Adjacent region sketching auxiliary method before hepatic segment resection

Abstract

The invention relates to the technical field of image feature recognition, in particular to an auxiliary method for sketching adjacent regions before hepatic segment resection. According to the method, a CT image under each layer of the liver is analyzed, and the respiration influence degree is quantified by using the shape change and area change of a blood vessel region. According to the method, matching analysis is carried out on target blood vessel areas between adjacent layers, then change trend characteristics of respiration influence degrees are determined, all target blood vessel areas are traversed in a mode of matching and then combining, and an optimal combination mode for representing portal veins can be determined based on the change trend characteristics; therefore, the blood vessel region represented by the portal vein is screened out, and the effective assistance of the sketching of the adjacent region can be realized through three-dimensional reconstruction. According to the method, the accurate portal vein region is screened out and three-dimensional reconstruction is carried out through correlation analysis on the blood vessel regions in the CT images of different layers, and sketching of adjacent regions can be effectively assisted.

Description

Technical Field

[0001] The present invention relates to the technical field of image feature recognition, and particularly to an adjacent region delineation assistance method before hepatectomy segmentectomy. Background Art

[0002] The delineation of adjacent regions before hepatectomy segmentectomy is a core link in surgical planning, and its technical development has experienced an evolution from traditional two-dimensional image analysis to three-dimensional visualization and intelligent segmentation. Currently, the delineation of adjacent regions before hepatectomy segmentectomy can be combined with comprehensive imaging techniques, intelligent algorithms, and intraoperative dynamic navigation. The prior art can utilize liver static CT image segmentation technology to remove the influence of adjacent organs such as the spine and ribs, observe the relationship between tumors and surrounding blood vessels, plan the resection range, and improve the segmentation accuracy; however, since each hepatic segment has its own blood inflow, outflow, and bile drainage. At the center of each hepatic segment, there are branches of the portal vein, hepatic artery, and bile duct. Around each hepatic segment, blood vessels flow out through the hepatic vein. Therefore, the hepatic veins and portal veins of different hepatic segments are prone to varying degrees of deformation under respiratory movement and intraoperative traction, confusing the boundaries of adjacent regions. If the types of blood vessel regions on CT images cannot be effectively distinguished and segmented, the delineation results will be inaccurate. Summary of the Invention

[0003] In order to solve the technical problem that the prior art cannot effectively segment portal vein information based on liver CT images, resulting in inaccurate delineation of adjacent regions, the purpose of the present invention is to provide an adjacent region delineation assistance method before hepatectomy segmentectomy, and the specific technical solution adopted is as follows: The present invention proposes an adjacent region delineation assistance method before hepatectomy segmentectomy, and the method includes: Obtaining consecutive multiple frames of CT images of the patient's liver at each level under CT layer scanning, where the blood vessel region in the first frame of CT image is used as the target blood vessel region; Obtaining the fusiform feature degree of each blood vessel region; at each level, sequentially matching the blood vessel regions in adjacent frames of CT images, and obtaining the respiratory influence degree of the target blood vessel region according to the area change and fusiform feature degree change between the matching regions; Matching the target blood vessel regions with the same hepatic segment between adjacent levels, and the target blood vessel regions with a matching relationship form a set of matching target blood vessel regions; in the set of matching target blood vessel regions, obtaining the mutual extension probability of the set of matching target blood vessel regions according to the distance and position distribution between regions; For each liver segment, traverse and combine all the sets of matching target vascular regions included, arrange them according to the mutual extension probability, and obtain multiple combination methods; obtain the change trend characteristics of the respiratory influence degree of the set of matching target vascular regions in each combination method; according to the change trend characteristics, the number of layers included in the set of matching target vascular regions, the fusiform feature degree of the target vascular region corresponding to the set of matching target vascular regions, and the number of elements in the combination method, screen out the optimal combination method; Take the target vascular region corresponding to the optimal combination method as the portal vein branch and perform three-dimensional reconstruction for assisting in the delineation of adjacent regions.

[0004] Further, the method for obtaining the fusiform feature degree includes: Obtain the minimum circumscribed rectangle of the vascular region, take the aspect ratio of the minimum circumscribed rectangle as the initial fusiform feature degree, and take the product of the average gray value of the vascular region and the initial fusiform feature degree as the fusiform feature degree.

[0005] Further, sequentially match the vascular regions in adjacent-frame CT images, including: For each pair of adjacent frames at each level, obtain the degree of coincidence between the already matched vascular regions in the previous-frame CT image and the vascular regions in the next-frame CT image, and select the vascular region with the highest degree of coincidence as the new already matched vascular region; among them, the already matched vascular region in the first-frame CT image is any one of the target vascular regions; Match the target vascular regions with the same liver segment between two adjacent levels, including: For any two adjacent levels, match the first-frame CT images between the two levels, and select a set of target vascular regions with the largest degree of coincidence as the pair of matching target vascular regions with a matching relationship; traverse all levels, and form a set of matching target vascular regions by including the target vascular regions included in the pairs of matching target vascular regions with a mutual matching relationship.

[0006] Further, the method for obtaining the respiratory influence degree includes: For each group of adjacent frames, obtain the difference in the average fusiform feature degree and the difference in the average area between the already matched vascular regions in the previous frame and the already matched vascular regions in the next frame, and obtain the morphological distortion degree; count the morphological distortion degrees of all groups of adjacent frames, and obtain the respiratory influence degree according to the average morphological distortion degree and the variance of the morphological distortion degree.

[0007] Further, the method for obtaining the morphological distortion degree includes: For each group of adjacent frames, obtain a first ratio of the average fusiform feature degree between the matched blood vessel regions of the previous frame and the matched blood vessel regions of the next frame, and a second ratio of the average area; obtain a first product between the first ratio and the second ratio, and normalize the difference between the first product and the positive integer 1 to obtain the morphological distortion degree.

[0008] Further, the method for obtaining the mutual extension probability includes: In the set of matched target blood vessel regions, for two matched target blood vessel regions corresponding to adjacent layers, obtain the minimum circumscribed rectangles of the two matched target blood vessel regions, extend the side lengths of the two minimum circumscribed rectangles, and use the minimum included angle formed by the extension lines as the pose difference angle; obtain the distance between the two matched target blood vessel regions, calculate the Euclidean norm based on the distance and the pose difference angle, and perform a negative correlation mapping on the Euclidean norm to obtain the initial mutual extension probability between the two matched target blood vessel regions; In the set of matched target blood vessel regions, count the initial mutual extension probabilities of each pair of matched target blood vessel regions corresponding to all adjacent layers, and use the average initial mutual extension probability as the mutual extension probability of the set of matched target blood vessel regions.

[0009] Further, the method for obtaining the change trend feature includes: Obtain the average degree of respiratory influence of each set of matched target blood vessel regions; Arrange the sets of matched target blood vessel regions in each combination method based on the mutual extension probability in descending order, and the average degrees of respiratory influence of the sets of matched target blood vessel regions form a sequence of degrees of respiratory influence; In the sequence of degrees of respiratory influence, obtain the element difference obtained by subtracting the previous element from the next element; use the cumulative value of the element differences as the numerator and the cumulative value of the absolute values of the element differences as the denominator to obtain the change trend feature of each combination method.

[0010] Further, the method for screening the optimal combination method includes: According to the change trend feature, the number of layers included in the set of matched target blood vessel regions, the fusiform feature degree of the target blood vessel regions corresponding to the set of matched target blood vessel regions, and the number of elements in the combination method, obtain the venous branching possibility of each combination method, and use the combination method with the largest venous branching possibility as the optimal combination method.

[0011] Further, the method for obtaining the venous branching possibility includes: For each combination method, obtain the maximum fusiform feature degree of the target blood vessel regions that match the target blood vessel region set; use the product of the maximum fusiform feature degree and the number of layers as the venous branch feature of the target blood vessel region set that matches; multiply the cumulative value of the venous branch features in the combination method by the change trend feature, and subtract the number of the target blood vessel region set that matches in the combination method from the product to obtain the venous branch possibility.

[0012] Further, the method for obtaining the blood vessel region includes: Perform edge detection on each frame of CT image, and use the closed region formed by the edges as the blood vessel region.

[0013] The present invention has the following beneficial effects: The present invention analyzes the CT images of each layer of the liver, considering that the patient's liver will produce respiratory state characteristics due to blood flow and physiological activities under normal conditions. The portal vein has relatively weak respiratory characteristics compared to the hepatic vein, and as the layer changes, the degree of respiratory influence of the portal vein will change with the position of the region. Therefore, the present invention quantifies the degree of respiratory influence by using the shape change and area change of the blood vessel region. By performing matching analysis on the target blood vessel regions between adjacent layers, the change trend feature of the degree of respiratory influence is further determined. The present invention traverses all target blood vessel regions in a way of matching and then combining, and based on the change trend feature, the optimal combination method representing the portal vein can be determined, and then the blood vessel regions represented by the portal vein can be screened out. Effective assistance for the delineation of adjacent regions can be achieved through three-dimensional reconstruction. Description of the Drawings

[0014] 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 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.

[0015] Figure 1 It is a flowchart of an adjacent region delineation assistance method before hepatectomy provided by an embodiment of the present invention. Detailed Embodiments

[0016] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following specifically describes, in conjunction with the accompanying drawings and preferred embodiments, a method for assisting in delineating adjacent regions before liver segment resection according to the present invention, including its specific implementation manner, structure, features, and effects. 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 may be combined in any suitable form.

[0017] 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.

[0018] The following specifically describes the specific solution of a method for assisting in delineating adjacent regions before liver segment resection provided by the present invention in conjunction with the accompanying drawings.

[0019] Please refer to Figure 1 , which shows a flowchart of a method for assisting in delineating adjacent regions before liver segment resection provided by an embodiment of the present invention. The method includes: Step S1: Obtain a series of consecutive CT images of each layer of the patient's liver under CT axial scanning, where the vascular region in the first CT image is used as the target vascular region.

[0020] The embodiment of the present invention is directed to the three-dimensional reconstruction scenario of the hepatic portal vein of the liver that has completed CT axial scanning. Through CT axial scanning of multiple layers, the state characteristics of each position of the liver can be identified. Conventional axial scanning has information on four layers. In the embodiment of the present invention, a series of consecutive CT images are collected for each layer, and then the images are transmitted to a computer for image analysis and recognition. The gray value of the CT image is the HU value, and different substances such as air, fat, water, soft tissue, and blood have different HU value ranges. Therefore, the vascular region can be effectively determined in the CT image for analyzing the vascular type.

[0021] It should be noted that the intrahepatic ducts can be divided into two parts: the Glisson system and the hepatic vein system. The branches and distributions of these two systems can divide the liver into 8 functionally independent liver segments, and each liver segment has its own blood vessel inflow, outflow, and bile drainage. At the center of each liver segment, there are branches of the portal vein, hepatic artery, and bile duct. There are blood vessels flowing out through the hepatic vein around each liver segment. That is, the information of multiple liver segments is included in a single CT image.

[0022] In each hepatic segment, the portal vein branches enter the core area of each segment, forming the main trunk of blood supply for that segment. Its branching pattern is relatively constant with fewer variations. For example, the right branch mainly supplies the right lobe, and the left branch covers areas such as the left lateral lobe, left medial lobe, and papillary process. This central distribution makes the portal vein the "core landmark" for hepatic segment division, and its branching path forms potential longitudinal fissures between hepatic segments in three-dimensional space. However, the portal vein branches may have "interdigitating" intersections with adjacent hepatic segments, resulting in an irregular actual boundary. Hepatic veins (such as the right, middle, and left hepatic veins) mainly run within the longitudinal fissures between hepatic segments, draining blood from multiple adjacent segments. For example, the right hepatic vein separates the right anterior lobe from the right posterior lobe, and the middle hepatic vein divides the left and right half livers. Since hepatic veins have more variations (such as double right hepatic veins accounting for 22 - 36%), their reliability as demarcation markers is relatively low. In addition, the hepatic veins form an acute angle with the inferior vena cava, further affecting their demarcation position in three-dimensional space. Therefore, it is necessary to distinguish the characteristic manifestations of portal vein branches and hepatic veins in the preliminarily divided hepatic segments to construct a clearer three-dimensional image of hepatic segment vessels.

[0023] For blood vessels, during CT image acquisition, techniques such as contrast agents can be used to make the blood vessel areas in CT images have a significant brightness difference from the hepatic parenchyma areas. Therefore, all blood vessel areas can be directly determined in each CT image. It should be noted that since each layer in the embodiments of the present invention contains a series of consecutive CT images, and the consecutive CT images are used to analyze the degree of respiratory influence on a certain blood vessel area in the subsequent process, the first CT image is selected as the reference image, and the blood vessel area therein is used as the target blood vessel area, and each target blood vessel area is used as the object for analysis in the subsequent process.

[0024] Preferably, in an embodiment of the present invention, since the use of a blood vessel contrast agent enhances the blood vessel characteristics, edge detection can be directly performed on each frame of CT image, and the closed area formed by the edges is used as the blood vessel area.

[0025] Step S2: Obtain the fusiform feature degree of each blood vessel area; at each layer, the blood vessel areas in adjacent frames of CT images are sequentially matched, and the degree of respiratory influence on the target blood vessel area is obtained based on the area change and fusiform feature degree change between the matching areas.

[0026] The hepatic veins and portal veins in different liver segments will undergo varying degrees of deformation under respiratory movement and intraoperative traction. Therefore, in a series of consecutive CT images at a single level, as the respiratory movement occurs, the morphology of the vascular region will change. The hepatic veins are usually located in the adjacent areas of the liver segments and are more significantly affected by respiration, resulting in a greater degree of deformation. In contrast, the deformation of the portal vein caused by respiration is relatively small. Therefore, the type characteristics can be characterized by analyzing the degree of respiratory influence on the vascular region. Further considering that when analyzing the morphology of the vascular region, one part is the area change and the other part should be the shape change. The vascular region has an obvious spindle-shaped feature in the CT image. Therefore, in the embodiments of the present invention, the spindle-shaped feature of each vascular region is first quantified, and then the degree of change in the spindle-shaped feature and the area change among a series of consecutive CT images at a single level are further analyzed. Starting from the first CT image, adjacent CT images are sequentially matched, and based on the change characteristics of the matched regions, the degree of respiratory influence on each target vascular region can be determined.

[0027] Preferably, in the embodiments of the present invention, considering that the vascular region has a relatively obvious gray value due to the influence of the contrast agent, and since the vascular region is a closed region at the edge of the CT image, there may be non-vascular regions misidentified as vascular regions. Therefore, the average gray value of the vascular region can be introduced during the process of analyzing the degree of the spindle-shaped feature. The minimum bounding rectangle of the vascular region is obtained, and the aspect ratio of the minimum bounding rectangle is used as the initial degree of the spindle-shaped feature. The product of the average gray value of the vascular region and the initial degree of the spindle-shaped feature is used as the degree of the spindle-shaped feature. That is, the greater the degree of the spindle-shaped feature, the more the vascular region conforms to the spindle-shaped feature, and the greater the average gray value, the greater the possibility of it being a hepatic vein.

[0028] In an embodiment of the present invention, since the degree of the spindle-shaped feature includes both the shape and gray value of the region, the vascular region can be screened based on the degree of the spindle-shaped feature to filter out the misidentified non-vascular regions. After obtaining the screened vascular region, the analysis of the degree of respiratory influence continues. The specific screening process can be set by methods such as threshold screening, which can be selected by those skilled in the art and is not limited herein.

[0029] Preferably, in an embodiment of the present invention, the vascular regions in adjacent CT images are sequentially matched, including: For each pair of adjacent frames at each level, obtain the degree of overlap between the matched blood vessel regions in the previous-frame CT image and the blood vessel regions in the next-frame CT image, and select the blood vessel region with the highest degree of overlap as the new matched blood vessel region; wherein the matched blood vessel region in the first-frame CT image is any one of the target blood vessel regions. For example, starting from a pair of adjacent frames composed of the first frame and the second frame, select a certain target blood vessel region on the first frame as the matched blood vessel region, determine the matched blood vessel region of this matched blood vessel region on the second frame, further analyze in the adjacent frames composed of the second frame and the third frame, determine the matched blood vessel regions of each matched blood vessel region of the second frame on the third frame, and then determine all the matched blood vessel regions and the matching relationships between these regions.

[0030] Further, after determining the matching relationships of the blood vessel regions between adjacent-frame CT images, the matching relationships corresponding to one target blood vessel region can successively form a set of region sets, and each set of region sets corresponds to one target blood vessel region. Then, in this set of region sets, for each pair of adjacent frames, obtain the difference in the average fusiform feature degree and the difference in the average area between the matched blood vessel region of the previous frame and the matched blood vessel region of the next frame, and obtain the morphological distortion degree. Further, statistically analyze the morphological distortion degrees of all sets of adjacent frames, and obtain the degree of respiratory influence based on the average morphological distortion degree and the variance of the morphological distortion degree. That is, the larger the average morphological distortion degree and the larger the variance of the morphological distortion degree, the stronger the respiratory deformation characteristics of the target blood vessel region at this level, the greater the degree of respiratory influence, and the more likely the target blood vessel region is the hepatic vein region and the less likely it is the portal vein region.

[0031] In the embodiments of the present invention, since both the average morphological distortion degree and the variance of the morphological distortion degree are positively correlated with the degree of respiratory influence, the product of the two can be directly used as the degree of respiratory influence.

[0032] Further, in an embodiment of the present invention, the morphological distortion degree is quantified. For each pair of adjacent frames, obtain the first ratio of the average fusiform feature degree between the matched blood vessel region of the previous frame and the matched blood vessel region of the next frame, and the second ratio of the average area; obtain the first product between the first ratio and the second ratio. The closer the first product is to 1, the closer the first ratio and the second ratio are to 1, indicating that the degree of distortion of the average fusiform feature degree and the average area between the two frames is smaller. Therefore, normalize the difference between the first product and the positive integer 1. The larger the difference, the greater the distortion between the two frames. Thus, the morphological distortion degree can be obtained. It should be noted that the normalization in the embodiments of the present invention can adopt the method of linear normalization, and those skilled in the art can select a suitable normalization method for processing by themselves, which will not be elaborated here.

[0033] After the processing of step S2, there is a corresponding degree of respiratory influence for each target vascular region under each layer.

[0034] Step S3: Match the target vascular regions with the same liver segment between adjacent layers. The target vascular regions with a matching relationship form a set of matching target vascular regions. In the set of matching target vascular regions, according to the distance and position distribution between regions, obtain the mutual extension probability of the set of matching target vascular regions.

[0035] The present invention aims to analyze the distribution of the portal vein in the liver at different layers, and then obtain an accurate and effective three-dimensional modeling result of the portal vein to assist in the delineation of adjacent regions. Therefore, it is also necessary to perform an association analysis on the target vascular regions between adjacent layers. Therefore, step S3 further matches the target vascular regions with the same liver segment between adjacent layers, and the target vascular regions with a matching relationship form a set of matching target vascular regions. That is, a set of matching target vascular regions is a set of target vascular regions at multiple layers. These target vascular regions can be regarded as local regions of a certain venous vessel. Therefore, ideally, a set of matching target vascular regions can correspond to a certain venous vessel. However, the successfully matched target vascular regions are not necessarily real regions of the same kind of venous vessel. It is necessary to further analyze their distance and position distribution, and quantify the mutual extension probability in the set of matching target vascular regions. That is, the greater the mutual extension probability, the more likely it is that the set of matching target vascular regions is a region of the same kind of venous vessel.

[0036] Preferably, in the embodiment of the present invention, similar to the matching of adjacent frame CT images in one layer, matching the target vascular regions with the same liver segment between two layers includes: For any two adjacent layers, match the first frame CT images between the two layers, and select a set of target vascular regions with the largest degree of coincidence as a pair of matching target vascular regions with a matching relationship; traverse all layers, and the target vascular regions included in the pairs of matching target vascular regions with a mutual matching relationship form a set of matching target vascular regions.

[0037] Preferably, in the embodiment of the present invention, considering that the regions of the same kind of venous vessel should show similar regional trends between different layers of the liver, the matching target vascular regions between adjacent layers should have similar postures and similar distances. Therefore, the method for obtaining the mutual extension probability includes: In the set of matching target vascular regions, for two matching target vascular regions corresponding to adjacent layers, obtain the minimum circumscribed rectangles of the two matching target vascular regions, extend the side lengths of the two minimum circumscribed rectangles, and the minimum included angle formed by the extension lines is used as the posture difference angle. The larger the posture difference angle, the greater the difference in the postures of the two regions.

[0038] Obtain the distance between two matching target vascular regions. Thus, the farther the distance, the larger the pose difference angle, indicating that the two regions are less likely to be the result of the extension of the same venous region between different levels. Therefore, calculate the Euclidean norm based on the distance and the pose difference angle, and perform a negative correlation mapping on the Euclidean norm to obtain the initial mutual extension probability between the two matching target vascular regions. It should be noted that in the embodiments of the present invention, the negative correlation mapping result can be in the form of a reciprocal, and those skilled in the art can also use other basic mathematical means to implement it, which will not be elaborated here.

[0039] In the set of matching target vascular regions, since there are analyses between multiple adjacent levels, statistically calculate the initial mutual extension probability of each pair of matching target vascular regions corresponding to all adjacent levels, and use the average initial mutual extension probability as the mutual extension probability of the set of matching target vascular regions. That is, the greater this mutual extension probability, the more likely it is that the set of matching target vascular regions is the result of the extension of the same venous region.

[0040] Step S4: For each liver segment, traverse and combine all the sets of matching target vascular regions included, and arrange them according to the mutual extension probability to obtain multiple combination methods; obtain the change trend characteristics of the respiratory influence degree of the sets of matching target vascular regions in each combination method; according to the change trend characteristics, the number of layers included in the set of matching target vascular regions, the fusiform feature degree of the target vascular regions corresponding to the set of matching target vascular regions, and the number of elements in the combination method, screen out the optimal combination method.

[0041] In step S3, sets of matching target vascular regions representing different level information are obtained. For different sets of matching target vascular regions, they can be regarded as venous regions at different positions within a liver segment. To obtain the complete portal vein region in the liver segment, it is necessary to determine which venous regions represented by the matching vascular region sets belong to the portal vein region. Therefore, in the embodiments of the present invention, a method of traversal analysis is adopted. For each liver segment, traverse and combine all the sets of matching target vascular regions included. For example, if there are four sets of matching target vascular regions A, B, C, and D, then nine combination methods such as AB, AC, AD, BC, BD, CD, ABC, ACD, and BCD will be formed.

[0042] Among CT images at multiple different levels, as the portal vein branches penetrate deeper into the liver segment area, the diameter of the blood vessels gradually decreases, and at this time, the degree of influence of respiration on them gradually increases; while for the hepatic vein, it usually has a relatively uniform diameter, so it is generally more affected by respiration and the values in different layers are approximately the same. That is, the degree of influence of respiration on the portal vein branches changes with the extension of the layers. The mutual extension probability of the matching target blood vessel regions represents the degree of extension of the regions in this set. The greater the degree of extension, the greater the overall amplification of the final respiratory influence on the matching target blood vessel regions, and vice versa, indicating that the respiratory influence remains at a small value. That is, for the matching target blood vessel regions, their overall respiratory influence is correlated with the mutual extension probability. Therefore, in the embodiments of the present invention, the matching target blood vessel region sets in the combination method are arranged according to the mutual extension probability, and then the change trend characteristics of the respiratory influence degree are determined. The more relevant the change trend characteristics are to the arrangement characteristics of the mutual extension probability, the more in line with the portal vein characteristics this combination method is.

[0043] Furthermore, for the portal vein region, it extends on multiple levels of the liver and has a relatively obvious fusiform degree. Since the portal vein branches are close to the end of the liver segment, the actual number of portal vein branches shown in the liver segment should be small, so the number of elements in the combination method should also be small. Therefore, the optimal combination method can be selected according to the change trend characteristics, the number of layers included in the matching target blood vessel region set, the fusiform characteristic degree of the target blood vessel region corresponding to the matching target blood vessel region set, and the number of elements in the combination method.

[0044] Preferably, in the embodiments of the present invention, the matching target blood vessel region sets in each combination method are arranged in descending order based on the mutual extension probability, that is, the obtained change trend characteristics should also be a decreasing downward trend characteristic. Therefore, the method for obtaining the change trend characteristics includes: Obtain the average respiratory influence degree of each matching target blood vessel region set; Arrange the matching target blood vessel region sets in each combination method in descending order based on the mutual extension probability, and the average respiratory influence degrees of the matching target blood vessel region sets form a respiratory influence degree sequence.

[0045] In the respiratory influence degree sequence, obtain the element difference of subtracting the previous element from the next element; use the cumulative value of the element differences as the numerator and the cumulative value of the absolute values of the element differences as the denominator to obtain the change trend characteristics of each combination method. That is, the value range of the ratio corresponding to the change trend characteristics is between -1 and 1, and the closer it is to 1, the more obvious the downward trend is, and the greater the possibility of representing the portal vein branches.

[0046] It should be noted that in other embodiments of the present invention, if the sets of target vascular regions in each combination mode are arranged in ascending order of the mutual extension probability, the rising trend characteristics of the average respiratory influence degree should be analyzed, which will not be elaborated here.

[0047] Furthermore, the method for screening the optimal combination mode includes: Based on the change trend characteristics, the number of layers included in the set of target vascular regions, the fusiform feature degree of the target vascular region corresponding to the set of target vascular regions, and the number of elements in the combination mode, obtain the venous branching possibility of each combination mode, and take the combination mode with the largest venous branching possibility as the optimal combination mode. The method for obtaining the venous branching possibility includes: For each combination mode, obtain the maximum fusiform feature degree of the target vascular region in each set of target vascular regions.

[0048] Take the product of the maximum fusiform feature degree and the number of layers as the venous branching feature of each set of target vascular regions. That is, the more the number of layers, the more it conforms to the characteristics of the portal vein extending in multiple layers of liver segments; the greater the maximum fusiform feature degree, the more it conforms to the shape characteristics of the portal vein.

[0049] Since there are multiple sets of target vascular regions in a combination mode, multiply the cumulative value of the venous branching features in the combination mode by the change trend characteristics, and subtract the number of sets of target vascular regions in the combination mode from the product to obtain the venous branching possibility. That is, the greater the change trend characteristics, the more the combination mode conforms to the change characteristics of the portal vein region; the greater the cumulative value of the venous branching features, the more the sets of target vascular regions in the combination mode conform to the branch region characteristics of the portal vein; the smaller the number of sets of target vascular regions in the combination mode, the more it conforms to the quantity distribution characteristics of the portal vein in the liver segment. Therefore, the greater the finally quantified venous branching possibility, the more likely the combination mode is the combination of the portal vein branches.

[0050] Step S5: Use the target vascular region corresponding to the optimal combination mode as the portal vein branches and perform three-dimensional reconstruction for assisting in the delineation of adjacent regions.

[0051] After determining the optimal combination method for representing the portal vein branches, three-dimensional reconstruction can be performed based on the corresponding target vascular area as the portal vein branch area. The three-dimensional reconstruction can be completed using the post-processing workstation that comes with the CT machine. The reconstruction methods include multiple planar reformation (MPR), volume rendering (VR) and maximum intensity projection (MIP). MPR reconstruction includes coronal, sagittal and curved planar reformation (CPR), with a layer thickness of 2.0 mm, an interval of 2.0 mm, a window width of 250 Hu, and a window position of 50 Hu.

[0052] For all liver segments, reconstructed images of the corresponding portal vein branches can be obtained; based on the segmentation results of the portal vein branches on the above liver segments, doctors can be assisted in outlining the adjacent areas between different liver segments, thereby improving the accuracy of the liver segmentation process.

[0053] In summary, the embodiment of the present invention analyzes the CT images at each level of the liver, and uses the shape changes and area changes of the vascular region to quantify the degree of respiratory influence. By performing matching analysis on the target vascular regions between adjacent levels, the changing trend characteristics of the degree of respiratory influence are determined, and then all target vascular regions are traversed by matching and then recombining. Based on the changing trend characteristics, the optimal combination method representing the portal vein can be determined, and the vascular region represented by the portal vein can be screened out, and effective assistance in delineating adjacent regions can be achieved through three-dimensional reconstruction. The present invention can effectively assist in delineating adjacent regions by analyzing the association of vascular regions in CT images of different levels, screening out accurate portal vein regions, and performing three-dimensional reconstruction.

[0054] It should be noted that the sequence of the above embodiments of the present invention is for description only and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying 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.

[0055] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. An adjacent area delineation assistance method before hepatic segment resection, characterized in that, The method includes: Obtaining consecutive multiple-frame CT images of each layer of the patient's liver under CT sectional scanning, where the vascular region in the first-frame CT image is used as the target vascular region; Obtaining the fusiform feature degree of each vascular region; at each layer, successively matching the vascular regions in adjacent-frame CT images, and obtaining the respiratory influence degree of the target vascular region according to the area change and fusiform feature degree change between the matching regions; Matching the target vascular regions with the same hepatic segment between adjacent layers, and the target vascular regions with a matching relationship form a set of matching target vascular regions; in the set of matching target vascular regions, obtaining the mutual extension probability of the set of matching target vascular regions according to the distance and position distribution between the regions; For each hepatic segment, traversing and combining all the sets of matching target vascular regions included, arranging them according to the mutual extension probability, and obtaining multiple combination methods; obtaining the change trend characteristics of the respiratory influence degree of the sets of matching target vascular regions in each combination method; according to the change trend characteristics, the number of layers included in the set of matching target vascular regions, the fusiform feature degree of the target vascular region corresponding to the set of matching target vascular regions, and the number of elements in the combination method, screening out the optimal combination method; Taking the target vascular region corresponding to the optimal combination method as the portal vein branch and performing three-dimensional reconstruction for assisting in the delineation of adjacent regions.

2. The adjacent area delineation assistance method before hepatic segment resection according to claim 1, wherein, The method for obtaining the fusiform feature degree includes: Obtaining the minimum bounding rectangle of the vascular region, taking the aspect ratio of the minimum bounding rectangle as the initial fusiform feature degree, and taking the product of the average gray value of the vascular region and the initial fusiform feature degree as the fusiform feature degree.

3. The adjacent region delineation assistance method before hepatectomy according to claim 1, wherein Successively matching the vascular regions in adjacent-frame CT images includes: For each pair of adjacent frames at each layer, obtaining the coincidence degree between the already-matched vascular region in the previous-frame CT image and the vascular region in the next-frame CT image, and selecting the vascular region with the highest coincidence degree as the new already-matched vascular region; where the already-matched vascular region in the first-frame CT image is any one of the target vascular regions; Matching the target vascular regions with the same hepatic segment between adjacent layers includes: For any two adjacent layers, matching the first-frame CT images between the two layers, and selecting a group of target vascular regions with the largest coincidence degree as the pair of matching target vascular regions with a matching relationship; traversing all layers, and forming a set of matching target vascular regions from the target vascular regions included in the pairs of matching target vascular regions with a mutual matching relationship.

4. A method for assisting in delineating adjacent regions before hepatic segmentectomy according to claim 3, characterized in that The method for obtaining the respiratory influence degree includes: For each group of adjacent frames, obtaining the difference in the average fusiform feature degree and the difference in the average area between the already-matched vascular region in the previous frame and the already-matched vascular region in the next frame, and obtaining the morphological distortion degree; statistically calculating the morphological distortion degrees of all groups of adjacent frames, and obtaining the respiratory influence degree according to the average morphological distortion degree and the variance of the morphological distortion degree.

5. The adjacent area delineation assistance method before hepatectomy according to claim 4, wherein, The method for obtaining the morphological distortion degree includes: For each set of adjacent frames, obtain a first ratio of the average fusiform feature degree between the matched vascular regions of the previous frame and the matched vascular regions of the next frame, and a second ratio of the average area; obtain a first product between the first ratio and the second ratio, and normalize the difference between the first product and the positive integer 1 to obtain the morphological distortion degree.

6. The adjacent region delineation assistance method before hepatectomy according to claim 1, characterized in that, The method for obtaining the mutual extension probability includes: In the set of matched target vascular regions, for two matched target vascular regions corresponding to adjacent layers, obtain the minimum circumscribed rectangles of the two matched target vascular regions, extend the side lengths of the two minimum circumscribed rectangles, and use the minimum included angle formed by the extension lines as the pose difference angle; obtain the distance between the two matched target vascular regions, calculate the Euclidean norm based on the distance and the pose difference angle, and perform a negative correlation mapping on the Euclidean norm to obtain the initial mutual extension probability between the two matched target vascular regions. In the set of matched target vascular regions, count the initial mutual extension probabilities of each pair of matched target vascular regions corresponding to all adjacent layers, and use the average initial mutual extension probability as the mutual extension probability of the set of matched target vascular regions.

7. The adjacent region delineation assistance method before hepatectomy according to claim 1, wherein, The method for obtaining the change trend feature includes: Obtain the average degree of respiratory influence of each set of matched target vascular regions. Arrange the sets of matched target vascular regions in each combination method in descending order based on the mutual extension probability, and the average degree of respiratory influence of the sets of matched target vascular regions constitutes a sequence of respiratory influence degrees. In the sequence of respiratory influence degrees, obtain the element difference obtained by subtracting the previous element from the next element; use the cumulative value of the element differences as the numerator and the cumulative value of the absolute values of the element differences as the denominator to obtain the change trend feature of each combination method.

8. A method for assisting in delineating adjacent regions before hepatectomy according to claim 7, characterized in that, The method for screening the optimal combination method includes: According to the change trend feature, the number of layers included in the set of matched target vascular regions, the fusiform feature degree of the target vascular regions corresponding to the set of matched target vascular regions, and the number of elements in the combination method, obtain the venous branching possibility of each combination method, and use the combination method with the largest venous branching possibility as the optimal combination method.

9. A method for assisting in delineating adjacent regions before hepatectomy according to claim 8, characterized in that, The method for obtaining the venous branching possibility includes: For each combination method, obtain the maximum fusiform feature degree of the target vascular regions in the set of matched target vascular regions; use the product of the maximum fusiform feature degree and the number of layers as the venous branching feature of the set of matched target vascular regions; multiply the cumulative value of the venous branching features in the combination method by the change trend feature, and subtract the number of sets of matched target vascular regions in the combination method from the product to obtain the venous branching possibility.

10. The adjacent region delineation assistance method before hepatectomy according to claim 1, characterized in that, The method for obtaining the vascular region includes: Perform edge detection on each frame of CT image, and use the closed region formed by the edges as the vascular region.

Images