A cardiovascular medicine OCT image filtering enhancement method

Abstract

The present application relates to the technical field of image processing, in particular to a kind of cardiovascular medicine OCT image filtering enhancement method, the method comprises: by obtaining optical coherence tomography OCT image, and the classification distribution parameter and tissue distribution parameter of image, wherein classification distribution parameter is based on the light and dark change of original image and is obtained by classifying original image, and tissue distribution parameter is based on the tissue distribution of original image and is obtained by segmenting original image.Then based on the difference of classification distribution parameter and tissue distribution parameter, determine the noise influence factor of original image, and according to noise influence factor, original image is carried out non-local mean filtering, clear target image that inhibits speckle noise is obtained, finally, target image is enhanced, it is ensured that the OCT image after enhancement retains important structure details, and unnecessary noise interference is reduced.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and more specifically to a method for filtering and enhancing cardiovascular OCT images. Background Technology

[0002] Optical coherence tomography (OCT) is a high-resolution imaging technique widely used in cardiovascular examinations, such as assessing the microstructure of blood vessels, particularly in evaluating atherosclerotic plaques and other cardiovascular diseases.

[0003] Since some minor or early lesions may not be easily detected in the original images, special enhancement processing is needed on cardiovascular OCT images to improve the doctor's ability to examine patients' cardiovascular OCT images, thereby helping to detect these subtle abnormalities.

[0004] However, due to the presence of speckle noise in OCT images, the speckle noise is also amplified during the enhancement process, which in turn affects the image clarity and enhancement effect. Summary of the Invention

[0005] To address the technical problem of unclear OCT images caused by speckle noise being amplified during enhancement, this invention aims to provide a filtering and enhancement method for cardiovascular OCT images. The specific technical solution adopted is as follows:

[0006] A method for filtering and enhancing OCT images in cardiovascular medicine, the method comprising:

[0007] Acquire the original image, which is an optical coherence tomography (OCT) image;

[0008] Obtain the classification distribution parameters of the original image, which are obtained by classifying the original image based on the brightness and darkness variations of the original image;

[0009] Obtain the tissue distribution parameters of the original image, which are obtained by segmenting the original image based on its tissue distribution.

[0010] The noise impact factor of the original image is determined based on the classification distribution parameter and the tissue distribution parameter. The noise impact factor is used to represent the difference between the classification distribution parameter and the tissue distribution parameter.

[0011] The target image is obtained by performing nonlocal mean filtering on the original image based on the noise impact factor.

[0012] Enhance the target image.

[0013] Optionally, the above steps include: obtaining the classification distribution parameters of the original image, including:

[0014] Obtain the brightness fluctuation parameters of each pixel in the original image. The brightness fluctuation parameters are determined based on the rate of change of the gray value of the pixel.

[0015] Based on the brightness fluctuation parameters of each pixel, each pixel is clustered to obtain multiple first category clusters and first pixel clusters contained in multiple first category clusters;

[0016] The image distribution factor of each first classification cluster is determined based on the relationship between the first pixel cluster in each first classification cluster and the center point of the first classification cluster.

[0017] The classification distribution parameters of the original image are determined based on the image distribution factor of each first classification cluster.

[0018] Optionally, the brightness fluctuation parameter is determined based on the distance between each maximum point in the original image, the gray value of the maximum point, the average gray value of the pixels between the maximum points, and the number of pixels between the maximum points. The maximum point is the pixel with the largest gray value in the neighborhood of each pixel.

[0019] Optionally, the image distribution factor is determined based on the information entropy of the angle between the first pixel cluster in each first classification cluster and the center point, and the variance of the distance from the center point.

[0020] Optionally, the classification distribution parameters are determined based on the average value of the image distribution factor for each first classification cluster and the information entropy of the average center distance of the first pixel cluster in each first classification cluster.

[0021] Optionally, the above steps include: obtaining the tissue distribution parameters of the original image, including:

[0022] The original image after the closing operation is segmented to obtain multiple clusters of second pixel points;

[0023] The trapping parameters of each second pixel cluster are determined based on multiple second pixel clusters;

[0024] The tissue distribution parameters of the original image are determined based on the loop parameters of each second pixel cluster.

[0025] Optionally, the trap parameters are determined based on the area of ​​the minimum circumcircle of the second pixel cluster and the number of pixels belonging to the second pixel cluster within the minimum circumcircle.

[0026] Optionally, the organization distribution parameters are determined based on the information entropy of the product of the minimum circumcircle diameter of the second pixel cluster and the loop parameter, as well as the variance of the center of the minimum circumcircle of the second pixel cluster.

[0027] Optionally, the above steps, including enhancing the target image, include:

[0028] Enhancement of target images based on histogram equalization.

[0029] Optionally, the original image is a cardiovascular image.

[0030] The present invention has the following beneficial effects:

[0031] Optical coherence tomography (OCT) images were acquired, along with their classification and tissue distribution parameters. The classification parameters were obtained by classifying the original image based on its brightness variations, while the tissue distribution parameters were obtained by segmenting the original image based on its tissue distribution. Then, the noise impact factor of the original image was determined based on the difference between the classification and tissue distribution parameters. Non-local mean filtering was then applied to the original image based on this noise impact factor to obtain a clear target image with suppressed speckle noise. Finally, the target image was enhanced to ensure that the enhanced OCT image retained important structural details while reducing unnecessary noise interference. Attached Figure Description

[0032] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a schematic diagram of a cardiovascular OCT image filtering and enhancement method provided in one embodiment of the present invention;

[0034] Figure 2 A schematic diagram of the original image provided for one embodiment of the present invention;

[0035] Figure 3 This is a schematic diagram of the maximum value points of pixels in the original image provided in an embodiment of the present invention;

[0036] Figure 4 This is a schematic diagram of the result obtained by performing k-means segmentation on the original image after performing overclosing operation, according to an embodiment of the present invention.

[0037] Figure 5 This is a schematic diagram of the minimum circumcircle provided in one embodiment of the present invention. Detailed Implementation

[0038] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a cardiovascular OCT image filtering enhancement method proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0039] 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 this invention pertains.

[0040] The specific scheme of the cardiovascular OCT image filtering and enhancement method provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0041] Please see Figure 1 The diagram illustrates a flowchart of a cardiovascular OCT image filtering and enhancement method according to an embodiment of the present invention, the method comprising:

[0042] 101. Obtain the original image.

[0043] It should be understood that the cardiovascular OCT image filtering enhancement method provided by the present invention can be applied to a variety of devices or instruments, such as computing devices or other units or modules with computing functions.

[0044] In this application, the original image is an OCT image of the cardiovascular system. The specific acquisition method and scenario are described below with examples. It is not excluded that the acquisition method and application scenario of the original image provided by this invention may be other possibilities.

[0045] like Figure 2 As shown, for patients requiring cardiovascular examinations, doctors insert a guide into the patient's vascular system via puncture or intravascular cannulation to determine its location and orientation in OCT imaging. Once the guide is in the target position, the doctor introduces an optical probe through it. This probe contains a light source and optical elements that transmit a light beam to the area of ​​interest in the blood vessel and receive the reflected light signals. The optical probe scans the vessel wall very quickly, recording the reflected light signals at each location. These signals are then converted into digital signals and transmitted to a computer system for processing. The computer system processes the received light signals and uses OCT algorithms to reconstruct images of the internal structures of the blood vessel. These images reveal detailed information about the internal structures such as the vessel wall, plaque, and thrombi—these are the original images.

[0046] Optionally, for the acquired cardiovascular OCT images, since the OCT images need to be enhanced, in order to reduce the impact of noise after enhancement, it is necessary to distinguish between noise and normal tissue and to distinguish the texture in the image. Therefore, an edge detection algorithm (Canny) can be applied to the OCT image first to obtain the edges in the image. The edges contain vascular tissue texture and noise texture, and the edges in the image are used as the original image for subsequent processing.

[0047] For the edges in the cardiovascular OCT images obtained by the above operations, some of the edges are generated by noise. Since the original image needs to be enhanced, it is necessary to first obtain the degree of noise interference in the original image due to the noise interference in the enhancement process.

[0048] Because OCT technology utilizes the coherence of laser light for imaging, when laser light shines on different tissues, the reflected light waves will interfere due to different reflection paths and scattering, resulting in random variations in light intensity. This manifests as speckle noise in the original image. Speckle noise affects the visualization of plaques within the coronary arteries, including calcified plaques, soft plaques, and fibrous plaques, thus impacting patient diagnosis.

[0049] The following section provides a detailed explanation of how to eliminate the effects of speckle noise.

[0050] 102. Obtain the classification distribution parameters of the original image.

[0051] In this embodiment, the classification distribution parameters are obtained by classifying the original image based on the brightness and darkness variations of the original image. Since speckle noise in OCT images is formed by the superposition and interference of reflected light from various tissues, it is known from the cause of speckle noise that it will cause some areas in the original image to appear as alternating bright and dark areas. Therefore, the brightness and darkness fluctuation parameters of each pixel can be obtained based on the gray value fluctuations of the pixels in the original image. That is, the brightness and darkness fluctuation parameters of each pixel in the original image are first obtained, and the brightness and darkness fluctuation parameters are determined based on the rate of change of the gray value of the pixel.

[0052] For example, the brightness fluctuation parameter is determined based on the distance between each maximum point in the original image, the gray value of the maximum point, the average gray value of the pixels between the maximum points, and the number of pixels between the maximum points. The maximum point is the pixel with the largest gray value in the neighborhood of each pixel.

[0053] Specifically, such as Figure 3As shown, for each pixel, the pixel with the largest gray value among its eight neighbors and itself is designated as the maximum. A 5×5 neighborhood is constructed around each pixel. Since each pixel in the 5×5 neighborhood has a maximum in its corresponding eight neighbors, and a pixel's 5×5 neighborhood can completely cover 9 pixels and their eight neighbors, there are several maximums in a pixel's 5×5 neighborhood. Similarly, for each pixel, the pixel with the smallest gray value among its eight neighbors and itself is designated as the minimum. The brightness fluctuation parameter of the pixel is quantified based on the distance between the maximums and the difference in gray values ​​between pixels with maximums.

[0054]

[0055] In the formula, Indicates the first The brightness fluctuation parameters of each pixel It is a pixel in any OCT image. It is any local maximum point in the 5×5 neighborhood of any pixel. Is with the first The nearest maximum point to a given maximum point It is the first In the neighborhood of the nth pixel The distance between a local maximum and its nearest local maximum It is the first In the neighborhood of the nth pixel The gray value of the maximum point It is the first In the neighborhood of the nth pixel The gray value of the nearest maximum point of a given maximum point. It is the first In the neighborhood of the nth pixel The average gray value of pixels between a given maximum point and its nearest maximum point. It is the first In the neighborhood of the nth pixel The number of minimum points between a maximum point and its nearest maximum point. It is a hyperbolic function.

[0056] For each pixel in the original image, the distance between any local maxima and its nearest local maxima is... The smaller the value, the faster the grayscale values ​​change in the original image. Furthermore, the ratio of the average grayscale value of the local maximum point and its nearest local maximum to the average grayscale value of the pixels in the neighborhood of a pixel is... The larger the value, the smaller the grayscale value of the pixel between the maxima, indicating a faster change in grayscale value. Furthermore, it indicates the number of minima between any maxima and its nearest neighbor within the pixel's neighborhood. The larger the value, the more minima there are between the maxima, which indicates that the grayscale value changes faster, and therefore the larger the brightness fluctuation parameter.

[0057] It should be understood that the above steps for obtaining brightness fluctuation parameters are only an example, and brightness fluctuation parameters can also be obtained through other methods or means. This application embodiment does not limit this.

[0058] Furthermore, since different tissues in cardiovascular OCT images have different scattering coefficients and refractive indices, the speckle noise in different tissue parts will vary. Therefore, the brightness coefficients in the original image will change with tissue variations, resulting in some regularity in the brightness coefficients. However, due to the influence of lesions inside the cardiovascular system, the regularity of speckle noise is affected. In order to obtain the true speckle noise characteristics, after first obtaining the brightness fluctuation parameters of the pixels in the image, it is also necessary to classify the original image.

[0059] First, clustering is performed on the brightness fluctuation parameters of each pixel in the original image to obtain... Pixel clustering involves grouping pixels based on their brightness fluctuation parameters to obtain multiple first-class clusters. Then, based on the distribution of pixels, neighboring pixels within each first-class cluster are grouped as first-pixel clusters. For example, clustering can be based on the distance distribution between pixels to obtain multiple first-pixel clusters, meaning each first-class cluster contains multiple first-pixel clusters. For instance, k-means clustering (a clustering method) can be used based on brightness fluctuation parameters. Next, the image distribution factors of each first-class cluster are obtained, and finally, the classification distribution parameters of the original image are obtained. This is achieved by determining the image distribution factor of each first-class cluster based on the relationship between the first-pixel clusters within each first-class cluster and the cluster center point, and then using the image distribution factors of each first-class cluster to determine the classification distribution parameters of the original image.

[0060] Specifically, since each first classification cluster contains several first pixel clusters, in order to obtain the image distribution factor of the first classification cluster, for each first pixel cluster in each first classification cluster, the centroid of the first pixel cluster is determined, and the distance between the centroid and the center point of the original image is obtained, denoted as the center distance. Then, the image distribution factor is obtained according to the relationship between each first pixel cluster in the first classification cluster and the center point. That is, the image distribution factor is determined based on the information entropy of the angle between the first pixel cluster in each first classification cluster and the center point and the variance of the distance from the center point.

[0061]

[0062] In the formula, It is the first Image distribution factor of the first classification cluster, It is any first category cluster, It is the first The information entropy of the angle between the first pixel cluster and the center point in the first category cluster. It is the first The variance of the distance between the first pixel cluster and the center in each first category cluster. It is an exponential function with the natural constant as its base.

[0063] The variance of the distance between each first pixel cluster and the center point in each first category cluster The larger the value, the more irregular the distances between the first pixel clusters, and the worse the image distribution factor. The information entropy of the angle between each first pixel cluster and the center point... The larger the value, the more likely that each first category cluster exists in the original image from all angles, and therefore the better its image distribution factor.

[0064] Then, the classification distribution parameters of the original image are obtained based on the image distribution factors of each first classification cluster in the original image. That is, the classification distribution parameters are determined based on the average value of the image distribution factors of each first classification cluster and the information entropy of the average center distance of the first pixel cluster in each first classification cluster.

[0065]

[0066] In the formula, These are the classification distribution parameters of the original image. It is the number of classification clusters in the image. It is the information entropy of the average center distance of the first pixel clusters in each first category of the original image. It is a hyperbolic function.

[0067] The average value of the image distribution factor for each first category in the original image. The larger the value, the better the overall classification distribution of the original image, and the more information entropy it represents for the average center distance of each first pixel cluster in each first classification cluster of the original image. The larger the value, the wider the center distance distribution of each first category cluster, indicating that the distribution of each first category cluster is more uniform, and therefore the better the classification distribution parameters of the original image.

[0068] It should be understood that the above steps for obtaining classification distribution parameters are only an example, and classification distribution parameters can also be obtained through other methods or means. This application embodiment does not limit this.

[0069] 103. Obtain the tissue distribution parameters of the original image.

[0070] Regarding the image distribution factors of each first category cluster in the original image obtained through the above operations, since the image distribution factors obtained based on the brightness fluctuation parameters within different tissues are similar, the distribution of the first category clusters obtained by each brightness parameter should be relatively clustered, because tissues are clustered. However, due to factors such as lesions in the tissue, the distribution of the brightness fluctuation parameters is affected, resulting in poor clustering of the first distribution clusters obtained by each brightness fluctuation parameter. Therefore, it is necessary to obtain the degree of influence of the brightness fluctuation parameters to obtain the degree of noise influence of the original image itself. That is, since the brightness distribution of the tissue itself is relatively concentrated and uniform, it is necessary to judge the distribution of the original image itself and obtain the tissue distribution parameters of the original image, where the tissue distribution parameters are obtained by segmenting the original image based on the tissue distribution of the original image.

[0071] In this embodiment, in order to obtain the tissue distribution parameters of the original image, the original image after the closing operation is first segmented to obtain multiple second classification clusters. Each second classification cluster contains multiple second pixel clusters. Then, the looping parameter of each second pixel cluster is determined according to the looping parameter of each second pixel cluster. Finally, the tissue distribution parameters of the original image can be determined according to the looping parameter of each second pixel cluster.

[0072] Specifically, firstly, a closing operation is performed on the original image to eliminate speckle noise. Then, k-means segmentation is performed on the original image after the closing operation, dividing it into several clusters of second-pixel points. For example... Figure 4 As shown, in OCT technology, the grayscale values ​​of the same tissue differ due to varying reflections of laser light. Therefore, k-means can be used to segment pixels with similar grayscale values ​​into clusters, thus segmenting the same tissue into a second pixel cluster (a second pixel cluster can be understood as corresponding to one tissue). Then, the tissue distribution factor in the original image is obtained based on the segmented second pixel clusters. It should be understood that the tissues in the original image are generally circular blood vessels; therefore, different tissues should be distributed in concentric circles. To obtain the influence of noise on the distribution of different tissues, it is necessary to first determine... The nested parameters of each class in the class (obtained by k-means clustering after the closing operation). The trap parameter is determined based on the area of ​​the minimum circumcircle of the second pixel cluster and the number of pixels belonging to the second pixel cluster in the minimum circumcircle.

[0073] Furthermore, for each cluster of second-pixel points, it is necessary to determine its minimum circumcircle, such as... Figure 5 As shown, the minimum bounding circle of each second pixel cluster is obtained, and the center of the minimum bounding circle is... The diameter of the smallest circumcircle is ,have:

[0074]

[0075] In the formula, It is any cluster of second pixel points. It is the first The minimum bounding circle area of ​​the second pixel cluster. It is the first The number of pixels belonging to the second pixel cluster within the smallest outer circle of the second pixel cluster.

[0076] Then, based on the diameter distribution and center of different tissues in the original image, the tissue distribution parameters are obtained. That is, the tissue distribution parameters are determined based on the information entropy of the product of the minimum circumcircle diameter of the second pixel cluster and the loop parameter, and the variance of the center of the minimum circumcircle of the second pixel cluster.

[0077]

[0078] In the formula, ZF is the tissue distribution parameter. It is based on the minimum circumcircle diameter of each second pixel cluster in the original image and the loop parameter. The information entropy obtained by calculating the product is... It is the variance of the center of the smallest circumcircle of each second pixel cluster in the original image.

[0079] For each tissue in the original image, since blood vessels are circular and different tissues, such as the blood vessel wall, are arranged in a concentric circle, the ratio of the number of pixels inside the smallest circumcircle of each tissue distribution in the original image to the area of ​​the smallest circumcircle is used. To obtain its trap parameters The larger the value, the more internal structure there is within the smallest circumcircle, and therefore its encirclement parameter... The higher. And based on the radius and loop parameters of different tissues (different second pixel clusters) in the original image. Information entropy of the product A higher information entropy value indicates a more uniform distribution of the circumradii of the various tissue parts, thus suggesting that the tissues in the original image are distributed in a circumradius pattern. Furthermore, based on the variance of the center of the minimum circumcircle of each tissue in the original image... The smaller the value, the better the tissue distribution parameter, indicating that the tissues are distributed in concentric circles.

[0080] It should be noted that the method for calculating information entropy is a well-known technique. In this embodiment, the entropy is calculated using the nesting parameters of each second pixel cluster. and minimum circumcircle diameter The product result is used as a random variable to calculate information entropy. The larger the information entropy value, the more uniform the probability distribution of the product result (that is, the more uniform the probability distribution of each second pixel cluster). The values ​​are close, with no extreme outliers. The trapping characteristics and sizes of all tissues in the original image conform to normal physiological patterns, and the structures are relatively similar. The smaller the information entropy value, the more uneven the probability distribution of the product calculation result (i.e., the probability of a certain second pixel cluster...). The proportion is extremely high, such as abnormally dispersed clusters caused by noise, or abnormally compact clusters caused by lesions. The corresponding tissue structure is disrupted from normal distribution, resulting in extreme abnormalities. This may be due to pathological or noise-induced disorder.

[0081] It should be understood that the pixel coordinates of the center of the minimum circumcircle of each second pixel cluster in the original image are obtained, the variance of the x-coordinate and the variance of the y-coordinate of the corresponding center of all second pixel clusters are calculated respectively, and the average value between the variance of the x-coordinate and the variance of the y-coordinate is calculated to obtain the variance of the center of the minimum circumcircle of all second pixel clusters in the original image.

[0082] It should be understood that the above steps for obtaining tissue distribution parameters are only an example, and tissue distribution parameters can also be obtained through other methods or means. This application embodiment does not limit this.

[0083] 104. Determine the noise influence factor of the original image based on classification distribution parameters and tissue distribution parameters.

[0084] The difference between the classification distribution parameters and tissue distribution parameters of the original image obtained through the above operations reflects the distribution of each category in the original image after noise influence, and the original tissue distribution information. If there is a large difference between the two, it means that the noise has a significant impact on the tissue distribution of the original image itself. Therefore, the noise influence factor is obtained.

[0085]

[0086] In the formula, It is a noise influence factor, which shows that it is used to represent the difference between classification distribution parameters and tissue distribution parameters.

[0087] In cardiovascular OCT images, texture patterns and geometric shapes should ideally match closely; noise can disrupt this match and affect classification distribution parameters. A larger value indicates a more regular texture classification and a better match with the normal tissue texture distribution; tissue distribution parameter The larger the value, the more regular the geometric shape and the better it matches the geometric distribution of normal tissue; differences This indicates the degree of deviation between two features under normal tissue conditions. The greater the deviation, the more severe the texture and geometric misalignment caused by noise, and the greater the impact of noise. The smaller the deviation, the higher the matching between the texture regularity and geometric morphology feature distributions, and the smaller the degree of noise influence.

[0088] 105. Perform non-local mean filtering on the original image based on the noise impact factor to obtain the target image.

[0089] After obtaining the noise impact factor, the similarity weight of the nonlocal mean filter can be adjusted using the obtained noise impact factor to perform nonlocal mean filtering on the original image, and finally obtain a smooth and clear target image.

[0090] For example, the larger the noise impact factor, the lower the similarity weight in the nonlocal mean filter should be, thereby eliminating the influence of speckle noise. For instance, when the noise impact factor is 1 and the similarity weight in the nonlocal mean filter is 3, the similarity weight in the nonlocal mean filter can be adaptively reduced to 2; similarly, when the noise impact factor is 0.5 and the similarity weight in the nonlocal mean filter is 3, the similarity weight in the nonlocal mean filter can be reduced to 2.5.

[0091] It should be understood that the numerical descriptions of the noise impact factor and similarity weights in nonlocal value filtering are merely illustrative. In practical applications, the settings should be tailored to the specific application scenario, and no restrictions are imposed here.

[0092] It should be understood that similarity weights can also be adjusted according to other rules or mapping relationships, and this application embodiment does not limit this.

[0093] 106. Enhance the target image.

[0094] After acquiring the target image, the target image is enhanced based on histogram equalization.

[0095] Specifically, the grayscale histogram of the target image is normalized to obtain the cumulative probability density function for each grayscale level. Based on the cumulative density function, the grayscale level of each pixel in the target image is replaced with the corresponding cumulative probability density function value. This completes the redistribution of grayscale levels in the target image. Based on the above mapping relationship, the grayscale level of each pixel in the target image is replaced with the corresponding new grayscale value to obtain the enhanced image.

[0096] In this invention, optical coherence tomography (OCT) images are acquired, along with their classification and tissue distribution parameters. The classification parameters are obtained by classifying the original image based on its brightness variations, while the tissue distribution parameters are obtained by segmenting the original image based on its tissue distribution. Then, the noise impact factor of the original image is determined based on the difference between the classification and tissue distribution parameters. Non-local mean filtering is then applied to the original image based on this noise impact factor to obtain a clear target image with suppressed speckle noise. Finally, the target image is enhanced to ensure that the enhanced OCT image retains important structural details while reducing unnecessary noise interference.

[0097] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0098] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for filtering and enhancing OCT images in cardiovascular medicine, characterized in that, The method includes: Acquire the original image, which is an optical coherence tomography (OCT) image; Obtain the classification distribution parameters of the original image, wherein the classification distribution parameters are obtained by classifying the original image based on the brightness and darkness variations of the original image; Obtain the tissue distribution parameters of the original image, wherein the tissue distribution parameters are obtained by segmenting the original image based on the tissue distribution of the original image; The noise influence factor of the original image is determined based on the classification distribution parameters and the tissue distribution parameters, and the noise influence factor is used to represent the difference between the classification distribution parameters and the tissue distribution parameters; The original image is subjected to nonlocal mean filtering based on the noise impact factor to obtain the target image; Enhance the target image; The process of obtaining the classification distribution parameters of the original image includes: Obtain the brightness fluctuation parameters of each pixel in the original image, and the brightness fluctuation parameters are determined based on the rate of change of the gray value of the pixel. Based on the brightness fluctuation parameters of each pixel, the pixels are clustered to obtain multiple first classification clusters and a first pixel cluster contained in the multiple first classification clusters; The image distribution factor of each first classification cluster is determined based on the relationship between the first pixel cluster in each of the plurality of first classification clusters and the center point of the first classification cluster. The classification distribution parameters of the original image are determined based on the image distribution factor of each first classification cluster; The process of obtaining the tissue distribution parameters of the original image includes: The original image after the closing operation is segmented to obtain multiple clusters of second pixel points; The trapping parameters of each second pixel cluster are determined based on the plurality of second pixel clusters; The tissue distribution parameters of the original image are determined based on the loop parameters of each second pixel cluster.

2. The method for filtering and enhancing cardiovascular OCT images according to claim 1, characterized in that, The brightness fluctuation parameter is determined based on the distance between each maximum point in the original image, the gray value of the maximum point, the average gray value of the pixels between the maximum points, and the number of pixels between the maximum points. The maximum point is the pixel with the largest gray value in the neighborhood of each pixel.

3. The method for filtering and enhancing cardiovascular OCT images according to claim 2, characterized in that, The image distribution factor is determined based on the information entropy of the angle between the first pixel cluster in each first classification cluster and the center point, and the variance of the distance from the center point.

4. The method for filtering and enhancing cardiovascular OCT images according to claim 3, characterized in that, The classification distribution parameters are determined based on the average value of the image distribution factor of each first classification cluster and the information entropy of the average center distance of the first pixel cluster in each first classification cluster.

5. The method for filtering and enhancing cardiovascular OCT images according to claim 1, characterized in that, The trap parameters are determined based on the area of ​​the minimum circumcircle of the second pixel cluster and the number of pixels belonging to the second pixel cluster in the minimum circumcircle.

6. The method for filtering and enhancing cardiovascular OCT images according to claim 1, characterized in that, The tissue distribution parameters are determined based on the information entropy of the product of the minimum circumcircle diameter of the second pixel cluster and the loop parameter, as well as the variance of the center of the minimum circumcircle of the second pixel cluster.

7. The method for filtering and enhancing cardiovascular OCT images according to claim 3, characterized in that, The enhancement of the target image includes: The target image is enhanced based on histogram equalization.

8. The method for filtering and enhancing cardiovascular OCT images according to claim 3, characterized in that, The original image is a cardiovascular image.

Images