby obtaining a .svg format file containing all brain region boundary information of the current image. By traversing each image and performing the above operations, the .svg format file corresponding to each image of the old atlas can be obtained. At this time, the .svg format file constitutes the sixth image sequence A6 of the old version of the atlas.

[0070] Traverse each image bi in B1, use vector graphics drawing software, import bi as the background image, and then use the mapping function f obtained in step S1 to select the corresponding image aj in A6, and import it as a new layer and superimpose it on the background layer. In this way, the brain area boundary information of the old version of the atlas can be accurately matched to the image of the new atlas to be drawn.

[0071] The second embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, and the computer program executes the above-mentioned atlas-assisted drawing method when read and executed by a processor.

[0072] The third embodiment of the present invention also provides an electronic device, including: a processor, a memory and a communication bus, wherein the processor is connected to the memory via the communication bus to communicate with each other so as to execute the above-mentioned map-assisted drawing method.

[0073] The atlas-assisted drawing method, computer-readable storage medium and electronic device described in the present invention, through step S1, constructing the axial correspondence relationship of brain slices; step S2, normalizing the old atlas data; step S3, extracting regional features; step S4, nonlinear alignment; step S5, integrating the old atlas information, can achieve accurate matching of the brain area position and boundary information on the old atlas to the image of the new atlas to be drawn.

[0074] It should be understood that the above description is for illustration and not for limitation. Many embodiments and many applications beyond the examples provided will be apparent to those skilled in the art upon reading the above description. Therefore, the scope of the present teachings should not be determined with reference to the above description, but rather with reference to the foregoing claims and the full scope of equivalents to which such claims are entitled. For the purpose of comprehensiveness, all articles and references, including disclosures of patent applications and publications, are incorporated herein by reference. The omission of any aspect of the subject matter disclosed herein in the foregoing claims is not intended to be a waiver of such subject matter, nor should it be considered that the applicant has not considered such subject matter to be part of the disclosed application subject matter.

Claims

1. A method for assisting atlas drawing, characterized in that: include: Step S1, constructing the axial correspondence of brain slices: constructing the correspondence according to the landmark anatomical structures on the old atlas and the new atlas to be drawn; Step S2, normalizing the old version of the atlas data: converting the old version of the atlas file into a reusable .xml data format; Step S3, regional feature extraction: marking the landmark anatomical structures on the old atlas and the new atlas to be drawn as regional features; Step S4, nonlinear registration: based on the regional features and the corresponding relationship between the old map and the new map to be drawn, the old map is nonlinearly registered to the new map to be drawn to obtain the deformation field; Step S5, integrating old map information: superimposing the old map onto the new map to be drawn according to the correspondence between the old map and the new map to be drawn, the old map in .xml data format, and the deformation field; The step S2, normalizing the old version of the atlas data, comprises: The first image sequence of the old version of the atlas is converted into the second image sequence of the old version of the atlas in .svg format using vector graphics software. The .svg format file stores the boundary information of each brain region in the form of ascii code. The boundary of each brain region is independently represented by a pair of <path>< / path> The path label is used to record the brain region boundary. In the path label, the brain region boundary is divided into multiple line segments and curve segments. Different segments are recorded with different letters as the beginning and followed by the coordinate values ​​of several points. Different letters indicate different numbers of control points and nodes. Converting a .svg format file to a .xml format file includes: traversing each image in the second image sequence of the old version of the atlas; for each image, traversing all its path labels; for each path label, first creating two empty queue data structures, respectively recorded as the first queue and the second queue, and then using letters as markers, using a string segmentation function to divide it into a number of line segments and curve segments, and then traversing each segment in turn, storing the coordinate value of each point of the segment at the end of the first queue, and then storing it at the end of the second queue with different characters according to whether the point is a control point or a node; performing the above operations until After traversing all line segments and curve segments under the path label, the first queue records the coordinate values ​​of all points on the boundary of the brain region, and the second queue records whether the attributes of these points are control points or nodes; after traversing all path labels of each image, create an .xml format file, in which a brain region label is created for each brain region to store its path information, and in the label, the information of the first queue and the second queue obtained previously is placed; after traversing all images and performing the above operations, the third image sequence of the old version of the atlas in the .xml format file corresponding to the first image sequence of the old version of the atlas is formed; The step S3, regional feature extraction, comprises: Using vector graphics software to convert each image in the second image sequence of the old version of the atlas into a bitmap, and then marking the landmark anatomical structure selected in step S1, while leaving the remaining structures unmarked, to form a marked fourth image sequence of the old version of the atlas; Marking the boundary of the landmark anatomical structure on each image of the first image sequence of the new atlas to be drawn as a regional feature used in subsequent registration calculation, thereby forming a second image sequence of the new atlas to be drawn with marked regional features; The step S4, nonlinear registration, comprises: Based on the regional features extracted in step S3, for each image in the second image sequence of the new atlas to be drawn, according to the mapping function obtained in step S1, select the image corresponding to it in the fourth image sequence of the old atlas, and use a nonlinear registration algorithm based on a differential homeomorphic large deformation mapping model to register the fourth image sequence of the old atlas to the second image sequence of the new atlas to be drawn, so as to obtain a deformation field; The step S5, integrating old version atlas information, includes: Traverse all images of the first image sequence of the new atlas to be drawn, and for any image therein, select the image corresponding to it in the third image sequence of the old atlas through the mapping function obtained in step S1, traverse all control points and nodes in the image, use the deformation field obtained in step S4 to obtain the coordinates of the new positions to which these control points and nodes need to be moved after nonlinear alignment, use the coordinate values ​​of these new positions to replace the original coordinate values ​​in the first queue of the corresponding control points and nodes in the third image sequence of the old atlas, and then save them as new images to form the fifth image sequence of the old atlas; Traverse each image in the fifth image sequence of the old version of the atlas, for each image, traverse all its brain region labels, and create an empty .svg format file for storing all brain region information contained in the image; for each brain region label, sequentially read the second queue information of each point, create line segments or curve segments according to the distribution rules of control points and nodes, start with different letters, and record the coordinate values ​​of each point; finally, concatenate the strings corresponding to different segments in sequence, so as to put the entire brain region boundary information into a string, and put the string into a <path>< / path> path label; perform the above operation on each brain region label in turn, thereby obtaining a .svg format file containing all brain region boundary information of each image; traverse each image, perform the above operation, and form the sixth image sequence of the old version of the atlas of the .svg format file; Traverse each image in the first image sequence of the new map to be drawn, use vector graphics drawing software to import the image in the first image sequence of the new map to be drawn as the background layer, and then use the mapping function obtained in step S1 to select the corresponding image in the sixth image sequence of the old version of the map, import it as a new layer, superimpose it on the background layer, and match the brain area boundary information of the old version of the map to the image of the new map to be drawn.

2. The atlas-assisted drawing method according to claim 1, characterized in that: The step S1, constructing the axial correspondence of the brain slices, comprises: Acquire the first image sequence of the old atlas; acquire the first image sequence of the new atlas to be drawn; select landmark anatomical structures on the first image sequence of the old atlas and the first image sequence of the new atlas to be drawn, and for each image in the first image sequence of the new atlas to be drawn, find an image in the first image sequence of the old atlas that is closest to the image in terms of brain area distribution according to the morphology of the landmark anatomical structures and their changes in the anterior-posterior direction, thereby constructing a mapping function.

3. The atlas-assisted drawing method according to claim 2, characterized in that: The landmark anatomical structures include the hippocampus, cerebellum, and anterior commissure.

4. The atlas-assisted drawing method according to claim 2, characterized in that: The different letters represent different numbers of control points and nodes, including: C represents a cubic Bezier curve, including two control points and one node; S represents an abbreviated Bezier curve, including one control point and one node; L represents a straight line segment, including only one node.

5. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is read and executed by a processor, the atlas-assisted drawing method as described in any one of claims 1-4 is executed.

6. An electronic device, characterized in that: include: A processor, a memory and a communication bus, wherein the processor is connected to the memory via the communication bus to communicate with each other so as to execute the map-assisted drawing method as described in any one of claims 1 to 4.

Images