Real-time magnetic resonance navigation system for neurosurgery operation

Abstract

The invention relates to the technical field of surgical navigation, in particular to a real-time magnetic resonance navigation system for neurosurgery. The system comprises an image acquisition module used for acquiring intraoperative continuous frame MRI (Magnetic Resonance Imaging) images; the dynamic region identification module is used for determining an instrument connected domain according to the boundary rule distribution condition of the connected domain and determining an operation receiving position in combination with the endpoint movement condition of continuous frames; the method comprises the following steps: determining a traction area range through local tissue edge distribution of an operated position, considering a multi-frame range dynamic change condition on a preorder, and determining an intra-operation dynamic area according to approximate distribution of the operated position; and the quality adjusting module is used for performing gradient attenuation analysis on the distribution of the intraoperative dynamic region and other regions under the multi-frame image before the current moment, and determining the transmission quality of the connected region for transmission display navigation. According to the method, an intra-operative affected area is positioned after instrument area tracking, the quality of different areas is dynamically adjusted, and the transmission quality and the real-time performance are balanced.

Description

Technical Field

[0001] The present invention relates to the technical field of surgical navigation, and particularly to a real-time magnetic resonance navigation system for neurosurgery. Background Art

[0002] Neurosurgery is a type of surgical operation involving the central nervous system (brain and spinal cord) and its related structures, mainly used for treating diseases related to the brain, spinal cord, nerves, and their blood vessels. Neurosurgery not only requires extremely high skills from surgeons, but also the support of imaging techniques and surgical navigation systems is crucial to assist in improving the accuracy of the operation.

[0003] Navigation during the surgical process can be accurately carried out through imaging techniques. However, since real-time image scanning is required for navigation, the overall computational gap is relatively large. If the hardware is upgraded, the cost will be greatly increased. Usually, in order to meet the real-time requirement, the image quality needs to be reduced, and the reduction of image quality will in turn affect surgical navigation, resulting in a poor balance between image quality and real-time performance during transmission. Summary of the Invention

[0004] In order to solve the technical problem of the poor balance between image quality and real-time performance during transmission in the prior art, the purpose of the present invention is to provide a real-time magnetic resonance navigation system for neurosurgery, and the specific technical solution adopted is as follows: The present invention provides a real-time magnetic resonance navigation system for neurosurgery, and the system includes: An image acquisition module, configured to acquire intraoperative consecutive-frame MRI images; A dynamic region recognition module, configured to determine the instrument connected domain of each frame of MRI image according to the boundary rule distribution of the connected domain in each frame of MRI image; and determine the position of the operation area of each frame of MRI image by analyzing the movement degree of the end points of the instrument connected domain in the consecutive-frame MRI images. Determine the traction influence area of the operation area according to the direction distribution of the edges within the local range at the operation area of each frame of MRI image; and obtain the intraoperative dynamic region of each frame of MRI image through the proximity degree between the consecutive-frame MRI images before each frame of MRI image at the operation areas. A quality adjustment module, configured to determine the gradient attenuation quality of the connected domain in each frame of MRI image according to the position distribution of the intraoperative dynamic region of the operation area and other connected domains in each frame of MRI image; and determine the transmission quality of each connected domain at the current moment according to the gradient attenuation quality of each connected domain except the instrument connected domain in the MRI images at different times. A navigation execution module, configured to update the MRI image based on the transmission quality of different connected domains at the current moment to obtain an intraoperative transmission image and acquire a navigation path.

[0005] Further, the method for obtaining the instrument connected region includes: For any connected region in each frame of MRI image, taking any pixel point on the boundary of the connected region as the starting point, and forming a sequence of pixel points continuously distributed clockwise along the boundary as the boundary sequence of the connected region; In the boundary sequence, when the gradient direction between each pixel point and the next pixel point is different, the corresponding pixel point is taken as the turning pixel point of the connected region; count the number of turning pixel points and perform normalization processing to obtain the turning degree of the connected region. Taking the connected region with the smallest turning degree in each frame of MRI image as the instrument connected region.

[0006] Further, the method for obtaining the treatment position includes: Performing linear fitting on the instrument connected region in each frame of MRI image to obtain the instrument straight line of each frame of MRI image; For any frame of MRI image, successively taking one endpoint of the instrument straight line in the MRI image as the analysis point; calculating the displacement between the analysis points of the instrument straight lines of the adjacent frames of MRI before the MRI image and calculating the mean value to obtain the moment of inertia value of the MRI image at the analysis point; Taking the endpoint with the smallest moment of inertia value in the MRI image as the treatment position of the MRI image.

[0007] Further, the method for obtaining the traction influence region includes: For the treatment position of any MRI image, within the preset local range at the treatment position, marking the existing edges as the involved edges of the treatment position; obtaining the extension direction of each involved edge of the treatment position; For any involved edge of the treatment position, calculating the mean similarity between the involved edge and the adjacent involved edges in the extension direction to obtain the involvement significance of the involved edge; When the involvement significance is greater than the preset significance threshold, taking the corresponding involved edge as the influence edge; merging the connected regions where the image edges are located at the treatment position to obtain the traction influence region of the treatment position.

[0008] Further, the method for obtaining the intraoperative dynamic region includes: For any frame of MRI image, successively taking each frame of MRI image before the MRI image as the analysis image; Calculating the Euclidean distance between the treatment positions of the MRI image and the analysis image as the similarity parameter of the analysis image; taking the analysis image with the similarity parameter less than the preset similarity threshold as the approximate image of the MRI image; Overlay the traction influence region in the approximate image with the traction influence region in the MRI image to obtain the intraoperative dynamic region of the MRI image.

[0009] Further, the method for obtaining the gradient attenuation quality includes: For any frame of MRI image, take the surgical position in the MRI image as the center of a circle, and use the farthest distance from the center of the circle to the boundary of the intraoperative dynamic region as the radius to draw a circle, and determine the boundary of the highest quality range of the MRI image; the quality coefficient within the boundary of the highest quality range is the preset maximum quality coefficient; Calculate the distance between the surgical position in the MRI image and the boundary of each other connected domain, draw a circle with the length of the farthest distance from the boundary of other connected domains as the radius, and determine the boundary of the maximum loss quality range of the MRI image; the quality coefficient on the boundary of the maximum loss quality range is the preset minimum quality coefficient; Decay the quality coefficient between the boundary of the highest quality range and the boundary of the maximum loss quality range to obtain the quality coefficient at each position; Calculate the average value of the quality coefficients existing in each connected domain in the MRI image as the gradient attenuation quality of each connected domain.

[0010] Further, the step of decaying the quality coefficient between the boundary of the highest quality range and the boundary of the maximum loss quality range to obtain the quality coefficient at each position includes: Take the shortest connection line from the boundary of the highest quality range to the boundary of the maximum loss quality range as the gradient descent line; Starting from the quality coefficient at the boundary of the highest quality region, use the attenuation function to decay the quality coefficient along the gradient descent line to the quality coefficient on the boundary of the maximum loss quality range, and determine the quality coefficient at each position on the gradient descent line.

[0011] Further, the method for obtaining the transmission quality includes: In the MRI image at the current moment, take the maximum gradient attenuation quality of each connected domain except the instrument connected domain in all other MRI images as the transmission quality of each connected domain at the current moment.

[0012] Further, the method for obtaining the intraoperative transmission image includes: Take the MR image and the transmission quality situation at the current moment as inputs, and output the intraoperative transmission image through the trained image transmission model.

[0013] Further, the step of obtaining the extension direction of each traction edge at the surgical position includes: Perform linear fitting on each traction edge, and take the direction of the fitting line as the extension direction of each traction edge.

[0014] The present invention has the following beneficial effects: The present invention determines the surgical tool part based on the boundary regular distribution of the connected region and the movement of consecutive frames, and then tracks and locates the position of the operation site. Subsequently, it is convenient to analyze the operation area at the operation site to ensure high-quality development of local images. Further, it is ensured that the areas affected by tissue traction during the operation can also be observed and analyzed. According to the local deformation edge generated at the operation site during the operation, the range of the traction area of the operation is determined through the distribution of the local tissue edge at the operation site. Considering the operation range involved many times in the previous sequence, the intraoperative dynamic area affected during the operation is determined according to the distribution proximity of the operation site, making the key attention area in each frame of image more accurate. Then, through the gradient attenuation analysis of the distribution between the intraoperative dynamic area and the remaining areas that are key attention areas in each frame of image, the attenuation quality is obtained. Considering the integrity and consistency of the regional tissue during the operation, the transmission quality of the final connected region is determined through the gradient attenuation quality in different image frames before the current moment for transmission display navigation. The present invention locates the intraoperative affected area after tracking the instrument area, dynamically adjusts the quality of different areas, and balances the transmission quality and real-time issues. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 It is a structural block diagram of a real-time magnetic resonance navigation system for a neurosurgical operation provided by an embodiment of the present invention; Figure 2 It is a basic operation flowchart of real-time navigation in an existing neurological operation provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of a global high-quality MRI image provided by an embodiment of the present invention; Figure 4 It is a schematic diagram of some surgical tool instruments provided by an embodiment of the present invention; Figure 5 It is a schematic diagram of a single connected region provided by an embodiment of the present invention; Figure 6 It is a side schematic diagram of the moment of inertia of a surgical tool provided by an embodiment of the present invention; Figure 7 It is a schematic diagram of the traction situation of the operation site provided by an embodiment of the present invention; Figure 8Schematic diagram of the boundary of the highest quality range and the boundary of the maximum loss quality range provided by an embodiment of the present invention. Detailed implementation manners

[0017] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following combines the accompanying drawings and preferred embodiments to detail the specific implementation manners, structures, features and effects of a real-time magnetic resonance navigation system for neurosurgery proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0019] The following specifically describes the specific solution of a real-time magnetic resonance navigation system for neurosurgery provided by the present invention with reference to the accompanying drawings.

[0020] Please refer to Figure 1 , which shows a structural block diagram of a real-time magnetic resonance navigation system for neurosurgery provided by an embodiment of the present invention. The system includes: an image acquisition module 101, a dynamic area recognition module 102, a quality adjustment module 103, and a navigation execution module 104.

[0021] The image acquisition module 101 is used to acquire intraoperative continuous frame MRI images.

[0022] In the existing real-time magnetic resonance navigation system in neurosurgery, mainly by using MRI technology to perform scans before surgery, and then establishing a technical plan for neurosurgery according to the scan results, and setting navigation parameters in the navigation model navigation parameters according to the technical plan for surgery, and using MRI technology to perform real-time scans, and integrating real-time MRI images through positioning and tracking for magnetic resonance navigation of surgery. For the basic operation process, please refer to Figure 2 , which shows a basic operation flow chart of real-time navigation in an existing neurosurgery provided by an embodiment of the present invention.

[0023] Since the quality of the image transmission update after scanning will affect the accuracy of subsequent navigation planning, usually providing overall high-quality images for navigation, but the transmission of overall high-quality images will generate a large computational gap and affect the timeliness. To balance real-time calculation and quality, when further analyzing during tool positioning and tracking after real-time scanning, through dynamic adjustment of regional positioning, local high-quality updates are achieved, reducing the balance problem between image quality and real-time performance caused by global high quality.

[0024] In an embodiment of the present invention, during a neurosurgical operation, the magnetic resonance scanner needs to perform two processes of scanning on the patient. The first scan is the pre-operative MRI scan process, the purpose of which is to obtain the comprehensive condition of the patient to be operated on for formulating a treatment plan. Therefore, global high-quality images are required for information collection to judge the diseased focus area of the condition. Please refer to Figure 3 , which shows a schematic diagram of a global high-quality MRI image provided by an embodiment of the present invention.

[0025] And further, during the operation, continuous-frame MRI images are collected to provide an analysis basis for real-time feedback. In neurosurgery, a magnetic resonance wireless coil is used to enhance the image quality and clarity of the original MRI images, improving the accuracy of subsequent image analysis. It should be noted that the frequency of collecting MRI images during the patient's operation can be adjusted independently. In the embodiment of the present invention, the sampling frequency can be set to 2HZ, and the continuous-frame images within the first 2 seconds before the current moment are used as auxiliary analysis images.

[0026] The dynamic region recognition module 102 is used to determine the instrument connected domain of each frame of MRI image according to the boundary rule distribution of the connected domains in each frame of MRI image; determine the operation position of each frame of MRI image by analyzing the endpoint movement degree of the instrument connected domain in the continuous-frame MRI images; determine the traction influence region of the operation position according to the direction distribution of the edges within the local range at the operation position in each frame of MRI image; obtain the intraoperative dynamic region of each frame of MRI image through the proximity degree between the continuous-frame MRI images before each frame of MRI image at the operation positions.

[0027] In the patient's MRI image, there are operation regions and non-operation regions. Since the actual operation region does not involve the entire scanned surgical local region, it is necessary to distinguish and determine. First, track the tool part during the operation. Compared with the patient's local tissue, the surgical tool has, first of all, the uniformity of texture and the regularity of shape. Therefore, this feature is used to track the surgical tool in the MRI image. Please refer to Figure 4 , which shows a schematic diagram of a partial surgical tool provided by an embodiment of the present invention.

[0028] The possibility that the region is an instrument tool is reflected by the morphological regularity and uniformity of the region in the MRI image, so as to facilitate the subsequent determination of the operation position. First, obtain the connected domain for each frame of MRI image. It should be noted that the method for obtaining the connected domain is a well-known technical means for those skilled in the art and will not be elaborated here. Please refer to Figure 5 , which shows a schematic diagram of a single connected domain provided by an embodiment of the present invention.

[0029] After that, the connected regions where the instrument exists are initially determined according to the boundaries. Preferably, in the embodiments of the present invention, the method for obtaining the connected regions of the instrument includes: First, for any connected region in each frame of MRI image, taking any pixel point on the boundary of the connected region as the starting point, the pixel points continuously distributed clockwise along the boundary are formed into a sequence, which is used as the boundary sequence of the connected region. The shape complexity of the connected region is reflected by the changes between consecutive points on the boundary.

[0030] Compared with surgical tools, brain tissues have a more complex shape. Therefore, the complexity of direction turning is much higher than that of surgical tool parts. Therefore, the direction changes of adjacent pixel points can be used to distinguish surgical tools from brain tissues. In the boundary sequence, when the gradient directions between each pixel point and the next pixel point are different, it indicates that a direction turn occurs. The corresponding pixel point is used as the turning pixel point of the connected region. The number of turning pixel points is counted and normalized to obtain the turning degree of the connected region. The more the number of turning pixel points, the higher the boundary complexity, and the more likely it is to be a tissue region.

[0031] It should be noted that normalization is a well-known technical means to those skilled in the art. The choice of normalization can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0032] Finally, the connected region with the minimum turning degree in each frame of MRI image is used as the connected region of the instrument, which reflects the surgical tool part. The surgical tool part can be used for positioning the area to be operated on. Generally, tools have a certain length, and the action point of the tool is directed at human tissues. Therefore, on the basis of tool tracking, the area to be operated on can be obtained through the moment of inertia of the tool. The end with a larger moment of inertia is the end far from the area to be operated on, and the end with a smaller moment of inertia is the end close to the area to be operated on. Please refer to Figure 6 which shows a side schematic view of the moment of inertia of a surgical tool provided by an embodiment of the present invention.

[0033] The position to be operated on is determined through the movement of the endpoint part. In the embodiments of the present invention, the method for obtaining the position to be operated on includes: First, the connected regions of the instrument in each frame of MRI image are linearly fitted to obtain the instrument line of each frame of MRI image. In the embodiments of the present invention, the linear fitting of the connected region is performed by the least squares method, and Hough transform, etc. can also be selected. The process of linear fitting is a well-known technical means to those skilled in the art and will not be elaborated and limited here.

[0034] For any frame of MRI image, one end point of the instrument straight line in the MRI image is sequentially used as an analysis point, the displacements between the analysis points of the instrument straight lines in each adjacent frame before the MRI image are calculated, and the mean value is calculated to obtain the moment of inertia value of the MRI image at the analysis point. The moment of inertia situation is reflected by the displacement distance changing between consecutive frames. When the moment of inertia is smaller, it indicates that it is more likely to be a point closer to the tissue site.

[0035] The displacement of the tool at the treatment point position is usually small, and the displacement of the tool at the non-treatment point position is usually long. Therefore, the end point with the smallest moment of inertia value in the MRI image is used as the treatment position of the MRI image, and the true operation point is locked based on the end point displacement difference.

[0036] The traction effect refers to the process of using a tool during the operation. When using the tool for the operation, because the tissue is locally connected and the nerve-related tissues of the human body are generally relatively soft, and there is only one treatment point under the action of the surgical tool. When the treatment point is acted on by the surgical tool, it often produces a pulling effect on the nerve tissues connected to the treatment point. These pulled areas need to be key attention areas during the operation. Please refer to Figure 7 , which shows a schematic diagram of the traction situation of a treatment position provided by an embodiment of the present invention.

[0037] The tissues pulled by the treatment point obviously have edge bulges and highly similar directions. Therefore, the affected area is determined through the edge distribution in the image. Preferably, in the embodiment of the present invention, the method for determining the traction influence area of the treatment position includes: First, for the treatment position of any MRI image, within a preset local range at the treatment position, the existing edges are recorded as the pulled edges of the treatment position to obtain the local existing edge situation. The edges may be generated due to tissue pulling. In the embodiment of the present invention, the preset local range is set to a circular range centered on the treatment position with a radius of 6. The complete edge line determined to have edges through edge detection within the local range is used as the traction edge, and the specific numerical value can be adjusted by the implementer himself.

[0038] Furthermore, the extension direction of each pulled edge at the treatment position is obtained, and analysis is performed based on the characteristics of highly similar directions. In the embodiment of the present invention, each pulled edge is linearly fitted, and the direction of the fitted straight line is used as the extension direction of each pulled edge, reflecting the main direction vector of the edge distribution.

[0039] Furthermore, for any one of the involved edges of the operative position, calculate the average similarity between the involved edge and the adjacent involved edges in the extension direction to obtain the involvement significance of the involved edge. In the embodiments of the present invention, the similarity between the extension directions can be calculated by cosine similarity. If the direction similarity is high, it indicates that both this edge and the adjacent edges are likely to be parts of the regions where involvement occurs.

[0040] Therefore, when the involvement significance is greater than a preset significance threshold, it is considered that there is an edge affected by involvement. The corresponding involved edge is used as the affected edge, and the connected regions where the image edges are located at the operative position are merged to obtain the traction influence region at the operative position. The overall part affected by the traction effect is determined by the region covered by the existing affected edges. In the embodiments of the present invention, the preset significance threshold is set to 0.85, and its value can be adjusted by the implementer himself.

[0041] In neurosurgery, only focusing on the operative point and the traction region is very limited. It is often necessary to perform auxiliary analysis of the changes in adjacent consecutive frames to more comprehensively control the real-time state. Therefore, it is necessary to obtain a more complete dynamic region through the operative point and the traction region in time series to ensure the quality of the images around the operative point, and further ensure the efficiency of intraoperative navigation.

[0042] Preferably, in the embodiments of the present invention, the method for obtaining the intraoperative dynamic region includes: First, for any frame of MRI image, sequentially use each frame of MRI image before this MRI image as the analysis image, and analyze each frame of image in the time series before. Calculate the Euclidean distance between the operative positions of this MRI image and the analysis image as the proximity parameter of the analysis image. The closer the positions are, the greater the possibility that the operative positions of the images at this moment involve the same nerve tissue.

[0043] Furthermore, use the analysis images with proximity parameters less than the preset proximity threshold as the approximate images of this MRI image. The approximate images reflect other influence ranges of this region that the operative point position may involve when performing surgery in this region. In the embodiments of the present invention, the preset proximity threshold is set to 3, and the specific value can be adjusted by the implementer according to the specific implementation situation and is not limited here.

[0044] Therefore, superimpose the traction influence region in the approximate images and the traction influence region of this MRI image to obtain the intraoperative dynamic region of this MRI image. By superimposing the regions of multiple operative positions that are close to the moment of this MRI image, it reflects the historical overall traction influence range that exists when performing surgery near this operative position, and the quality of the region needs to be ensured in subsequent considerations.

[0045] The quality adjustment module 103 is configured to determine the gradient attenuation quality of the connected regions in each frame of the MRI image according to the positional distribution of the intraoperative dynamic region and other connected regions at the surgical position; and determine the transmission quality of each connected region at the current moment according to the gradient attenuation quality of each connected region except the instrument connected region in the MRI image at the current moment.

[0046] The gradient quality degradation radiates around based on the region where traction exists, and the image quality gradually degrades during the radiation process. The high quality of the affected dynamic region is ensured, and the quality degrades faster when the distance from the intraoperative dynamic region is farther. Therefore, the degradation situation of each frame of MRI image is analyzed first.

[0047] Preferably, in the embodiment of the present invention, the method for obtaining the gradient attenuation quality includes: For any frame of the MRI image, a circle is drawn with the surgical position in the MRI image as the center and the maximum distance from the center to the boundary of the intraoperative dynamic region as the radius to determine the boundary of the highest quality range of the MRI image. This range includes the overall range of the connected regions involved in the operation, and high-quality transmission needs to be retained to ensure the accuracy of subsequent navigation. The quality coefficient within the boundary of the highest quality range is the preset maximum quality coefficient. In the embodiment of the present invention, the preset maximum quality coefficient is 1, and the image quality is fully retained.

[0048] Further, calculate the distance between the surgical position in the MRI image and the boundary of each other connected region, and draw a circle with the length of the farthest distance from the boundary of the other connected region as the radius to determine the boundary of the maximum loss quality range of the MRI image. This range is the farthest range where the quality of the adjustable connected region in the image can be adjusted, that is, the boundary of the maximum range of quality loss. The quality coefficient on the boundary of the maximum loss quality range is the preset minimum quality coefficient. In the embodiment of the present invention, the preset minimum quality coefficient is 0.5, and only half of the image quality needs to be retained to reduce the data display time during transmission. Please refer to Figure 8 , which shows a schematic diagram of the boundary of the highest quality range and the boundary of the maximum loss quality range provided by an embodiment of the present invention.

[0049] Between the boundary ranges, the quality coefficient decreases gradually. The quality coefficient is attenuated between the boundary of the highest quality range and the boundary of the maximum loss quality range to obtain the quality coefficient at each position. In the embodiment of the present invention, the method for obtaining the quality coefficient includes: The shortest line connecting the highest quality range boundary to the maximum loss quality range boundary is taken as the gradient descent line. Starting from the quality coefficient at the boundary of the highest quality region, along the gradient descent line, the quality coefficient is attenuated to the quality coefficient on the boundary of the maximum loss quality range using an attenuation function, and the quality coefficient at each position on the gradient descent line is determined. In the embodiments of the present invention, the selected attenuation function is a negative exponential function, the starting quality coefficient is 1, and the ending quality coefficient is 0.5. As the length from the starting point increases, the quality coefficient exponentially decays to 0.5. The distance length between each position on the gradient descent line and the treated position is used as the radius to form the corresponding exponential coefficient on the circular boundary.

[0050] Since the tissue region needs to ensure the same quality, further through combination, the mean value of the quality coefficients existing in each connected domain in the MRI image is calculated as the gradient attenuation quality of each connected domain, maintaining the tissue integrity criterion.

[0051] By analyzing the possible quality attenuation situations of all regions before the current moment, the quality retention situation of each connected domain at the final current moment is further determined. In the embodiments of the present invention, the method for obtaining the transmission quality includes: In the MRI image at the current moment, the maximum gradient attenuation quality of each connected domain (excluding the instrument connected domain) in all other MRI images is taken as the transmission quality of each connected domain at the current moment. Through analysis of different frames involved, the maximum required gradient attenuation quality is taken as the transmission quality for current adjustment to ensure the integrity of dynamic information.

[0052] The navigation execution module 104 is used to update the MRI image based on the transmission quality of different connected domains at the current moment to obtain the intraoperative transmission image and acquire the navigation path.

[0053] After determining the quality conditions of different regions in the image at the current moment, the real-time scanned image is transmitted and updated and then output. In the embodiments of the present invention, the MRI image at the current moment and the transmission quality situation are used as inputs, and through the trained image transmission model, the intraoperative transmission image is output for subsequent navigation planning. In the embodiments of the present invention, the image transmission model adjusts the local image resolution of the image based on the transmission quality of different regions, retaining more regions with higher transmission quality. Model training is a well-known technical means for those skilled in the art and will not be elaborated here.

[0054] In the embodiments of the present invention, first, the preoperatively obtained MRI image is registered with the intraoperative transmission image to ensure that the intraoperative transmission image is consistent with the actual position of the head. Secondly, tool tracking is performed through the real-time transmitted intraoperative transmission image, and then the route marking is completed using the neurosurgery implementation plan established based on the preoperative MRI image. Then, positioning is performed using the tracked surgical tool according to the marked route, and the navigation path is displayed to complete the navigation.

[0055] The present invention determines the surgical tool part based on the boundary regular distribution of the connected components and the movement of consecutive frames, and then tracks and locates the position of the operation site. Subsequently, it is convenient to analyze the operation area at the operation site to ensure high-quality development of local images. Further, it is ensured that the areas affected by tissue traction during the operation can also be observed and analyzed. According to the local deformation edge generated at the operation site during the operation, the range of the traction area of the operation is determined based on the distribution of the local tissue edge at the operation site, and considering the operation range situation involved many times in the previous steps, the intraoperative dynamic area affected during the operation is determined according to the proximity of the distribution of the operation site, making the key attention area in each frame of image more accurate. Then, a gradient attenuation analysis is performed on the distribution of the intraoperative dynamic area of key attention in each frame of image and the remaining areas to obtain the attenuation quality situation, and considering the integrity and consistency of the regional tissue during the operation, the transmission quality of the final connected area is determined based on the gradient attenuation quality situation in different image frames before the current moment for transmission display navigation. The present invention locates the intraoperative affected area after tracking the instrument area, dynamically adjusts the quality of different areas, and balances the transmission quality and real-time issues.

[0056] It should be noted that the above sequence of embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0057] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments.

Claims

1. A real-time magnetic resonance navigation system for neurosurgery, characterized in that, The system includes: An image acquisition module for acquiring intraoperative continuous-frame MRI images; A dynamic region recognition module for determining the instrument connected region of each frame of MRI image according to the boundary regular distribution of the connected regions in each frame of MRI image; determining the operative position of each frame of MRI image by analyzing the endpoint movement degree of the instrument connected region in the continuous-frame MRI images; Determining the traction influence region of the operative position according to the direction distribution of the edges in the local range at the operative position of each frame of MRI image; obtaining the intraoperative dynamic region of each frame of MRI image through the proximity degree between the continuous-frame MRI images before each frame of MRI image at the operative positions; A quality adjustment module for determining the gradient attenuation quality of the connected regions in each frame of MRI image according to the position distribution of the intraoperative dynamic region of the operative position and other connected regions in each frame of MRI image; determining the transmission quality of each connected region at the current moment according to the gradient attenuation quality of each connected region except the instrument connected region in the MRI images at different times; A navigation execution module for updating the MRI image based on the transmission quality of different connected regions at the current moment to obtain an intraoperative transmission image and acquire a navigation path.

2. The real-time magnetic resonance navigation system for neurosurgery according to claim 1, wherein The method for obtaining the instrument connected region includes: For any connected region in each frame of MRI image, taking any pixel point on the boundary of the connected region as a starting point, and forming a sequence of pixel points continuously distributed clockwise along the boundary as the boundary sequence of the connected region; In the boundary sequence, when the gradient direction between each pixel point and the next pixel point is different, taking the corresponding pixel point as the turning pixel point of the connected region; counting the number of turning pixel points and performing normalization processing to obtain the turning degree of the connected region; Taking the connected region with the smallest turning degree in each frame of MRI image as the instrument connected region.

3. The real-time magnetic resonance navigation system for neurosurgery according to claim 1, characterized in that The method for obtaining the operative position includes: Performing linear fitting on the instrument connected regions in each frame of MRI image to obtain the instrument line of each frame of MRI image; For any frame of MRI image, successively taking one endpoint of the instrument line in this MRI image as an analysis point; calculating the displacement between the instrument lines of adjacent frames of MRI before this MRI image at the analysis points and calculating the mean value to obtain the moment of inertia value of this MRI image at the analysis point; Taking the endpoint with the smallest moment of inertia value in this MRI image as the operative position of this MRI image.

4. The real-time magnetic resonance navigation system for neurosurgery according to claim 1, characterized in that, The method for obtaining the traction influence region includes: For the operative position of any MRI image, within the preset local range at the operative position, marking the existing edges as the involved edges of the operative position; obtaining the extension direction of each involved edge of the operative position; For any involved edge of the operative position, calculating the similarity mean value between the involved edge and the adjacent involved edges in the extension direction to obtain the involvement significance of the involved edge; When the involvement significance is greater than the preset significance threshold, taking the corresponding involved edge as an influencing edge; merging the connected regions where the image edges are located at the operative position to obtain the traction influence region of the operative position.

5. The real-time magnetic resonance navigation system for neurosurgery according to claim 1, characterized in that, The method for obtaining the intraoperative dynamic region includes: For any frame of MRI image, each frame of MRI image before this MRI image is sequentially used as the analysis image; Calculate the Euclidean distance between the operation positions of this MRI image and the analysis image as the proximity parameter of the analysis image; The analysis image with the proximity parameter less than the preset proximity threshold is used as the approximate image of this MRI image; Overlay the traction influence area in the approximate image with the traction influence area of this MRI image to obtain the intraoperative dynamic area of this MRI image.

6. The real-time magnetic resonance navigation system for neurosurgery according to claim 1, wherein The method for obtaining the gradient attenuation quality includes: For any frame of MRI image, take the operation position in this MRI image as the center of the circle, and use the farthest distance from the center of the circle to the boundary of the intraoperative dynamic area as the radius to draw a circle to determine the boundary of the highest quality range of this MRI image; The quality coefficient within the boundary of the highest quality range is the preset maximum quality coefficient; Calculate the distance between the operation position of this MRI image and the boundary of each other connected domain, and use the length farthest from the boundary of the other connected domain as the radius to draw a circle to determine the boundary of the maximum loss quality range of this MRI image; The quality coefficient on the boundary of the maximum loss quality range is the preset minimum quality coefficient; Decay the quality coefficient between the boundary of the highest quality range and the boundary of the maximum loss quality range to obtain the quality coefficient at each position; Calculate the mean value of the quality coefficients existing in each connected domain in this MRI image as the gradient attenuation quality of each connected domain.

7. The real-time magnetic resonance navigation system for neurosurgery according to claim 6, characterized in that The decaying the quality coefficient between the boundary of the highest quality range and the boundary of the maximum loss quality range to obtain the quality coefficient at each position includes: Take the shortest connection line from the boundary of the highest quality range to the boundary of the maximum loss quality range as the gradient descent line; Starting from the quality coefficient at the boundary of the highest quality area, use the attenuation function to decay the quality coefficient along the gradient descent line to the quality coefficient on the boundary of the maximum loss quality range to determine the quality coefficient at each position on the gradient descent line.

8. The real-time magnetic resonance navigation system for neurosurgery according to claim 1, wherein The method for obtaining the transmission quality includes: In the MRI image at the current moment, take the maximum gradient attenuation quality of each connected domain except the instrument connected domain in all other MRI images as the transmission quality of each connected domain at the current moment.

9. The real-time magnetic resonance navigation system for neurosurgery according to claim 1, characterized in that The method for obtaining the intraoperative transmission image includes: Take the MR image and the transmission quality situation at the current moment as the input, and output the intraoperative transmission image through the trained image transmission model.

10. The real-time magnetic resonance navigation system for neurosurgery according to claim 4, characterized in that, The method for obtaining the extension direction of each traction edge at the operation position includes: Perform linear fitting on each traction edge, and use the direction of the fitting line as the extension direction of each traction edge.

Images