Myocardial infarction image segmentation method based on data processing

Abstract

This invention discloses a method for segmenting myocardial infarction images based on data processing. This invention relates to the field of medical image processing technology and addresses the problem that existing methods do not fully consider individual differences in myocardial infarction images, leading to low segmentation accuracy due to mismatched segmentation parameters. This invention performs differentiated pre-segmentation processing based on the initial feature parameters of the myocardial infarction image, and adapts and adjusts the edge detection gradient by combining the feature trend intervals of historical segmentation data to restore the true feature distribution of the image. This avoids missed detections and over-segmentation caused by mismatched initial segmentation parameters, ensuring the pre-segmentation effect while laying the foundation for subsequent accurate segmentation. Based on historical segmentation data and similar labeled data, a stable feature interval of the image is determined. This interval corresponds to the pixel range where the lesion region features are most stable in the image. Segmentation matching based on the feature values ​​of this interval can accurately locate the core region of the myocardial infarction lesion, improving the targeting and accuracy of segmentation.

Description

Technical Field

[0001] This invention relates to the field of medical image processing technology, specifically to a method for segmenting myocardial infarction images based on data processing. Background Technology

[0002] In the clinical diagnosis of myocardial infarction, accurate segmentation of medical images (such as cardiac MRI and CT images) is a core element for lesion localization and disease assessment. By segmenting the contour, area, and texture features of the myocardial infarction region, quantitative evidence can be provided for the formulation of clinical treatment plans. Current myocardial infarction image segmentation methods mostly use preset segmentation algorithms and fixed thresholds to process images uniformly, failing to fully consider the imaging characteristics of different patients (such as tissue grayscale differences, lesion edge blurring, and noise interference levels) and the differences in imaging parameters of different scanning devices. This easily leads to problems such as missed detections, oversegmentation, or edge localization deviations, resulting in insufficient accuracy of segmentation results. Consequently, reliable quantitative data cannot be provided for clinical practice, affecting the diagnosis and treatment outcomes of myocardial infarction.

[0003] For the segmentation process of myocardial infarction images, existing technologies are mostly based on common image segmentation standards to perform segmentation operations, sequentially completing steps such as feature extraction, threshold screening, and region segmentation. However, this processing method does not fully consider the individual characteristics of different myocardial infarction images. Due to the mismatch of segmentation parameters, the accuracy of image segmentation results will be reduced, which in turn will affect the clinical judgment of myocardial infarction. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a data processing-based image segmentation method for myocardial infarction, which solves the problem that existing methods do not fully consider individual differences in myocardial infarction images and result in low segmentation accuracy due to mismatched segmentation parameters.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a myocardial infarction image segmentation method based on data processing, comprising the following steps:

[0006] Initial feature monitoring of images: Initial feature monitoring is performed on the myocardial infarction images to be segmented to obtain initial feature parameters such as the initial gray-scale mean, the edge gradient of the suspected lesion area, and the image noise value.

[0007] Image pre-segmentation: Based on the monitored initial feature parameters, the myocardial infarction images are pre-segmented, and different pre-segmentation methods are executed based on the specific values ​​of the initial feature parameters.

[0008] Constant feature segmentation control: Based on historical segmentation data of myocardial infarction images and labeled data of similar lesion images, the image feature change curve is identified, and the segmentation process in this stage is controlled based on the feature change curve to achieve accurate segmentation of the lesion area;

[0009] Constant threshold fine segmentation: A constant segmentation threshold is used to refine the image regions after pre-segmentation and constant feature segmentation. When the matching degree of the segmented regions meets the standard, the entire segmentation process of the myocardial infarction image is completed.

[0010] Preferably, the step of performing different pre-segmentation methods based on the specific values ​​of the initial feature parameters specifically includes: labeling the initial grayscale mean as G, the image noise value as N, setting the preset grayscale threshold as G0, the noise threshold as N0, ΔG as the preset grayscale tolerance value, and K as the standard edge detection gradient.

[0011] If G∈[G0-ΔG,G0+ΔG] and N≤N0, then a constant edge detection gradient of 0.1K is used to pre-segment the myocardial infarction image. The pixel traversal range of the pre-segmentation is S, and the pre-segmentation ends, where S is a preset value.

[0012] If G∉[G0-ΔG,G0+ΔG] or N>N0, obtain relevant feature data of similar myocardial infarction images in the corresponding gray / noise interval from historical segmentation data. Based on the image pixel coordinates, confirm the correlation feature change curve of its relevant feature data. The horizontal axis of the curve is the pixel horizontal coordinate, and the vertical axis is the feature parameter. Sequentially confirm the feature trend value T of adjacent pixels within the correlation feature change curve. i Its trend value T i = (Vertical coordinate difference between adjacent pixels) ÷ Horizontal coordinate difference between adjacent pixels, where the coordinate difference is the coordinate parameter of the next pixel minus the coordinate parameter of the previous pixel, and the identified several feature trend values ​​T are then used. i After performing absolute value processing, the trend interval [T] is determined. min T max ], and T is the trend value after absolute value processing, where i represents the line segment between different adjacent pixels;

[0013] A constant edge detection gradient of 0.06K is preferentially used for pre-segmentation of myocardial infarction images, and the feature change trend F of the images is monitored in real time during the segmentation process. If F∈[T] min T max If F < T, then maintain a constant edge detection gradient for pre-segmentation; min Then, the edge detection gradient is gradually increased up to a maximum of 0.1K, so that F∈[T] min T max [Time ends; if F > T] max Then, gradually reduce the edge detection gradient to 0.02K, so that F∈[T] min T max [Time limit]; When F cannot be adjusted to fall within this trend range, an abnormal image signal is directly generated for display;

[0014] When the clarity of the outline of the suspected lesion area in the pre-segmented image is greater than or equal to the preset clarity threshold, the pre-segmentation ends and constant feature segmentation control is executed.

[0015] Preferably, the constant feature segmentation control includes three sub-steps: feature stability interval confirmation, feature matching and segmentation, and real-time feature adjustment, specifically:

[0016] Feature stability interval confirmation: From historical segmentation data of myocardial infarction images and labeled data of similar lesion images, relevant feature data with image grayscale values ​​between G1 and G2 were confirmed. G1 is the lower limit of grayscale value of normal myocardial tissue, and G2 is the upper limit of grayscale value of infarct lesions. The minimum value F of the feature parameter was determined from this relevant feature data. min and the maximum value of the characteristic parameter F max And determine the characteristic interval [F] min F max Based on the relevant feature data corresponding to different pixels, and according to the spatial relationship of the pixels, a curve showing the change of their relevant features is generated from the feature interval [F]. min F max Randomly select a set of feature values ​​F t And generate a set of undetermined intervals [F t-X1 F t+X1 ], where X1 is the preset feature tolerance value, and t represents different feature values, and F t ∈[F min+X1 F max-X1 Based on the undetermined interval [F] t-X1 F t+X1 Identify the relevant line segments within the undetermined interval from the relevant characteristic change curve, and record the line length Ct of the corresponding relevant line segments. Select the maximum value Ct from several line lengths Ct. max , will Ct max The corresponding undetermined interval [F] t-X1 F t+X1 [This is] calibrated as a characteristic stable interval;

[0017] Feature matching segmentation: Based on the determined feature stability interval, the relevant feature values ​​within the interval are used as segmentation matching parameters to perform feature matching segmentation on the pre-segmented myocardial infarction image. When the gray mean of the segmented region reaches G2, feature matching segmentation is paused and the real-time feature adjustment step is entered.

[0018] Real-time feature adjustment: The image feature parameters during the feature matching and segmentation stage are monitored in real time, and the monitored feature parameters are transmitted to the real-time adjustment module. The real-time adjustment module adjusts the segmentation and matching parameters in real time based on the changes in the image feature parameters.

[0019] Preferably, the specific method for real-time feature adjustment is as follows: the image feature parameters of the previous pixel are calibrated as D1, the image feature parameters of the current pixel are calibrated as D2, the feature difference CZ between adjacent pixels is determined, and CZ = (D2 - D1). Based on this feature difference CZ, the segmentation matching parameter is adjusted, and the adjusted parameter value is (CZ + Y1), where Y1 is a preset value for parameter adjustment. It is also determined whether the segmentation matching degree of the corresponding pixel belongs to the preset matching degree range. If it does, when the image feature parameter changes by a certain value R, the segmentation matching parameter is adjusted by the value of (R + Y1). If it does not belong, when the image feature parameter changes by a certain value R, the segmentation matching parameter is also adjusted by the value of R accordingly.

[0020] Preferably, the specific process of constant threshold fine segmentation is as follows: when the gray-scale mean of the segmented region reaches G2, the segmentation threshold corresponding to the feature stable interval remains unchanged, and the image segmented region is finely processed using a constant segmentation threshold. The matching degree monitoring unit monitors the matching degree between the segmented region and the labeled lesion region until the matching degree is higher than the set value Y2, at which point the fine segmentation stops and the entire segmentation process of the myocardial infarction image is completed, where Y2 is the preset matching degree value.

[0021] Preferably, the constant threshold fine segmentation adopts an intermittent fine segmentation method, specifically: when the gray-scale mean of the segmented region reaches G2, the segmentation threshold corresponding to the stable feature interval remains unchanged, and a constant segmentation threshold is used to fine-tune the segmented region of the image. The matching degree monitoring unit monitors the matching degree between the segmented region and the labeled lesion region until the matching degree is higher than the set value Y2, and then the subsequent steps are executed; every preset number of pixels M, the edge of the segmented region is fine-tuned by N pixels with a constant segmentation threshold, and after the fine segmentation is completed, the initial matching degree I1 and the final matching degree I2 of the fine segmented region are confirmed, and the matching degree interval [I1, I2] of the corresponding fine segmentation stage is determined; when I1-Y2 < Y2-I2, the fine segmentation process is stopped, and the segmentation process of the corresponding myocardial infarction image is completed.

[0022] This invention provides a method for image segmentation of myocardial infarction based on data processing. Compared with existing technologies, it has the following advantages:

[0023] This invention performs differentiated pre-segmentation processing based on the initial feature parameters of myocardial infarction images, and adapts and adjusts the edge detection gradient by combining the feature trend interval of historical segmentation data to restore the true feature distribution of the image. This avoids the problems of missed detection and over-segmentation caused by the mismatch of initial segmentation parameters, ensuring the pre-segmentation effect while laying the foundation for subsequent accurate segmentation.

[0024] Based on historical segmentation data and similar labeled data, this invention determines the stable feature range of the image. This range corresponds to the pixel range where the lesion region features in the image are most stable. Based on the feature values ​​of this range, segmentation matching can accurately locate the core region of myocardial infarction lesions, improving the targeting and accuracy of segmentation.

[0025] This invention monitors image feature parameters in real time and dynamically adjusts segmentation matching parameters during the core segmentation stage, ensuring that the feature matching degree is always in the optimal range during the segmentation process. This not only improves segmentation efficiency but also avoids segmentation deviations caused by sudden changes in local image features.

[0026] In the final segmentation stage, this invention employs a constant threshold refinement or intermittent refinement method to finely process the edges of the segmented region, further improving the matching degree between the segmented region and the actual lesion region, while avoiding edge distortion caused by excessive refinement, thus providing high-precision myocardial infarction image segmentation data for clinical use. Attached Figure Description

[0027] Figure 1 This is a flowchart of the present invention; Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Example

[0030] Please see Figure 1 This application provides a method for segmenting myocardial infarction images based on data processing, including the following steps:

[0031] Initial feature monitoring of images: Initial feature monitoring is performed on the myocardial infarction images to be segmented (mainly cardiac MRI and CT images) to obtain initial feature parameters such as the initial gray-scale mean, edge gradient of suspected lesion areas, and image noise value.

[0032] Image pre-segmentation: Based on the monitored initial feature parameters, the myocardial infarction image is pre-segmented, and different pre-segmentation methods are executed based on the specific values ​​of the initial feature parameters. Specifically, the different pre-segmentation methods based on the specific values ​​of the initial feature parameters include: labeling the initial grayscale mean as G, the image noise value as N, setting a preset grayscale threshold as G0, a noise threshold as N0, ΔG as a preset grayscale tolerance value, and K as the standard edge detection gradient.

[0033] If G∈[G0-ΔG,G0+ΔG] and N≤N0, then a constant edge detection gradient of 0.1K is used to pre-segment the myocardial infarction image. The pixel traversal range of the pre-segmentation is S, and the pre-segmentation ends, where S is a preset value.

[0034] If G∉[G0-ΔG,G0+ΔG] or N>N0, obtain relevant feature data of similar myocardial infarction images in the corresponding gray / noise interval from historical segmentation data. Based on the image pixel coordinates, confirm the correlation feature change curve of its relevant feature data. The horizontal axis of the curve is the pixel horizontal coordinate, and the vertical axis is the feature parameter. Sequentially confirm the feature trend value T of adjacent pixels within the correlation feature change curve. i Its trend value T i = (Vertical coordinate difference between adjacent pixels) ÷ Horizontal coordinate difference between adjacent pixels, where the coordinate difference is the coordinate parameter of the next pixel minus the coordinate parameter of the previous pixel, and the identified several feature trend values ​​T are then used. i After performing absolute value processing, the trend interval [T] is determined. min T max ], and T is the trend value after absolute value processing, where i represents the line segment between different adjacent pixels;

[0035] A constant edge detection gradient of 0.06K is preferentially used for pre-segmentation of myocardial infarction images, and the feature change trend F of the images is monitored in real time during the segmentation process. If F∈[T] min T max If F < T, then maintain a constant edge detection gradient for pre-segmentation; min Then, the edge detection gradient is gradually increased up to a maximum of 0.1K, so that F∈[T] min T max [Time ends; if F > T] max Then, gradually reduce the edge detection gradient to 0.02K, so that F∈[T] min T max [Time limit]; When F cannot be adjusted to fall within this trend range, an abnormal image signal is directly generated for display;

[0036] When the clarity of the outline of the suspected lesion area in the pre-segmented image is greater than or equal to the preset clarity threshold, the pre-segmentation ends and constant feature segmentation control is executed.

[0037] Constant feature segmentation control: Based on historical segmentation data of myocardial infarction images and labeled data of similar lesion images, the image feature change curve is confirmed, and the segmentation process in this stage is controlled based on the feature change curve to achieve accurate segmentation of the lesion region; the constant feature segmentation control includes three sub-steps: feature stability interval confirmation, feature matching segmentation, and real-time feature adjustment, specifically as follows:

[0038] Feature stability interval confirmation: From historical segmentation data of myocardial infarction images and labeled data of similar lesion images, relevant feature data with image grayscale values ​​between G1 and G2 were confirmed. G1 is the lower limit of grayscale value of normal myocardial tissue, and G2 is the upper limit of grayscale value of infarct lesions. The minimum value F of the feature parameter was determined from this relevant feature data. min and the maximum value of the characteristic parameter F max And determine the characteristic interval [F] min F max Based on the relevant feature data corresponding to different pixels, and according to the spatial relationship of the pixels, a curve showing the change of their relevant features is generated from the feature interval [F]. min F max Randomly select a set of feature values ​​F t And generate a set of undetermined intervals [F t-X1 F t+X1 ], where X1 is the preset feature tolerance value, and t represents different feature values, and F t ∈[F min+X1 F max-X1 Based on the undetermined interval [F] t-X1 F t+X1 Identify the relevant line segments within the undetermined interval from the relevant characteristic change curve, and record the line length Ct of the corresponding relevant line segments. Select the maximum value Ct from several line lengths Ct. max , will Ct max The corresponding undetermined interval [F] t-X1 F t+X1 [This is] calibrated as a characteristic stable interval;

[0039] Feature matching segmentation: Based on the determined feature stability interval, the relevant feature values ​​within the interval are used as segmentation matching parameters to perform feature matching segmentation on the pre-segmented myocardial infarction image. When the gray mean of the segmented region reaches G2, feature matching segmentation is paused and the real-time feature adjustment step is entered.

[0040] Real-time feature adjustment: The image feature parameters during the feature matching and segmentation stage are monitored in real time, and the monitored feature parameters are transmitted to the real-time adjustment module. The real-time adjustment module adjusts the segmentation matching parameters in real time based on the changes in the image feature parameters. The specific method of real-time feature adjustment is as follows: The image feature parameters of the previous pixel are labeled as D1, and the image feature parameters of the current pixel are labeled as D2. The feature difference CZ between adjacent pixels is determined, and CZ = (D2 - D1). Based on this feature difference CZ, the segmentation matching parameters are adjusted. The adjusted parameter value is (CZ + Y1), where Y1 is a preset value for parameter adjustment. It is determined whether the segmentation matching degree of the corresponding pixel belongs to the preset matching degree range. If it does, when the image feature parameters change by a certain value R, the segmentation matching parameters are adjusted by the value of (R + Y1). If it does not belong, when the image feature parameters change by a certain value R, the segmentation matching parameters are also adjusted by the value of R accordingly.

[0041] Constant Threshold Refined Segmentation: A constant segmentation threshold is used to refine the image regions after pre-segmentation and constant feature segmentation. When the matching degree of the segmented regions reaches the target, the entire segmentation process of the myocardial infarction image is completed. The specific process of constant threshold refined segmentation is as follows: when the gray-scale mean of the segmented region reaches G2, the segmentation threshold corresponding to the stable feature interval remains unchanged. A constant segmentation threshold is used to refine the segmented image region, and the matching degree monitoring unit monitors the matching degree between the segmented region and the labeled lesion region. When the matching degree is higher than the set value Y2, the refined segmentation stops, and the entire segmentation process of the myocardial infarction image is completed, where Y2 is the preset matching degree value. The constant threshold fine segmentation adopts an intermittent fine segmentation method, specifically: when the gray-scale mean of the segmented region reaches G2, the segmentation threshold corresponding to the stable feature interval remains unchanged, and a constant segmentation threshold is used to refine the segmented region of the image. The matching degree monitoring unit monitors the matching degree between the segmented region and the labeled lesion region until the matching degree is higher than the set value Y2, and then the subsequent steps are executed; every preset number of pixels M, the edge of the segmented region is refined by N pixels with a constant segmentation threshold, and after the fine segmentation is completed, the initial matching degree I1 and the final matching degree I2 of the refined region are confirmed, and the matching degree interval [I1, I2] of the corresponding fine segmentation stage is determined; when I1-Y2 < Y2-I2, the fine segmentation process is stopped, and the segmentation process of the corresponding myocardial infarction image is completed.

[0042] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0043] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for segmenting myocardial infarction images based on data processing, characterized in that, Includes the following steps: Initial feature monitoring of images: Initial feature monitoring is performed on the myocardial infarction images to be segmented to obtain initial feature parameters such as the initial gray-scale mean, the edge gradient of the suspected lesion area, and the image noise value. Image pre-segmentation: Based on the monitored initial feature parameters, the myocardial infarction images are pre-segmented, and different pre-segmentation methods are executed based on the specific values ​​of the initial feature parameters. Constant feature segmentation control: Based on historical segmentation data of myocardial infarction images and labeled data of similar lesion images, the image feature change curve is identified, and the segmentation process in this stage is controlled based on the feature change curve to achieve accurate segmentation of the lesion area; Constant threshold fine segmentation: A constant segmentation threshold is used to refine the image regions after pre-segmentation and constant feature segmentation. When the matching degree of the segmented regions meets the standard, the entire segmentation process of the myocardial infarction image is completed.

2. The image segmentation method for myocardial infarction based on data processing according to claim 1, characterized in that, The specific pre-segmentation method based on the initial feature parameters includes: labeling the initial grayscale mean as G, the image noise value as N, the preset grayscale threshold as G0, the noise threshold as N0, ΔG as the preset grayscale tolerance value, and K as the standard edge detection gradient. If G∈[G0-ΔG,G0+ΔG] and N≤N0, then a constant edge detection gradient of 0.1K is used to pre-segment the myocardial infarction image. The pixel traversal range of the pre-segmentation is S, and the pre-segmentation ends, where S is a preset value. If G∉[G0-ΔG,G0+ΔG] or N>N0, obtain relevant feature data of similar myocardial infarction images in the corresponding gray / noise interval from historical segmentation data. Based on the image pixel coordinates, confirm the correlation feature change curve of its relevant feature data. The horizontal axis of the curve is the pixel horizontal coordinate, and the vertical axis is the feature parameter. Sequentially confirm the feature trend value T of adjacent pixels within the correlation feature change curve. i Its trend value T i = (Vertical coordinate difference between adjacent pixels) ÷ Horizontal coordinate difference between adjacent pixels, where the coordinate difference is the coordinate parameter of the next pixel minus the coordinate parameter of the previous pixel, and the identified several feature trend values ​​T are then used. i After performing absolute value processing, the trend interval [T] is determined. min T max ], and T is the trend value after absolute value processing, where i represents the line segment between different adjacent pixels; A constant edge detection gradient of 0.06K is preferentially used for pre-segmentation of myocardial infarction images, and the feature change trend F of the images is monitored in real time during the segmentation process. If F∈[T] min T max If F < T, then maintain a constant edge detection gradient for pre-segmentation; min Then, the edge detection gradient is gradually increased up to a maximum of 0.1K, so that F∈[T] min T max [Time ends; if F > T] max Then, gradually reduce the edge detection gradient to 0.02K, so that F∈[T] min T max [Time limit]; When F cannot be adjusted to fall within this trend range, an abnormal image signal is directly generated for display; When the clarity of the outline of the suspected lesion area in the pre-segmented image is greater than or equal to the preset clarity threshold, the pre-segmentation ends and constant feature segmentation control is executed.

3. The image segmentation method for myocardial infarction based on data processing according to claim 1, characterized in that, The constant feature segmentation control includes three sub-steps: feature stability interval confirmation, feature matching and segmentation, and real-time feature adjustment. Specifically: Feature stability interval confirmation: From historical segmentation data of myocardial infarction images and labeled data of similar lesion images, relevant feature data with image grayscale values ​​between G1 and G2 were confirmed. G1 is the lower limit of grayscale value of normal myocardial tissue, and G2 is the upper limit of grayscale value of infarct lesions. The minimum value F of the feature parameter was determined from this relevant feature data. min and the maximum value of the characteristic parameter F max And determine the characteristic interval [F] min F max Based on the relevant feature data corresponding to different pixels, and according to the spatial relationship of the pixels, a curve showing the change of their relevant features is generated from the feature interval [F]. min F max Randomly select a set of feature values ​​F t And generate a set of undetermined intervals [F t-X1 F t+X1 ], where X1 is the preset feature tolerance value, and t represents different feature values, and F t ∈[F min+X1 F max-X1 Based on the undetermined interval [F] t-X1 F t+X1 Identify the relevant line segments within the undetermined interval from the relevant characteristic change curve, and record the line length Ct of the corresponding relevant line segments. Select the maximum value Ct from several line lengths Ct. max , will Ct max The corresponding undetermined interval [F] t-X1 F t+X1 [This is] calibrated as a characteristic stable interval; Feature matching segmentation: Based on the determined feature stability interval, the relevant feature values ​​within the interval are used as segmentation matching parameters to perform feature matching segmentation on the pre-segmented myocardial infarction image. When the gray mean of the segmented region reaches G2, feature matching segmentation is paused and the real-time feature adjustment step is entered. Real-time feature adjustment: The image feature parameters during the feature matching and segmentation stage are monitored in real time, and the monitored feature parameters are transmitted to the real-time adjustment module. The real-time adjustment module adjusts the segmentation and matching parameters in real time based on the changes in the image feature parameters.

4. The image segmentation method for myocardial infarction based on data processing according to claim 3, characterized in that, The specific method for real-time feature adjustment is as follows: the image feature parameters of the previous pixel are calibrated as D1, and the image feature parameters of the current pixel are calibrated as D2. The feature difference CZ between adjacent pixels is determined, and CZ = (D2 - D1). Based on this feature difference CZ, the segmentation matching parameters are adjusted, and the adjusted parameter value is (CZ + Y1), where Y1 is a preset value for parameter adjustment. It is also determined whether the segmentation matching degree of the corresponding pixel belongs to the preset matching degree range. If it does, when the image feature parameters change by a certain value R, the segmentation matching parameters are adjusted by the value of (R + Y1). If it does not belong, when the image feature parameters change by a certain value R, the segmentation matching parameters are also adjusted by the value of R accordingly.

5. The image segmentation method for myocardial infarction based on data processing according to claim 1, characterized in that, The specific process of constant threshold fine segmentation is as follows: when the gray average value of the segmented region reaches G2, the segmentation threshold corresponding to the stable feature interval remains unchanged, and the segmented region of the image is finely processed using a constant segmentation threshold. The matching degree monitoring unit monitors the matching degree between the segmented region and the labeled lesion region until the matching degree is higher than the set value Y2, at which point the fine segmentation stops and the entire segmentation process of the myocardial infarction image is completed, where Y2 is the preset matching degree value.

6. The image segmentation method for myocardial infarction based on data processing according to claim 5, characterized in that, The constant threshold fine segmentation adopts an intermittent fine segmentation method, specifically: when the gray-scale mean of the segmented region reaches G2, the segmentation threshold corresponding to the stable feature interval remains unchanged, and a constant segmentation threshold is used to refine the segmented region of the image. The matching degree monitoring unit monitors the matching degree between the segmented region and the labeled lesion region until the matching degree is higher than the set value Y2, and then the subsequent steps are executed; every preset number of pixels M, the edge of the segmented region is refined by N pixels with a constant segmentation threshold, and after the fine segmentation is completed, the initial matching degree I1 and the final matching degree I2 of the refined region are confirmed, and the matching degree interval [I1, I2] of the corresponding fine segmentation stage is determined; when I1-Y2 < Y2-I2, the fine segmentation process is stopped, and the segmentation process of the corresponding myocardial infarction image is completed.

Images