A system for identifying calcified nodule lesion regions in coronary angiography

Abstract

This application relates to the field of medical image recognition, and particularly to a system for identifying calcified nodule lesion regions in coronary angiography. The system includes: a coronary angiography image acquisition module that acquires pre-contrast key-frame images and post-contrast key-frame images; a first coronary angiography image analysis module that analyzes the gray-scale distribution and gray-scale change conditions of pixel points in the pre-contrast key-frame images and the post-contrast key-frame images to obtain a set of suspected calcification points; a second coronary angiography image analysis module that combines the distance distribution of elements in the set of suspected calcification points to the edge to obtain a contrast coefficient; obtains the relative displacement of each pixel point through the difference characteristics of the gray-scale distribution between adjacent pixel points in the post-contrast key-frame images, and filters to obtain a suspected calcification distance; a calcified nodule lesion region image recognition module that obtains a cardiac motion fluctuation coefficient and identifies the calcified lesion regions in coronary angiography. This application aims to enhance the recognition ability of tiny calcified lesion regions.

Description

Technical Field

[0001] This application relates to the field of medical image recognition, and specifically to a system for identifying calcified nodule lesion regions in coronary angiography. Background Art

[0002] Cardiovascular diseases are one of the main causes of death globally, especially coronary artery diseases, which are of particular importance. The coronary arteries supply essential oxygen and nutrients to the myocardium. When these blood vessels become narrowed or blocked due to factors such as atherosclerosis, it may lead to serious complications such as angina pectoris and myocardial infarction. Coronary artery calcification is a key factor in arteriosclerosis, further exacerbating vascular stiffness and stenosis. Among various coronary artery examination techniques, coronary angiography is widely regarded as the "gold standard" for diagnosing coronary atherosclerotic heart disease because it can directly display coronary artery lesions, greatly reducing the risk of misdiagnosis and missed diagnosis, and having high diagnostic accuracy.

[0003] Although coronary angiography has been quite mature in clinical applications, there are still some deficiencies in the existing technology for identifying calcified lesions. Currently, the annotation of lesion regions in coronary angiography images still mainly relies on manual annotation by physicians. Although there are already some automated tools for assisting diagnosis, these tools have good effects on lesions with relatively clear features such as thrombus, but the annotation and recognition effects for lesions with relatively low feature clarity like calcification are not ideal; at the same time, the existing automated tools for assisting diagnosis often process and identify single-frame images, resulting in poor recognition effects. Analyzing only based on single-frame images may even overlook calcification in some smaller regions of the coronary arteries. Summary of the Invention

[0004] In view of the above, it is necessary to provide a system for identifying calcified nodule lesion regions in coronary angiography to solve the above problems.

[0005] An embodiment of this application provides a system for identifying calcified nodule lesion regions in coronary angiography, and the system includes:

[0006] A coronary angiography image acquisition module, configured to obtain coronary angiography dynamic images, and classify all frames of coronary angiography images obtained within one cardiac cycle before injecting the contrast agent as pre-contrast key frame images; and classify all frames of coronary angiography images obtained within one cardiac cycle after injecting the contrast agent as post-contrast key frame images.

[0007] The first coronary angiography image analysis module is used to analyze the gray value distribution of pixel points in the key pre-contrast agent frame images, so as to obtain the high-density shadow set and the suspected blood vessel set of the key pre-contrast agent frame images; form a pair of the key pre-contrast agent frame images and the key post-contrast agent frame images with the same index within the cardiac cycle, analyze the gray value change of the pixel points at the same position in each pair of images, and combine the high-density shadow set to obtain the suspected calcification point set of the key post-contrast agent frame images; perform edge detection on each key post-contrast agent frame image, and according to the distance distribution of the elements in the suspected calcification point set of the key post-contrast agent frame image to the edges in the image and the edge pixel point distribution, combine the suspected blood vessel set of the key pre-contrast agent frame images to obtain the angiography coefficient;

[0008] The second coronary angiography image analysis module is used to calculate the matching index between pixel points by sequentially comparing the difference features of the gray value distributions between the pixel points of all adjacent key post-contrast agent frame images, obtain the matching point set of each pixel point in the initial key post-contrast agent frame image of the cardiac cycle, analyze the matching index between adjacent elements in the matching point set, and combine the position distribution of the elements in the matching point set to obtain the relative displacement of each pixel point, and screen to obtain the suspected calcification distance;

[0009] The calcification nodule lesion area image recognition module is used to analyze the distribution characteristics and quantity characteristics of the suspected calcification distance to obtain the cardiac fluctuation coefficient, and combine the angiography coefficient to identify the calcification lesion area in the coronary angiography.

[0010] Among them, the specific steps for obtaining the high-density shadow set and the suspected blood vessel set of the key pre-contrast agent frame images include:

[0011] In each key pre-contrast agent frame image, obtain the set composed of all pixel point coordinates whose gray value is greater than the upper quartile of the gray values of all pixel points in the corresponding image, and record it as the high-density shadow set;

[0012] Obtain the set composed of all pixel point coordinates whose gray value is less than the lower quartile of the gray values of all pixel points in the corresponding image, and record it as the suspected blood vessel set.

[0013] Among them, the specific process for obtaining the suspected calcification point set of the key post-contrast agent frame images is:

[0014] In each pair of images, calculate the gray value change amount of the pixel points at the same position, and form the set composed of the pixel point coordinates whose gray value change amount is less than the lower quartile of all gray value change amounts in the corresponding image, and record it as the gray value micro-change point set;

[0015] Take the intersection of the set of gray-scale micro-change points of each pair of images and the set of high-density shadows of the pre-contrast key-frame images in each pair of images as the set of suspected calcification points of the post-contrast key-frame images in each pair of images.

[0016] Among them, the obtaining of the contrast coefficient is specifically as follows:

[0017] For each post-contrast key-frame image, obtain the set of suspected calcification point distance distributions of the post-contrast key-frame image according to the distance distribution of the elements in the set of suspected calcification points to the edge in the image;

[0018] Calculate the intersection-over-union ratio of the set of suspected blood vessels of each pre-contrast key-frame image and the set composed of all edge pixel points in the post-contrast key-frame image with the same index;

[0019] Obtain the dispersion of the set of suspected calcification point distance distributions of each post-contrast key-frame image;

[0020] Fuse the negative correlation mapping results of the dispersions of all pre-contrast key-frame images and the intersection-over-union ratios of all post-contrast key-frame images to obtain the contrast coefficient.

[0021] Among them, the process of obtaining the set of suspected calcification point distance distributions of the post-contrast key-frame image is as follows:

[0022] In each post-contrast key-frame image, calculate the minimum distance between each element in the set of suspected calcification points and the edge line in the image, and take the set composed of the minimum distances as the set of suspected calcification point distance distributions of each post-contrast key-frame image.

[0023] Among them, the calculation of the matching index between pixel points is specifically as follows:

[0024] For each post-contrast key-frame image, obtain the gray-scale values of the pixel points in the neighborhood of each element in the set of suspected calcification points to form a gray-scale value sequence for each element;

[0025] For a pixel point in each post-contrast key-frame image and a pixel point in the next adjacent post-contrast key-frame image, obtain the Euclidean distance between the two pixel point coordinates and the similarity measure between the two corresponding gray-scale value sequences; fuse the negative correlation mapping of the Euclidean distance and the similarity measure to obtain the matching index between the two pixel points.

[0026] Among them, the obtaining of the set of matching points of each pixel point in the initial post-contrast key-frame image of the cardiac cycle is specifically as follows:

[0027] For any pixel point corresponding to an element in the set of suspected calcification points of each post-contrast key frame image, the pixel point corresponding to the element with the largest matching index among the suspected calcification points of the adjacent next post-contrast key frame image is used as the matching point of the pixel point corresponding to the said any element;

[0028] For each pixel point in the initial contrast key frame image of the cardiac cycle, according to the transitive feature, all the matching points of each pixel point form a set of matching points of each pixel point;

[0029] The transitive feature is: for pixel point a, if pixel point b is the matching point of pixel point a and pixel point c is the matching point of pixel point b, then both pixel points b and c are the matching points of pixel point a.

[0030] Among them, the obtaining of the relative displacement of each pixel point is specifically as follows: ; where is the relative displacement of each pixel point, is the matching index between two pixel points belonging to the th and th post-contrast key frame images in the set of matching points of each pixel point, is the distance weight of the th matching index; where , is the number of elements in the set of matching points, is the index of the matching index.

[0031] Among them, the obtaining of the suspected calcification distance is specifically as follows:

[0032] Taking the relative displacements of all pixel points in the initial post-contrast key frame image of the cardiac cycle as the input, using a threshold segmentation algorithm to obtain a segmentation threshold, and taking the relative displacements less than the segmentation threshold as the suspected calcification distance.

[0033] Among them, the obtaining of the cardiac fluctuation coefficient, in combination with the contrast coefficient, to identify the calcified lesion area in coronary angiography, the specific process is as follows:

[0034] Calculating the proportion of the number of suspected calcification distances in the total number of all relative displacements; dividing by the mean value of all suspected calcification distances, and then dividing by the variance of all suspected calcification distances; obtaining the cardiac fluctuation coefficient;

[0035] Obtaining samples of a preset number of coronary angiography dynamic images, and respectively performing pre-segmentation on the contrast coefficients and cardiac fluctuation coefficients of all samples to obtain a contrast coefficient threshold and a cardiac fluctuation coefficient threshold;

[0036] If the contrast coefficient of the dynamic coronary angiography image is greater than the contrast coefficient threshold and the cardiac motion coefficient is greater than the cardiac motion coefficient threshold, it is determined that there is a calcified lesion area in the dynamic coronary angiography image, and the lesion area is all pixel points corresponding to the suspected calcification distance.

[0037] The present application has at least the following beneficial effects:

[0038] In the embodiment of the present application, by analyzing the changes in the coronary artery images before and after the injection of the contrast agent, a set of suspected calcification points is obtained. Based on the distribution characteristics of the calcification points, pixel points are screened, providing an important basis for subsequent calcification point judgment and lesion evaluation; according to the edge detection results in the post-contrast key frame image, the distance characteristics between the suspected calcification points and the edge are analyzed, and combined with the set of suspected blood vessels in the pre-contrast key frame image, the contrast coefficient is obtained to evaluate the change in the gray level of the pixel points in the coronary artery images before and after the injection of the contrast agent, thereby initially measuring whether there is a calcification condition and the degree of calcification, enhancing the recognition ability of calcification characteristics; then, according to the temporal difference caused by the cardiac motion of the coronary artery vessels after the injection of the contrast agent, the difference characteristics of the gray level distribution between the pixel points of all adjacent post-contrast key frame images are compared, the matching index between the pixel points is calculated, and the matching point set of each pixel point in the initial post-contrast key frame image of the cardiac cycle is obtained. The beneficial effect is to analyze the similarity between the pixel points in adjacent post-contrast key frame images, facilitating the subsequent description of the displacement of each pixel point over time; by analyzing the matching index between adjacent elements in the matching point set and combining the position distribution of the elements in the matching point set, the relative displacement of each pixel point is obtained, and the suspected calcification distance is screened. The beneficial effect is to identify the displacement of the suspected calcification points over time, and it can also identify the calcified area and accurately calculate its position and range; finally, based on the distribution of the suspected calcification distance, the cardiac motion coefficient is calculated, and combined with the contrast coefficient, the calcified lesion area in the coronary angiography is identified, enhancing the recognition ability of the calcified lesion characteristics in the coronary artery vessels, and at the same time helping to enhance the comparative recognition ability of the tiny calcified lesion area. Description of the Drawings

[0039] Figure 1 It is a block diagram of a system for identifying calcified nodule lesion areas in coronary angiography provided by the present application;

[0040] Figure 2 It is a specific flowchart for identifying calcified lesion areas in coronary angiography provided by the present application. Detailed Embodiments

[0041] In the description of the embodiments of the present application, words such as "exemplary", "or", "for example" and the like are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary", "or", "for example" and the like is intended to present related concepts in a concrete manner.

[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art in the present application. The terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application.

[0043] It should also be noted that the terms "first" and "second" in this application and its drawings are used to distinguish similar objects, rather than to describe a specific order or sequence. The method disclosed in the embodiments of the present application or the method shown in the flow chart includes one or more steps for implementing the method. Without departing from the scope of protection of this application, the execution order of multiple steps can be interchanged with each other, and some steps can also be deleted.

[0044] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0045] The specific scheme of the system for identifying calcified nodular lesion areas in coronary angiography provided by the present application is described in detail below with reference to the accompanying drawings.

[0046] See also Figure 1 , which shows a block diagram of a calcified nodule lesion area identification system for coronary angiography provided by an embodiment of the present application. The system includes: a coronary angiography image acquisition module, a first coronary angiography image analysis module, a second coronary angiography image analysis module, and a calcified nodule lesion area image recognition module.

[0047] The present application embodiment first proposes a system for identifying calcified nodular lesion areas in coronary angiography, which is applied to the field of medical image recognition. The system includes:

[0048] Coronary angiography image acquisition module: obtains dynamic coronary angiography images, and classifies all frames of coronary angiography images obtained within a cardiac cycle before the injection of contrast agent as pre-contrast agent key frame images; within a cardiac cycle after the injection of contrast agent, classifies all frames of coronary angiography images obtained as post-contrast agent key frame images.

[0049] This application adopts the posteroanterior head position Position, insert a catheter through the radial artery puncture, inject an iodide contrast agent into the catheter, and observe the imaging of blood vessels under X-ray to obtain the original coronary angiography data. The original data is a Dicom file with continuous angiography frames. Then, in the Dicom file, all frame images within one cardiac cycle before injecting the contrast agent are used as pre-contrast key frames, and all frame images within one cardiac cycle after injecting the contrast agent are used as post-contrast key frames. The key frame sequences before and after contrast agent injection are respectively composed of pre-contrast image sequences and post-contrast image sequences in chronological order. Among them, the cardiac cycle refers to the time interval from the start of one heart beat to the start of the next heart beat. Enhance each key frame image. In this embodiment, a multi-scale Retinex enhancement algorithm is used. In other embodiments, image enhancement methods that can be used include image enhancement based on the Laplacian operator, image enhancement based on logarithmic transformation, image enhancement based on gamma transformation, etc. This application does not limit this.

[0050] First coronary angiography image analysis module: Analyze the gray value distribution of pixel points in the pre-contrast key frame images to obtain the high-density shadow set and suspected blood vessel set of the pre-contrast key frame images; form a pair of pre-contrast key frame images and post-contrast key frame images with the same index within the cardiac cycle, analyze the gray value changes of pixel points at the same position in each pair of images, and combine with the high-density shadow set to obtain the suspected calcification point set of the post-contrast key frame images; perform edge detection on each post-contrast key frame image, and according to the distance distribution of elements in the suspected calcification point set of the post-contrast key frame image to the edges in the image and the edge pixel point distribution, combine with the suspected blood vessel set of the pre-contrast key frame images to obtain the angiography coefficient.

[0051] Coronary artery calcification refers to a phenomenon in which calcium salts are abnormally deposited on the blood vessel wall in the coronary artery. As the calcium deposition becomes more and more serious, the scope of atherosclerotic lesions will also become wider and wider. Therefore, timely identification of the calcification area and degree in the coronary artery is crucial for coronary artery examination. Coronary angiography is widely used in clinical practice. At present, most lesion markings and identifications in coronary angiography still mainly rely on manual operations by doctors. However, due to the characteristics of calcified lesions often not being significant enough, there is a relatively high risk of missed diagnosis; although there are already automated tools to assist in diagnosis to enhance the sensitivity of identification, this tool often focuses too much on the information in a single frame image and ignores the change process of coronary angiography, thus the accuracy still needs to be improved; especially for calcified lesions in some small areas, it is even more difficult to detect only relying on a single frame image.

[0052] The characteristics of calcified regions in coronary arteries are closely related to the degree of calcification. Before injecting the contrast agent, mild calcification appears as faint and blurred high-density images during cardiac pulsation, while no calcification image can be observed during cardiac rest, and the coronary artery outline is hardly visible; in moderate calcification, relatively clear and distinguishable high-density images can be observed during cardiac pulsation, and the outline of some coronary arteries is visible; severe calcification is manifested as obvious high-density images both during cardiac pulsation and rest, and the coronary artery outline is relatively clear. In X-ray images, due to the high atomic number of calcium in the calcified region, the ability to absorb X-rays is stronger than that of surrounding soft tissues. Therefore, the high-density region usually appears as a brighter region than the surrounding tissues, that is, a bright white region with a larger gray value. At the same time, since calcium salts are deposited on the blood vessel wall, these regions tend to be distributed along the blood vessels. While normal blood vessel soft tissues present gray or black images with lower gray values.

[0053] After injecting the contrast agent, since the contrast agent is a high-density substance, it can significantly enhance the X-ray absorption of blood vessels and other soft tissues, thus making the gray value changes of these structures relatively obvious on the image; for the region with calcification foci, since calcification itself is a high-density substance and already has a high X-ray absorption rate without the contrast agent, even after injecting the contrast agent, the increase in brightness of the calcified region is relatively small.

[0054] In the pre-contrast agent image sequence, taking the th image as an example, the gray value of each pixel point in the image is obtained. The set composed of all pixel points with gray values greater than the upper quartile in the th image is denoted as the high-density shadow set, and the set composed of all pixel points with gray values less than the lower quartile in the th image is the suspected blood vessel set. In this way, the high-density shadow set and the suspected blood vessel set of each image in the pre-contrast agent image sequence are obtained respectively.

[0055] Then, the images with the same index in the pre-contrast agent image sequence and the post-contrast agent image sequence are paired. In each pair of images, the absolute value of the difference in gray values of the pixel points at the same position in the two images is used as the gray value change amount. All gray value change amounts are obtained. The set of pixel points with all gray value change amounts less than the lower quartile of all gray value change amounts of the corresponding image constitutes the gray value micro-change point set; the gray value micro-change point sets of all image pairs are obtained; then, for each image in the post-contrast agent image sequence after injecting the contrast agent, the Canny edge detection algorithm is used to respectively output the corresponding edge detection images. Canny edge detection is a well-known technology and will not be elaborated here.

[0056] Take the intersection of the set of high-density shadow collections of an image in the pre-contrast agent image sequence and the set of gray-scale micro-variation points of the same-index image pair as the set of suspected calcification points of the post-contrast agent key-frame image at the same index, and calculate the minimum distance between each pixel point in the set of suspected calcification points of each post-contrast agent key-frame image and the edge line in the same post-contrast agent key-frame image. All the minimum distances in a post-contrast agent key-frame image constitute the set of suspected calcification point distance distributions of this post-contrast agent key-frame image.

[0057] According to the distribution of the set of suspected blood vessels in each pre-contrast agent key-frame image and the suspected calcification points in the post-contrast agent key-frame image with the same index, obtain the contrast coefficient, which is used to evaluate the change in the gray scale of pixel points in the coronary artery images before and after the injection of the contrast agent, so as to preliminarily measure whether there is a calcification situation and the degree of calcification: calculate the intersection-over-union ratio of the set of suspected blood vessels in each pre-contrast agent key-frame image and the set composed of all edge pixel points in the post-contrast agent key-frame image with the same index; obtain the dispersion of the set of suspected calcification point distance distributions of each post-contrast agent key-frame image; fuse the negative correlation mapping results of the dispersions of all pre-contrast agent key-frame images with the intersection-over-union ratios of all post-contrast agent key-frame images to obtain the contrast coefficient.

[0058] In this embodiment, the intersection-over-union ratio is specifically the ratio of the number of elements in the intersection of two sets to the number of elements in the union of the two sets; the dispersion of the set of suspected calcification point distances of the post-contrast agent key-frame image is calculated using variance, and the formula form of the contrast coefficient is: ; where represents the intersection-over-union ratio of the i-th pre-contrast agent key-frame image; represents the dispersion of the i-th post-contrast agent key-frame image; N represents the total number of pre-contrast agent key-frame images, and also represents the total number of post-contrast agent key-frame images; A represents the contrast coefficient.

[0059] It should be understood that when there is a calcification situation, the calcification is distributed along the blood vessels, that is, the distance between the calcified pixel points and the blood vessel edge line is relatively small, and at the same time, the severity of the calcification is proportional to the visibility of the blood vessels before angiography, that is, the more severe the calcification, the clearer the blood vessels before angiography, so the intersection-over-union ratio before and after angiography is larger, so the contrast coefficient is larger at this time. On the contrary, when there is no calcification or the degree of calcification is lower, the contrast coefficient is smaller.

[0060] Coronary angiography image second analysis module: By sequentially comparing the difference features of the gray-scale distributions between the pixel points of all adjacent post-contrast key-frame images, calculating the matching index between the pixel points, obtaining the set of matching points for each pixel point in the initial post-contrast key-frame image of the cardiac cycle, analyzing the matching index between adjacent elements in the set of matching points, and combining the position distribution of the elements in the set of matching points, the relative displacement of each pixel point is obtained, and the suspected calcification distance is screened out.

[0061] Under normal circumstances, with the rhythm of the heartbeat, the blood vessels will be pushed by the blood flow and undergo dilation and contraction. When a contrast agent is injected, the normal blood vessels will change the lumen with the contraction and relaxation of the heart, the contrast agent gradually fills the blood vessels, or dynamic vascular stenosis appears. Therefore, in different image frames, the position and shape of the blood vessel wall will undergo obvious displacement and change. When calcification occurs, since the calcium component is hard and not easily changed, the calcified area appears as a relatively stable high-density area in different frame images; at the same time, calcification is usually accompanied by the pathological process of atherosclerosis, and the deposition of calcium in the blood vessel wall will cause the blood vessel wall to harden and the elasticity to decrease significantly, making the calcified area unable to effectively perform the movement of dilation and contraction, so the amplitude of dilation and contraction of the blood vessels in the calcified area becomes smaller, showing a smaller displacement of the calcified area in different frame images.

[0062] In the post-contrast image sequence, taking the th post-contrast key-frame image as an example, obtain the gray-scale values within the eight-neighborhood of each pixel point in the set of suspected calcification points of this image, and form the gray-scale value sequence of this pixel point in the order from left to right and from top to bottom. The sequence obtained by this construction method can measure the gray-scale value distribution within the neighborhood of each pixel point.

[0063] By comparing the difference features of the gray-scale distributions between the pixel points of two adjacent post-contrast key-frame images, the matching index between the pixel points in two adjacent post-contrast key-frame images is obtained, and its formula form is: ; is the matching index between pixel point in the th and pixel point in the th post-contrast key-frame images, where pixel point belongs to the th post-contrast key-frame image, pixel point belongs to the th post-contrast key-frame image, is the Euclidean distance between the coordinates of pixel point a and pixel point b, It is a similarity measure between the grayscale value sequences of these two pixel points. In this embodiment, cosine similarity is adopted; in other embodiments, the similarity measure methods that can be used include Pearson correlation coefficient, Spearman correlation coefficient, etc., and the present application does not limit this; It represents a parameter preset to be greater than zero. In this embodiment, the value is 0.01, and its function is to avoid the denominator being zero.

[0064] It can be understood that since the pixel points in the calcified area have small displacements and are relatively stable in the high-density area, the positions and grayscale value distributions between the pixel points at the same calcified position in different frame images are relatively close, so the matching index The larger it is, the more likely it is that the two pixel points are the pixel points at the same calcified position in different frame images. Therefore, for each pixel point in the set of suspected calcification points of the th post-contrast key frame image, the one with the largest matching index in the set of suspected calcification points of the th post-contrast key frame image is used as the matching point.

[0065] In the above manner, the matching points of each pixel point in the set of suspected calcification points of all post-contrast key frame images are obtained. The matching points have transitivity. According to the transitivity feature, the set of matching points of all the matching points of a pixel point is formed. Transitivity means that, assuming that for the pixel point in the set of suspected calcification points of the th image, the matching point is the pixel point in the set of suspected calcification points of the th image, and 's matching point is the pixel point in the set of suspected calcification points of the th image, then is also 's matching point. After obtaining the set of matching points of each pixel point in the set of suspected calcification points of the first image in the post-contrast image sequence, the relative displacement of each pixel point is calculated, and its formula form is: ; where is the relative displacement of each pixel point, is the matching index between two pixel points belonging to the th and th post-contrast key frame images in the set of matching points of each pixel point, is the distance weight of the th matching index. Among them , is the number of elements in the set of matching points.

[0066] It can be understood that as the number of frames between the first image of the post-contrast agent image sequence increases, the difference between coronary vasoconstriction and vasodilation becomes more and more obvious. At this time, the distance displacement of the calcified area is also relatively larger. According to the characteristics of the vascular calcified area, its displacement change should be relatively small. Therefore, the more the number of frames, the more the displacement size of the pixel points helps to distinguish whether it is calcification. So the corresponding weight is larger, and the relative displacement calculated is more able to reflect the displacement characteristics of the calcified area.

[0067] Calculate the relative displacement of each pixel point in the set of suspected calcification points of the first key frame image of the post-contrast agent, and then use all the relative displacements as inputs, and adopt Otsu threshold segmentation to obtain the segmentation threshold. The relative displacements less than the segmentation threshold are used as the suspected calcification distances. Otsu threshold segmentation is a well-known technology and will not be elaborated here.

[0068] Since the pixel point displacement of the calcified area is relatively small and the gray level distribution around the calcification point is relatively stable, obtaining the suspected calcification distance in the above way helps to measure the displacement situation and gray level distribution situation of the suspected calcification points in different frame images, thus contributing to the accuracy of subsequent measurement of the calcification situation.

[0069] Calcification nodule lesion area image recognition module: Analyze the distribution characteristics and quantity characteristics of the suspected calcification distances to obtain the cardiac fluctuation coefficient, and combine the contrast coefficient to identify the calcified lesion area in the coronary angiogram.

[0070] Analyze the distribution characteristics and quantity characteristics of the suspected calcification distances to obtain the cardiac fluctuation coefficient, which is used to measure the diastolic and systolic displacement changes of the blood vessels with cardiac changes: , is the cardiac fluctuation coefficient, is the mean value of all suspected calcification distances, is the variance of all suspected calcification distances, is the ratio of the number of suspected calcification distances to the number of all relative displacements. It can be understood that when there is a calcified lesion, the calcified area is relatively stable and the displacement is small, so the suspected calcification distance is relatively small. At the same time, the calcified areas are all relatively stable, and the larger the proportion of such calcification characteristics, the more serious the calcified lesion, so the cardiac fluctuation coefficient is larger. On the contrary, when there is no calcified lesion or the calcification is relatively mild, the cardiac fluctuation coefficient is smaller.

[0071] Obtain coronary angiography dynamic image samples, and obtain each frame image before and after the contrast of each sample. These It is necessary to include normal coronary angiography, slightly calcified coronary angiography, moderately calcified coronary angiography, and severely calcified coronary angiography. In this embodiment The value is 200, and each type of sample accounts for ; The implementer can adjust it according to the actual situation. Then calculate the angiography coefficient and heart rate fluctuation coefficient of each sample, and then use the Otsu threshold segmentation method with all the angiography coefficients and heart rate fluctuation coefficients as inputs respectively to obtain the angiography coefficient threshold and heart rate fluctuation coefficient threshold. When performing calcification judgment later, when the calculated angiography coefficient and heart rate fluctuation coefficient are greater than the angiography coefficient threshold and heart rate fluctuation coefficient threshold respectively, it is determined that there is a calcification lesion in the coronary angiography in this position, and the area that needs to be focused on is the pixel point corresponding to the suspected calcification distance.

[0072] In this way, by simultaneously considering the image changes before and after the contrast agent injection and the temporal changes after the contrast agent injection, it helps to enhance the ability to identify coronary calcification. At the same time, this temporal comparison method can achieve accurate identification based on the image changes before and after, even if the area of the calcification lesion is small, thus enhancing the ability to identify the calcification nodule lesions in the coronary angiography images.

[0073] Among them, the specific flowchart for identifying the calcification lesion area in coronary angiography is as Figure 2 shown.

[0074] In summary, in the embodiments of the present application, by analyzing the changes in coronary artery images before and after the injection of a contrast agent, a set of suspected calcification points is obtained. Based on the distribution characteristics of the calcification points, pixel points are screened, providing an important basis for subsequent determination of calcification points and lesion evaluation. According to the edge detection results in the post-contrast key frame image, the distance characteristics between the suspected calcification points and the edges are analyzed. Combining with the set of suspected blood vessels in the pre-contrast key frame image, a contrast coefficient is obtained to evaluate the change in the gray level of pixel points in the coronary artery images before and after the injection of the contrast agent, thereby preliminarily measuring whether there is a calcification condition and the degree of calcification, enhancing the recognition ability of calcification characteristics. Then, according to the temporal difference caused by the cardiac motion of the coronary artery blood vessels after the injection of the contrast agent, the difference characteristics of the gray level distribution between pixel points of all adjacent post-contrast key frame images are compared, and the matching index between pixel points is calculated to obtain the matching point set of each pixel point in the initial post-contrast key frame image of the cardiac cycle. The beneficial effect lies in analyzing the similarity between pixel points in adjacent post-contrast key frame images, facilitating the subsequent characterization of the displacement of each pixel point over time. By analyzing the matching index between adjacent elements in the matching point set and combining the position distribution of the elements in the matching point set, the relative displacement of each pixel point is obtained, and the suspected calcification distance is screened. The beneficial effect lies in identifying the displacement of the suspected calcification points over time, being able to identify the calcified area, and accurately calculating its position and range. Finally, based on the distribution of the suspected calcification distance, the cardiac fluctuation coefficient is calculated. Combining with the contrast coefficient, the calcified lesion area in the coronary angiography is identified, enhancing the recognition ability of calcified lesion characteristics in the coronary artery blood vessels and at the same time helping to enhance the comparative recognition ability of small calcified lesion areas.

[0075] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a part of code that contains one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in an order different from that marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. In the descriptions corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different blocks may also occur in an order different from that disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. Each block in the block diagram and / or flowchart, as well as combinations of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0076] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A system for identifying calcified nodule lesion areas in coronary angiography, characterized in that, The system includes: A coronary angiography image acquisition module, which is used to acquire dynamic coronary angiography images. Within one cardiac cycle before injecting the contrast agent, all the acquired frames of coronary angiography images are classified as pre-contrast key frame images; within one cardiac cycle after injecting the contrast agent, all the acquired frames of coronary angiography images are classified as post-contrast key frame images. A first coronary angiography image analysis module, which is used to analyze the gray value distribution of pixel points in the pre-contrast key frame images to obtain the high-density shadow set and the suspected blood vessel set of the pre-contrast key frame images; form a pair of the pre-contrast key frame images and the post-contrast key frame images with the same index within the cardiac cycle, analyze the gray value change of pixel points at the same position in each pair of images, and combine with the high-density shadow set to obtain the suspected calcification point set of the post-contrast key frame images; perform edge detection on each post-contrast key frame image, and according to the distance distribution of elements in the suspected calcification point set of the post-contrast key frame image to the edge in the image and the edge pixel point distribution, combine with the suspected blood vessel set of the pre-contrast key frame image to obtain the contrast coefficient. A second coronary angiography image analysis module, which is used to calculate the matching index between pixel points by sequentially comparing the difference features of the gray value distribution between all adjacent post-contrast key frame images, obtain the matching point set of each pixel point in the initial post-contrast key frame image of the cardiac cycle, analyze the matching index between adjacent elements in the matching point set, and combine with the position distribution of the elements in the matching point set to obtain the relative displacement of each pixel point, and screen to obtain the suspected calcification distance. A calcification nodule lesion area image recognition module, which is used to analyze the distribution characteristics and quantity characteristics of the suspected calcification distance to obtain the cardiac fluctuation coefficient, and combine with the contrast coefficient to identify the calcification lesion area in the coronary angiography.

2. The coronary angiography calcified nodule lesion area recognition system according to claim 1, characterized in that, The specific steps for obtaining the high-density shadow set and the suspected blood vessel set of the pre-contrast key frame images include: In each pre-contrast key frame image, obtain the set of all pixel point coordinates whose gray value is greater than the upper quartile of the gray values of all pixel points in the corresponding image, which is denoted as the high-density shadow set. Obtain the set of all pixel point coordinates whose gray value is less than the lower quartile of the gray values of all pixel points in the corresponding image, which is denoted as the suspected blood vessel set.

3. The calcified nodule lesion area recognition system for coronary angiography according to claim 1, characterized in that The specific process for obtaining the suspected calcification point set of the post-contrast key frame images is as follows: In each pair of images, calculate the gray value change amount of pixel points at the same position, and the set of pixel point coordinates whose gray value change amount is less than the lower quartile of all gray value change amounts in the corresponding image is denoted as the gray value micro-change point set. Take the intersection of the gray value micro-change point set of each pair of images and the high-density shadow set of the pre-contrast key frame image in each pair of images as the suspected calcification point set of the post-contrast key frame image in each pair of images.

4. A system for identifying calcified nodule lesion regions in coronary angiography according to claim 1, characterized in that, The specific method for obtaining the contrast coefficient is: For each post-contrast key frame image, according to the distance distribution of elements in the suspected calcification point set to the edge in the image, obtain the suspected calcification point distance distribution set of the post-contrast key frame image. Calculate the intersection over union of the set of suspected blood vessels in each pre-contrast key-frame image and the set composed of all edge pixel points in the post-contrast key-frame image with the same index; Obtain the dispersion of the set of suspected calcification point distance distributions in each post-contrast key-frame image; Fuse the negative correlation mapping results of the dispersions of all pre-contrast key-frame images with the intersection over union of all post-contrast key-frame images to obtain a contrast coefficient.

5. The coronary angiography calcification nodule lesion area recognition system according to claim 4, characterized in that, The process of obtaining the set of suspected calcification point distance distributions in the post-contrast key-frame image is as follows: In each post-contrast key-frame image, calculate the minimum distance between each element in the set of suspected calcification points and the edge line in the image, and use the set composed of the minimum distances as the set of suspected calcification point distance distributions in each post-contrast key-frame image.

6. The system for identifying the calcified nodule lesion area in coronary angiography according to claim 1, wherein, The calculation of the matching index between pixel points is specifically as follows: For each post-contrast key-frame image, obtain the gray values of the pixel points in the neighborhood of each element in the set of suspected calcification points to form a gray value sequence for each element; For a pixel point in each post-contrast key-frame image and a pixel point in the next adjacent post-contrast key-frame image, obtain the Euclidean distance between the coordinates of the two pixel points and the similarity measure between the gray value sequences corresponding to the two pixel points; fuse the negative correlation mapping of the Euclidean distance with the similarity measure to obtain the matching index between the two pixel points.

7. The system for identifying a calcified nodule lesion area in coronary angiography according to claim 1, characterized in that, The process of obtaining the set of matching points for each pixel point in the initial post-contrast key-frame image of the cardiac cycle is specifically as follows: For the pixel point corresponding to any element in the set of suspected calcification points in each post-contrast key-frame image, use the pixel point corresponding to the element with the largest matching index among the suspected calcification points in the next adjacent post-contrast key-frame image as the matching point for the pixel point corresponding to the said any element; For each pixel point in the initial contrast key-frame image of the cardiac cycle, form the set of matching points for each pixel point according to the transitive feature by all the matching points of each pixel point; The transitive feature is: for pixel point a, if pixel point b is the matching point of pixel point a and pixel point c is the matching point of pixel point b, then both pixel points b and c are the matching points of pixel point a.

8. The calcified nodule lesion area recognition system for coronary angiography according to claim 1, wherein The obtaining of the relative displacement of each pixel point is specifically as follows: ; where is the relative displacement of each pixel point, is the matching index between two pixel points belonging to the -th and -th contrast agent key frame images in the matching point set of each pixel point, is the distance weight of the -th matching index; where , is the number of elements in the matching point set, is the index of the matching index.

9. The calcified nodule lesion area recognition system for coronary angiography according to claim 1, characterized in that, The process of obtaining the suspected calcification distance is specifically as follows: Use the relative displacements of all pixel points in the initial post-contrast key-frame image of the cardiac cycle as input, adopt a threshold segmentation algorithm to obtain a segmentation threshold, and use the relative displacements less than the segmentation threshold as the suspected calcification distance.

10. A system for identifying calcified nodule lesion regions in coronary angiography according to claim 1, characterized in that, The process of obtaining the cardiac fluctuation coefficient and identifying the calcified lesion area in coronary angiography in combination with the contrast coefficient is as follows: Calculate the proportion of the number of suspected calcification distances in the total number of all relative displacements; divide by the mean value of all suspected calcification distances, and then divide by the variance of all suspected calcification distances; obtain the cardiac fluctuation coefficient; Obtain samples of a preset number of coronary angiography dynamic images, and respectively perform pre-segmentation on the contrast coefficients and cardiac fluctuation coefficients of all samples to obtain a contrast coefficient threshold and a cardiac fluctuation coefficient threshold; If the angiography coefficient of the dynamic coronary angiography image is greater than the angiography coefficient threshold and the cardiac fluctuation coefficient is greater than the cardiac fluctuation coefficient threshold, it is determined that there is a calcified lesion area in the dynamic coronary angiography image, and the lesion area is all pixel points corresponding to the distance from the suspected calcification.

Images