Intelligent recognition method for myocardial fibrosis area based on magnetic resonance image

Abstract

The application relates to the technical field of myocardial fibrosis region identification, in particular to a myocardial fibrosis region intelligent identification method based on magnetic resonance images. The method acquires multiple images of a cardiac magnetic resonance delayed gadolinium enhancement sequence; a last image is divided to obtain a subregion, a reference degree of the subregion is acquired according to the shape of the subregion and the signal value of the pixel points in the subregion; the reference signal value of the subregion in each image is acquired according to the reference degree and the signal value of the region corresponding to the subregion in each image; the fibrosis degree of the subregion is acquired according to the difference between the peak positions in the curve fitted by the reference signal value of the subregion and the overall signal value of each image, the difference between the reference signal value of the subregion in the last image and the overall signal value of the last image, and the myocardial fibrosis region is screened out. The fibrosis degree is accurately acquired, the myocardial fibrosis region is accurately screened out, and the accuracy of myocardial fibrosis region acquisition is effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of myocardial fibrosis region identification, and particularly relates to a myocardial fibrosis region intelligent identification method based on magnetic resonance images. BACKGROUND

[0002] Myocardial fibrosis is the core pathological change of cardiovascular diseases such as coronary heart disease, dilated cardiomyopathy and hypertensive heart disease, and its characteristic is the abnormal deposition of collagen in myocardial tissue, which leads to significant increase in systolic and diastolic dysfunction and arrhythmia risk. Cardiac magnetic resonance delayed gadolinium enhancement imaging (LGE) is currently the gold standard for clinical evaluation of myocardial fibrosis, which makes the lesion tissue show high signal in T1 weighted sequence through the retention effect of gadolinium contrast agent in the fibrosis region. However, the interpretation of LGE images is highly dependent on the experience of doctors, and there are subjective bias, diffuse fibrosis or micro-focal lesions that are difficult to identify by naked eye, and lack of automated tools to accurately calculate fibrosis load.

[0003] In the existing intelligent method, a linear iterative clustering algorithm is used to perform superpixel segmentation on the cardiac magnetic resonance delayed gadolinium enhancement image, and the pixels are aggregated into homogeneous blocks to divide the myocardial fibrosis region. However, in actual situations, the shape of the segmented region is affected by the distribution of the initial seed points and will generate irregular polygons (such as star-shaped, jagged edges), while the real myocardial fibrosis region is in the form of strips or patches, which leads to distortion of morphological feature extraction and cannot accurately identify the myocardial fibrosis region. SUMMARY

[0004] In order to solve the technical problem that the linear iterative clustering algorithm generates irregular polygons due to the influence of the initial seed point distribution on the shape of the segmented region, leading to distortion of morphological feature extraction and inability to accurately identify the myocardial fibrosis region, the purpose of the present application is to provide a myocardial fibrosis region intelligent identification method based on magnetic resonance images, and the technical solution adopted is as follows:

[0005] The embodiment of the present application provides a myocardial fibrosis region intelligent identification method based on magnetic resonance images, which comprises the following steps:

[0006] Obtaining multiple frames of images of a cardiac magnetic resonance delayed gadolinium enhancement sequence;

[0007] Dividing the last frame of image to obtain sub-regions, and obtaining the reference degree of each sub-region according to the shape of each sub-region and the distribution of signal values of the pixel points in each sub-region;

[0008] According to the reference degree of each sub-region and the signal value of the pixel points in the corresponding region of each sub-region in each frame of image, the reference signal value of each sub-region in each frame of image is obtained; the reference signal value of each sub-region and the overall signal value of each frame of image are fitted into curves according to the time sequence of the corresponding image, the difference between the peak value corresponding position of the curve of each sub-region and the overall signal value curve, and the difference between the reference signal value of each sub-region in the last frame of image and the overall signal value of the last frame of image, and the fibrosis degree of each sub-region is obtained;

[0009] Based on the fibrosis degree, the myocardial fibrosis region is screened out.

[0010] Further, the reference degree obtaining method is:

[0011] For any sub-region, the long axis and the short axis of the sub-region are obtained through the minimum circumscribed matrix of the sub-region.

[0012] When the ratio of the long axis to the short axis of the sub-region meets the preset long-to-short axis ratio range, the shape matching degree of the sub-region is set to 1.

[0013] When the ratio of the long axis to the short axis of the sub-region does not meet the preset long-to-short axis ratio range, the shape matching degree of the sub-region is set to 0.

[0014] The result of negative correlation and normalization of the variance of the signal values of all pixel points in the sub-region is taken as the signal stability degree of the sub-region.

[0015] The product of the shape matching degree and the signal stability degree of the sub-region is taken as the reference degree of the sub-region.

[0016] Further, the reference signal value obtaining method is:

[0017] For any sub-region and any frame of image, the mean value of the signal values of the pixel points in the corresponding region of the sub-region in the frame of image is taken as the initial signal value of the sub-region in the frame of image.

[0018] The product of the initial signal value and the reference degree of the sub-region is taken as the reference signal value of the sub-region in the frame of image.

[0019] Further, the overall signal value obtaining method is:

[0020] The mean value of the signal values of all pixel points in each frame of image is taken as the overall signal value of each frame of image.

[0021] Further, the fibrosis degree obtaining method is:

[0022] According to the difference between the peak corresponding position of the curve corresponding to each sub-region and the overall signal value curve, the peak lag degree of each sub-region is obtained;

[0023] According to the difference between the reference signal value of each sub-region in the last frame of image and the overall signal value of the last frame of image, the signal analysis value of each sub-region is obtained.

[0024] The product of the peak lag degree and the signal analysis value of each sub-region is taken as the fibrosis degree of each sub-region.

[0025] Further, the peak lag degree is obtained by:

[0026] All the frames of images are arranged according to the obtained time sequence to obtain an image sequence;

[0027] The images in the image sequence are sequentially labeled from small to large in the order from left to right to obtain the label of each frame of image;

[0028] For any sub-region, the label of the image corresponding to the peak in the curve corresponding to the sub-region is taken as the peak label of the sub-region;

[0029] The label of the image corresponding to the peak in the overall signal value curve is taken as the reference peak label;

[0030] The difference between the peak label of the sub-region and the reference peak label is normalized to obtain the peak lag degree of the sub-region.

[0031] Further, the signal analysis value is obtained by:

[0032] For any sub-region, the difference between the reference signal value of the sub-region in the last frame of image and the overall signal value of the last frame of image is normalized to obtain the signal analysis value of the sub-region.

[0033] Further, the myocardial fibrosis region is obtained by:

[0034] When the fibrosis degree is greater than a preset fibrosis degree threshold, the corresponding sub-region is taken as the myocardial fibrosis region.

[0035] Further, the sub-region is obtained by:

[0036] Each region after segmentation is taken as a sub-region by performing superpixel segmentation on the last frame of image through a linear iterative clustering algorithm.

[0037] Further, the region corresponding to each sub-region in each frame of image is obtained by:

[0038] For any sub-region and any frame image, the region corresponding to the sub-region in the last frame image is completely mapped to the region corresponding to the sub-region in the frame image as the region corresponding to the sub-region in the frame image.

[0039] The present application has the following advantages:

[0040] The present application first divides the last frame image to obtain sub-regions, which is conducive to more accurate and efficient subsequent acquisition of myocardial fibrosis regions; then, according to the shape of each sub-region and the distribution of signal values of the pixels in the sub-region, the reference degree of each sub-region is obtained, which preliminarily reflects the possibility of each sub-region being a myocardial fibrosis region, directly determines the reference significance of each sub-region, and is conducive to more efficient subsequent analysis of the possibility of each sub-region being a myocardial fibrosis region; further, according to the reference degree of each sub-region and the signal values of the pixels in the region corresponding to each sub-region in each frame image, the reference signal value of each sub-region in each frame image is obtained, which accurately reflects the signal intensity of each sub-region in each frame image and prepares for subsequent accurate analysis of whether each sub-region is a myocardial fibrosis region; then, the reference signal value of each sub-region and the overall signal value of each frame image are fitted into curves according to the time sequence of the corresponding images, which accurately analyzes the possibility of each sub-region being a myocardial fibrosis region; therefore, according to the difference between the peak value corresponding position of the curve corresponding to each sub-region and the overall signal value curve, and the difference between the reference signal value of each sub-region in the last frame image and the overall signal value of the last frame image, the fibrosis degree of each sub-region is obtained, which accurately reflects the possibility of each sub-region being a myocardial fibrosis region; then, based on the fibrosis degree, the myocardial fibrosis region is accurately screened, which effectively improves the accuracy of acquiring the myocardial fibrosis region and is conducive to the accurate analysis of the condition of patients with cardiovascular diseases by doctors. BRIEF DESCRIPTION OF DRAWINGS

[0041] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0042] Figure 1 A schematic flow chart of a myocardial fibrosis region intelligent identification method based on magnetic resonance images provided by an embodiment of the present application;

[0043] Figure 2 A fibrosis degree acquisition method flow chart provided by an embodiment of the present application;

[0044] Figure 3A structural diagram of a myocardial fibrosis region intelligent identification system based on magnetic resonance images provided by an embodiment of the present application;

[0045] Figure 4 A schematic diagram of a computer device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0046] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined purposes, the myocardial fibrosis region intelligent identification method based on magnetic resonance images according to the present application, its specific implementation, structure, features and effects are described in detail as follows in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0048] The specific scheme of the myocardial fibrosis region intelligent identification method based on magnetic resonance images provided by the present application is specifically described below in combination with the drawings.

[0049] Embodiment 1:

[0050] The present application proposes a myocardial fibrosis region intelligent identification method based on magnetic resonance images, please refer to Figure 1 , which shows a schematic flowchart of a myocardial fibrosis region intelligent identification method based on magnetic resonance images provided by an embodiment of the present application. The method comprises the following steps:

[0051] Step S1: Obtain multiple images of cardiac magnetic resonance delayed gadolinium enhancement sequence.

[0052] Specifically, the present embodiment takes a patient with cardiovascular disease as an example for analysis, uses a clinical magnetic resonance imaging device and adopts a delayed gadolinium enhancement imaging sequence to obtain multiple images of cardiac magnetic resonance delayed gadolinium enhancement sequence of the patient. By observing the myocardial fibrosis region in the images, the core pathology of the patient is analyzed. It should be noted that the delayed gadolinium enhancement imaging sequence is adopted in the present embodiment because the sequence has high sensitivity to myocardial fibrosis tissue, can make normal myocardial tissue and myocardial fibrosis tissue show different characteristics, and use the performance characteristics of myocardial fibrosis tissue in multiple images, combined with data processing method, which is beneficial to accurately quantifying the degree of regional fibrosis. The delayed gadolinium enhancement imaging sequence is a known technology and will not be described in detail.

[0053] It should be noted that when the plurality of images of the cardiac magnetic resonance delayed gadolinium enhancement sequence is acquired, the acquisition is performed in time sequence, and the time interval between the two adjacent images is set to 20 ms in this embodiment, and the time interval between the two adjacent images can be set according to actual conditions, which is not limited herein.

[0054] In order to accurately obtain the myocardial fibrosis region in the image subsequently and avoid the interference of noise, the image obtained is preprocessed in this embodiment, specifically, first, a median filtering algorithm is used to perform denoising processing on the image, so as to reduce random noise caused by factors such as non-uniform magnetic field or electronic noise in the scanning process; second, a histogram equalization algorithm is used to enhance the contrast of the image, so that the signal difference between the normal myocardial tissue and the myocardial fibrosis tissue is more obvious, which is beneficial to more accurate division into the myocardial fibrosis region. The median filtering algorithm and the histogram equalization algorithm are both known technologies, and will not be described in detail.

[0055] It should be noted that the subsequent images are all preprocessed images.

[0056] Step S2: The last image is divided to obtain sub-regions, and the reference degree of each sub-region is obtained according to the shape of each sub-region and the distribution of the signal values of the pixel points in the sub-region.

[0057] It is known that the signal corresponding to the normal myocardial tissue in the last image among all the images is in a normal state, and the signal corresponding to the myocardial fibrosis tissue is in a high active state, that is, the difference between the normal myocardial tissue and the myocardial fibrosis tissue in the last image is most obvious, therefore, the linear iterative clustering algorithm is used to perform superpixel segmentation on the last image in this embodiment, and the last image is divided to obtain sub-regions. The specific obtaining process of the sub-region is as follows: first, the linear iterative clustering algorithm is used to uniformly determine the initial seed points in the last image, wherein each initial seed point carries Lab color value and two-dimensional coordinate information; then the distance between each pixel point in the last image and the adjacent initial seed point in the 5-dimensional space is calculated, and the pixel point is assigned to the superpixel block to which the initial seed point belongs; further, the Lab mean value and the coordinate mean value in each superpixel block are recalculated, the seed point position is updated, and the iteration is repeated until the seed point offset is less than 1 pixel; finally, the signal intensity difference and the texture similarity of adjacent superpixel blocks are calculated, if the signal intensity difference satisfies the preset signal difference range and the texture similarity satisfies the preset texture similarity range, the adjacent superpixel blocks are merged, and the final each sub-region is determined. The linear iterative clustering algorithm is a known calculation, which will not be described in detail.

[0058] In actual situations, the shape of the sub-region obtained by linear iterative clustering algorithm is affected by the distribution of the initial seed points, and an irregular polygon is generated, which deviates from the real myocardial fibrosis region. In order to improve the accuracy of subsequent accurate acquisition of myocardial fibrosis region, therefore, the shape of each sub-region needs to be analyzed to preliminarily determine the possibility of each sub-region being a myocardial fibrosis region. It is known that the myocardial fibrosis region is a cord-like structure and the signal intensity inside is consistent, therefore, according to the shape of each sub-region and the distribution of the signal value of the pixel points inside, the reference degree of each sub-region is obtained. The greater the reference degree, the more likely the corresponding sub-region is a myocardial fibrosis region.

[0059] Preferably, in an implementable manner of the embodiment, the reference degree is obtained by: for any sub-region, obtaining the major axis and the minor axis of the sub-region through the minimum circumscribed matrix of the sub-region; wherein the minimum circumscribed matrix and the method of obtaining the major axis and the minor axis through the minimum circumscribed matrix are both known technologies, and will not be described in detail. It is known that the ratio of the major axis and the minor axis of the myocardial fibrosis region is usually 5 to 15, therefore, the embodiment sets the preset ratio range of the major axis and the minor axis as , and the implementer can set the preset ratio range of the major axis and the minor axis according to the actual situation, which is not limited here. When the ratio of the major axis and the minor axis of the sub-region meets the preset ratio range of the major axis and the minor axis, that is, is in , it means that the shape of the sub-region is more consistent with the myocardial fibrosis region, and the embodiment sets the shape matching degree of the sub-region as 1; when the ratio of the major axis and the minor axis of the sub-region does not meet the preset ratio range of the major axis and the minor axis, it means that the sub-region is less likely to be a myocardial fibrosis region, and the embodiment sets the shape matching degree of the sub-region as 0; further, the result of the negative correlation and normalization of the variance of the signal value of all pixel points in the sub-region is obtained as the signal stability degree of the sub-region; the greater the signal stability degree, the higher the accuracy of the analysis of the sub-region. The embodiment takes the inverse of the variance of the signal value as the power of an exponential function with a natural constant as the base, and the output result of the exponential function is the result of the negative correlation and normalization of the variance of the signal value. In order to represent the possibility of the sub-region being a myocardial fibrosis region, the product of the shape matching degree and the signal stability degree of the sub-region is taken as the reference degree of the sub-region.

[0060] So far, the reference degree of each sub-region is obtained, which is beneficial to the subsequent more accurate analysis of the possibility of each sub-region being a myocardial fibrosis region.

[0061] Step S3: According to the reference degree of each sub-region and the signal value of the pixel points in the corresponding region of each sub-region in each frame image, the reference signal value of each sub-region in each frame image is obtained; the reference signal value of each sub-region and the overall signal value of each frame image are fitted into curves according to the time sequence of the corresponding images, and the fibrosis degree of each sub-region is obtained according to the difference between the peak value corresponding positions of the curve corresponding to each sub-region and the overall signal value curve and the difference between the reference signal value of each sub-region in the last frame image and the overall signal value of the last frame image.

[0062] Specifically, the frame image sequence number corresponding to the signal peak value of the myocardial fibrosis region is lagging behind compared with the normal myocardial region, so the embodiment first obtains the reference signal value of each sub-region in each frame image according to the reference degree of each sub-region and the signal value of the pixel points in the corresponding region of each sub-region in each frame image, and accurately reflects the signal intensity of each sub-region in each frame image. Considering that in actual situations the myocardial fibrosis region is only a small part in the image, and the vast majority of the region in the image is the normal myocardial region, the frame image sequence number corresponding to the signal peak value of the normal myocardial region should be close to the frame image sequence number corresponding to the peak value of the overall signal value of each frame image, and the frame image sequence number corresponding to the signal peak value of the myocardial fibrosis region should lag behind the frame image sequence number corresponding to the peak value of the overall signal value of each frame image. In order to more accurately analyze the possibility of each sub-region being a myocardial fibrosis region, the reference signal value of each sub-region and the overall signal value of each frame image are fitted into curves according to the time sequence of the corresponding images, for example, the reference signal value of a certain sub-region is arranged according to the time sequence of the corresponding images from front to back and fitted into a curve. The method of fitting the curve is a known technology and will not be described in detail. Then, according to the difference between the peak value corresponding positions of the curve corresponding to each sub-region and the overall signal value curve, the lagging situation of the peak value of the reference signal value of each sub-region is analyzed, which is conducive to more accurately analyzing the possibility of each sub-region being a myocardial fibrosis region.

[0063] Another aspect considers the slow-in and slow-out characteristics of myocardial fibrosis tissue, so in the last frame image the myocardial fibrosis region still shows high signal, while the normal myocardial region has returned to normal signal, and then the embodiment further combines the difference between the reference signal value of each sub-region in the last frame image and the overall signal value of the last frame image to further determine the possibility of each sub-region being a myocardial fibrosis region. Therefore, according to the difference between the peak value corresponding positions of the curve corresponding to each sub-region and the overall signal value curve and the difference between the reference signal value of each sub-region in the last frame image and the overall signal value of the last frame image, the fibrosis degree of each sub-region is obtained. The greater the fibrosis degree, the more likely the corresponding sub-region is a myocardial fibrosis region.

[0064] Preferably, in one implementable manner of the present embodiment, the acquisition method of the region corresponding to each sub-region in each frame image is that, for any sub-region and any frame image, the region corresponding to the sub-region in the last frame image is completely mapped to the region corresponding to the sub-region in the frame image as the region corresponding to the sub-region in the frame image.

[0065] Preferably, in one implementable manner of the present embodiment, the acquisition method of the reference signal value is that, for any sub-region and any frame image, the mean value of the signal values of the pixels in the region corresponding to the sub-region in the frame image is taken as the initial signal value of the sub-region in the frame image; considering that the greater the reference degree of the sub-region is, the more likely the sub-region is a myocardial fibrosis region, which indirectly indicates that the signal values in the region corresponding to the sub-region in each frame image are more meaningful for reference, and then the product of the initial signal value and the reference degree of the sub-region is taken as the reference signal value of the sub-region in the frame image.

[0066] Up to now, the reference signal value of each sub-region in each frame image is acquired, which is beneficial to subsequent accurate analysis of whether each sub-region meets the characteristics of the myocardial fibrosis region.

[0067] Preferably, in one implementable manner of the present embodiment, the acquisition method of the overall signal value is that the mean value of the signal values of all the pixels in each frame image is taken as the overall signal value of the frame image.

[0068] Preferably, in one implementable manner of the present embodiment, the acquisition method of the fibrosis degree please refer to Figure 2 which shows a flowchart of the acquisition method of the fibrosis degree provided by the present embodiment, and the method comprises the following steps:

[0069] Step S201: acquiring the peak lag degree of each sub-region according to the difference between the peak value corresponding positions of the curve corresponding to each sub-region and the overall signal value curve.

[0070] The greater the peak lag degree is, the more likely the corresponding sub-region is a myocardial fibrosis region.

[0071] In an implementable manner of the embodiment, the peak lag degree obtaining method is: arranging all the frame images according to the obtained time sequence to obtain an image sequence; then sequentially labeling the images in the image sequence from left to right in ascending order to obtain the label of each frame image. The label sequence of the embodiment is set as 1, 2, 3… For any sub-region, the label of the image corresponding to the peak in the curve corresponding to the sub-region is taken as the peak label of the sub-region; the label of the image corresponding to the peak in the overall signal value curve is taken as the reference peak label; it should be noted that if there are multiple peaks in a curve, the largest peak is taken as the main peak, and if there are multiple largest peaks, the first appearing largest peak is taken as the main peak. The greater the peak label of the sub-region is greater than the reference peak label, the greater the signal peak lag degree of the sub-region is, and the more likely the sub-region is a myocardial fibrosis region. Further, the embodiment takes the normalized result of the difference between the peak label of the sub-region and the reference peak label as the peak lag degree of the sub-region. The embodiment normalizes the difference between the peak label of the sub-region and the reference peak label by using a norm normalization function.

[0072] At this point, the peak lag degree of each sub-region is obtained.

[0073] Step S202: According to the difference between the reference signal value of each sub-region in the last frame image and the overall signal value of the last frame image, the signal analysis value of each sub-region is obtained.

[0074] The greater the signal analysis value is, the greater the signal intensity of the corresponding sub-region in the last frame image is, and the more likely the corresponding sub-region is a myocardial fibrosis region.

[0075] In an implementable manner of the embodiment, the signal analysis value obtaining method is: for any sub-region, the normalized result of the difference between the reference signal value of the sub-region in the last frame image and the overall signal value of the last frame image is taken as the signal analysis value of the sub-region. The embodiment normalizes the difference between the reference signal value of the sub-region and the overall signal value of the last frame image by using a norm normalization function.

[0076] At this point, the signal analysis value of each sub-region is obtained.

[0077] Step S203: The product of the peak lag degree and the signal analysis value of each sub-region is taken as the fibrosis degree of each sub-region.

[0078] It is known that the greater the peak lag degree and the signal analysis value of a certain sub-region are, the more likely the sub-region is a myocardial fibrosis region. Therefore, the product of the peak lag degree and the signal analysis value of each sub-region is taken as the fibrosis degree of each sub-region.

[0079] Step S4: screening out the myocardial fibrosis region based on the fibrosis degree.

[0080] It is known that the greater the fibrosis degree is, the more likely the corresponding sub-region is the myocardial fibrosis region, and thus the embodiment screens out the myocardial fibrosis region based on the fibrosis degree.

[0081] Preferably, in an implementable manner of the embodiment, the method for obtaining the myocardial fibrosis region is that: the embodiment sets the preset fibrosis degree threshold value as 0.7, and the implementer can set the size of the preset fibrosis degree threshold value according to the actual situation, which is not limited herein. When the fibrosis degree is greater than the preset fibrosis degree threshold value, the corresponding sub-region is taken as the myocardial fibrosis region. Thus, the myocardial fibrosis region is accurately screened out.

[0082] Finally, the screened myocardial fibrosis region is sequentially subjected to open operation and close operation to smooth the edge of the myocardial fibrosis region, so that the myocardial fibrosis region is more accurate; then the myocardial fibrosis region after smoothing processing is marked in the original image to generate a visual result, which is beneficial to the doctor to more accurately analyze the patient's condition. The open operation and the close operation are both known technologies, and will not be described in detail.

[0083] To sum up, the embodiment obtains multiple images of a cardiac magnetic resonance delayed gadolinium enhancement sequence; divides the last image to obtain a sub-region, obtains a reference degree of the sub-region according to the shape of the sub-region and the signal value of the pixel points in the sub-region; obtains a reference signal value of the sub-region in each image according to the reference degree and the signal value of the region corresponding to the sub-region in each image; obtains a fibrosis degree of the sub-region according to the difference between the peak positions in the curve fitted by the reference signal value of the sub-region and the overall signal value of each image, and the difference between the reference signal value of the sub-region in the last image and the overall signal value of the last image, and screens out a myocardial fibrosis region. The present application accurately obtains the fibrosis degree, so that the myocardial fibrosis region is accurately screened out, and the accuracy of obtaining the myocardial fibrosis region is effectively improved.

[0084] Embodiment 2:

[0085] The present application also provides a myocardial fibrosis region intelligent identification system based on magnetic resonance images, please refer to Figure 3 which shows a myocardial fibrosis region intelligent identification system structure diagram provided by an embodiment of the present application, the system comprises: an image acquisition module 10, a reference degree acquisition module 20, a fibrosis degree acquisition module 30 and a myocardial fibrosis region acquisition module 40.

[0086] The image acquisition module 10 is used for obtaining multiple images of a cardiac magnetic resonance delayed gadolinium enhancement sequence.

[0087] The reference degree acquisition module 20 is configured to divide the last frame of image to obtain sub-regions, and acquire the reference degree of each sub-region according to the shape of each sub-region and the distribution of the signal values of the pixel points in each sub-region.

[0088] The fibrosis degree acquisition module 30 is configured to acquire the reference signal value of each sub-region in each frame of image according to the reference degree of each sub-region and the signal values of the pixel points in the region corresponding to each sub-region in each frame of image, fit the reference signal value of each sub-region and the overall signal value of each frame of image into a curve according to the time sequence of the corresponding image, and acquire the fibrosis degree of each sub-region according to the difference between the peak value corresponding positions of the curve corresponding to each sub-region and the overall signal value curve, and the difference between the reference signal value of each sub-region in the last frame of image and the overall signal value of the last frame of image.

[0089] The myocardial fibrosis region acquisition module 40 is configured to screen out the myocardial fibrosis region based on the fibrosis degree.

[0090] It should be noted that the system provided in the above embodiment is only exemplified by the division of the above functional modules, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the myocardial fibrosis region intelligent identification system based on magnetic resonance images and the myocardial fibrosis region intelligent identification method based on magnetic resonance images provided in the above embodiment belong to the same concept, and the specific implementation process is described in detail in the method embodiment, which will not be repeated here.

[0091] Embodiment 3:

[0092] The application further provides a myocardial fibrosis region intelligent identification device based on magnetic resonance images. The device comprises a memory and a processor. The memory stores executable program codes. The processor is configured to call and execute the executable program codes to execute the myocardial fibrosis region intelligent identification method based on magnetic resonance images provided in the embodiments. The device can be a chip, an assembly or a module. The chip can comprise a processor and a memory connected to each other. The memory is configured to store instructions. When the processor calls and executes the instructions, the chip can execute the myocardial fibrosis region intelligent identification method based on magnetic resonance images provided in the above embodiments.

[0093] In addition, the present embodiment also protects a computer device. Please refer to Figure 4The computer device comprises a memory 401, a processor 402, and a computer program 403 stored in the memory 401 and running on the processor 402, wherein the processor 402 executes the computer program 403, so that the computer device can execute any one of the aforementioned intelligent identification methods of myocardial fibrosis area based on magnetic resonance images.

[0094] Embodiment 4:

[0095] The application further provides a computer readable storage medium, which stores computer program codes, and when the computer program codes are run on a computer, the computer executes the above-mentioned related method steps to realize the intelligent identification method of myocardial fibrosis area based on magnetic resonance images provided by the above-mentioned embodiments.

[0096] Embodiment 5:

[0097] The application further provides a computer program product, and when the computer program product is run on a computer, the computer executes the above-mentioned related steps to realize the intelligent identification method of myocardial fibrosis area based on magnetic resonance images provided by the above-mentioned embodiments.

[0098] Wherein, the device, the computer readable storage medium, the computer program product or the chip provided by the embodiment are used to execute the corresponding method provided above, so the beneficial effects that can be achieved are referable to the beneficial effects in the corresponding method provided above, which will not be repeated here.

[0099] It should be noted that: the above-mentioned sequence of the embodiments of the application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0100] Each embodiment in the specification is described in a progressive manner, and the same and similar parts between each embodiment can be referred to each other, and each embodiment mainly describes the difference from other embodiments.

Claims

1. A method for intelligent recognition of myocardial fibrosis regions based on magnetic resonance images, characterized in that, The method comprises the following steps: Obtaining multiple images of a cardiac magnetic resonance delayed gadolinium enhancement sequence; Dividing the last image to obtain sub-regions, and obtaining a reference degree of each sub-region according to the shape of each sub-region and the distribution of signal values of pixels in each sub-region; Obtaining a reference signal value of each sub-region in each image according to the reference degree of each sub-region and the signal values of pixels in the region corresponding to each sub-region in each image; fitting the reference signal value of each sub-region and the overall signal value of each image into a curve according to the time sequence of the corresponding image, obtaining the fibrosis degree of each sub-region according to the difference between the peak value corresponding positions of the curve corresponding to each sub-region and the overall signal value curve, and the difference between the reference signal value of each sub-region in the last image and the overall signal value of the last image; Screening a myocardial fibrosis region based on the fibrosis degree; The reference degree is obtained by: For any sub-region, obtaining the long axis and the short axis of the sub-region through the minimum circumscribed matrix of the sub-region; When the ratio of the long axis to the short axis of the sub-region meets a preset long-to-short axis ratio range, setting the shape matching degree of the sub-region as 1; When the ratio of the long axis to the short axis of the sub-region does not meet the preset long-to-short axis ratio range, setting the shape matching degree of the sub-region as 0; The result of negatively correlating and normalizing the variance of the signal values of all pixels in the sub-region is taken as the signal stability degree of the sub-region; The product of the shape matching degree and the signal stability degree of the sub-region is taken as the reference degree of the sub-region; The reference signal value is obtained by: For any sub-region and any image, the mean value of the signal values of the pixels in the region corresponding to the sub-region in the image is taken as the initial signal value of the sub-region in the image; The product of the initial signal value and the reference degree of the sub-region is taken as the reference signal value of the sub-region in the image; The overall signal value is obtained by: The mean value of the signal values of all pixels in each image is taken as the overall signal value of each image. 2.The method of claim 1, wherein, The fibrosis degree is obtained by: The peak value lag degree of each sub-region is obtained according to the difference between the peak value corresponding positions of the curve corresponding to each sub-region and the overall signal value curve; The signal analysis value of each sub-region is obtained according to the difference between the reference signal value of each sub-region in the last image and the overall signal value of the last image; The product of the peak value lag degree and the signal analysis value of each sub-region is taken as the fibrosis degree of each sub-region. 3.The method of claim 2, wherein, The peak value lag degree is obtained by: Arranging all the images according to the obtained time sequence to obtain an image sequence; Labeling the images in the image sequence from small to large in order from left to right to obtain the label of each image; For any sub-region, the label of the image corresponding to the peak value in the curve corresponding to the sub-region is taken as the peak label of the sub-region; The label of the image corresponding to the peak value in the overall signal value curve is taken as the reference peak label; The result of normalizing the difference between the peak label of the sub-region and the reference peak label as the peak lag degree of the sub-region. 4.The method of claim 2, wherein, The method for obtaining the signal analysis value is: For any sub-region, the result of normalizing the difference between the reference signal value of the sub-region in the last frame of image and the overall signal value of the last frame of image as the signal analysis value of the sub-region. 5.The method of claim 1, wherein, The method for obtaining the myocardial fibrosis region is: When the fibrosis degree is greater than a preset fibrosis degree threshold, the corresponding sub-region is taken as the myocardial fibrosis region. 6.The method of claim 1, wherein, The method for obtaining the sub-region is: Each region obtained by performing superpixel segmentation on the last frame of image by a linear iterative clustering algorithm is taken as a sub-region. 7.The method of claim 1, wherein, The method for obtaining the region corresponding to each sub-region in each frame of image is: For any sub-region and any frame of image, the region corresponding to the sub-region in the last frame of image is completely mapped to the region corresponding to the sub-region in the frame of image, as the region corresponding to the sub-region in the frame of image.

Images