Lung CT image enhancement method

Abstract

The invention relates to the technical field of image enhancement processing, in particular to a lung CT image enhancement method, which comprises the following steps: acquiring a lung CT image, a lung target region and a to-be-enhanced region; according to the gray level distribution of each pixel point in the region to be enhanced in the neighborhood and the pixel point distribution of the lung target region, analyzing the relevance of the gray level extension trend between the pixel points in the region to be enhanced to obtain the extension similarity degree between every two pixel points in the region to be enhanced; according to the pixel point distribution in the gray extension direction of each pixel point in the to-be-enhanced area and the position distribution between other pixel points and the corresponding gray extension direction, combining the extension similarity degree to obtain the comprehensive similarity degree between every two pixel points in the to-be-enhanced area; and according to the comprehensive similarity between each pixel point and other pixel points in the to-be-enhanced area, adjusting the gray value of the pixel point to obtain an enhanced image of the lung CT. According to the method, the image enhancement effect on the lung CT image is relatively good.

Description

Technical Field

[0001] The present invention relates to the technical field of image enhancement processing, and in particular to a lung CT image enhancement method. Background Art

[0002] In the medical field, lung CT images are an important tool for diagnosing and treating lung diseases. Traditional lung CT images have limitations in clarity and noise control, which makes it difficult for doctors to diagnose and treat. Especially in the field of lung treatment such as radiotherapy and radiation therapy, accurate imaging information is crucial to assist doctors in formulating treatment plans.

[0003] In lung CT images, since the contrast between important tissues such as blood vessels and other lung tissues is low, the lung CT images need to be enhanced in practice to show the lung tissues more intuitively and clearly for doctors to observe. However, the existing technology for enhancing CT images is mainly through analyzing the grayscale distribution of pixels, and does not combine the characteristic extension direction of tissues such as blood vessels that are more densely distributed in lung CT images, resulting in poor enhancement effect on the denser blood vessel parts in lung CT images. Summary of the invention

[0004] In order to solve the technical problem that the existing method has a poor enhancement effect on the denser blood vessels in the lung CT image, the purpose of the present invention is to provide a lung CT image enhancement method, and the technical solution adopted is as follows: Acquire lung CT images, and perform segmentation preprocessing on the lung CT images to obtain lung target areas and areas to be enhanced; According to the grayscale distribution of each pixel in the area to be enhanced within the neighborhood and the pixel distribution in the lung target area, the correlation of the grayscale extension trend between the pixels in the area to be enhanced is analyzed to obtain the extension similarity between every two pixels in the area to be enhanced; According to the distribution of pixels in the gray extension direction of each pixel in the area to be enhanced and the position distribution between other pixels and the corresponding gray extension direction, combined with the extension similarity, the comprehensive similarity between every two pixels in the area to be enhanced is obtained; According to the comprehensive similarity between each pixel and other pixels in the area to be enhanced, the gray value of the pixel is adjusted to obtain an enhanced image of the lung CT.

[0005] Preferably, the step of analyzing the correlation of gray extension trends between pixels in the region to be enhanced according to the gray distribution of each pixel in the region to be enhanced within the neighborhood and the pixel distribution in the lung target region to obtain the extension similarity between every two pixels in the region to be enhanced specifically includes: According to the grayscale distribution of each pixel in the area to be enhanced within the neighborhood and the pixel distribution in the lung target area, region growing is performed on each pixel in the area to be enhanced to obtain the grayscale feature region of each pixel in the area to be enhanced; Obtaining a gray extension path for each pixel in the area to be enhanced according to the path distribution from the corresponding gray feature area to the lung target area for each pixel in the area to be enhanced; According to the grayscale distribution difference of the pixels on the grayscale extension path of each two pixels in the area to be enhanced, the chain code information difference of the pixels and the length difference of the grayscale extension path, the extension similarity between each two pixels in the area to be enhanced is obtained.

[0006] Preferably, the method for acquiring the grayscale feature area is specifically: Each pixel point in the area to be enhanced is used as an initial seed point, and region growing is performed according to a set region growing rule to obtain a grayscale feature region corresponding to each pixel point in the area to be enhanced; Among them, the regional growing rule is: if the difference between the grayscale values ​​of the initial seed point and the pixels in its neighborhood is less than or equal to the preset difference threshold, the pixel point corresponding to the minimum grayscale value in the neighborhood of the initial seed point is used as a new seed point for growth until there is a pixel point in the lung target area within the neighborhood of the growing seed point, or the difference between the grayscale values ​​of the growing seed point and the pixels in its neighborhood is greater than the difference threshold, then the growth is stopped.

[0007] Preferably, obtaining the gray extension path of each pixel point in the area to be enhanced according to the path distribution from the corresponding gray feature area to the lung target area of ​​each pixel point in the area to be enhanced specifically includes: Any pixel point in the area to be enhanced is taken as the selected pixel point, and the pixel point with the shortest distance to the lung target area in the grayscale feature area of ​​the selected pixel point is obtained as the feature pixel point. In the grayscale feature area of ​​the selected pixel point, the shortest path between the selected pixel point and the corresponding feature pixel point is taken as the grayscale extension path of the selected pixel point.

[0008] Preferably, obtaining the extension similarity between every two pixels in the area to be enhanced according to the grayscale distribution difference of the pixels on the grayscale extension path of every two pixels in the area to be enhanced, the chain code information difference of the pixels and the length difference of the grayscale extension path specifically includes: Obtain an 8-chain code sequence of the gray extension path of each pixel in the area to be enhanced; record any two different pixels in the area to be enhanced as the first pixel and the second pixel respectively; Determine a grayscale difference factor based on a difference between grayscale value means of all pixels on a grayscale extension path of a first pixel and a second pixel; Determine a chain code difference factor based on a difference between the means of all chain code values ​​in the 8-chain code sequence corresponding to the first pixel and the second pixel; Determining a path difference factor based on a length difference between grayscale extension paths of the first pixel and the second pixel; A negative correlation process is performed on the product of the grayscale difference factor, the chain code difference factor and the path difference factor to obtain the extension similarity between the first pixel point and the second pixel point.

[0009] Preferably, the method of obtaining the comprehensive similarity between every two pixels in the area to be enhanced based on the distribution of pixels in the grayscale extension direction of each pixel in the area to be enhanced and the position distribution between other pixels and the corresponding grayscale extension direction, combined with the extension similarity, specifically includes: According to the number distribution of other pixels outside the gray extension path of each pixel in the area to be enhanced and the fluctuation of the distance between other pixels and the gray extension path, combined with the length distribution of the gray extension path, the representative characteristic index of each pixel in the area to be enhanced is obtained; According to the representative characteristic index of every two pixels in the area to be enhanced, the extension similarity between every two pixels is adjusted to obtain the comprehensive similarity between every two pixels in the area to be enhanced.

[0010] Preferably, the method for acquiring the representative feature index of each pixel in the area to be enhanced is specifically as follows: In the grayscale feature area of ​​the selected pixel, the variance of the shortest distance between all other pixel points except the grayscale extension path of the selected pixel point and the grayscale extension path is calculated to obtain a first coefficient; In the grayscale feature area of ​​the selected pixel, the number of all other pixel points except the grayscale extension path of the selected pixel point is obtained to obtain a second coefficient; A representative characteristic index of the selected pixel point is obtained according to the first coefficient, the second coefficient and the length of the grayscale extension path of the selected pixel point, the length of the grayscale extension path is positively correlated with the representative characteristic index, and the first coefficient and the second coefficient are both negatively correlated with the representative characteristic index.

[0011] Preferably, the step of adjusting the extended similarity between every two pixels in the region to be enhanced according to the representative feature index of every two pixels in the region to be enhanced to obtain the comprehensive similarity between every two pixels in the region to be enhanced specifically includes: The product of the representative feature index of the first pixel, the representative feature index of the second pixel, and the extended similarity between the first pixel and the second pixel is taken as the comprehensive similarity between the first pixel and the second pixel.

[0012] Preferably, the gray value of the pixel is adjusted according to the comprehensive similarity between each pixel and other pixels in the area to be enhanced to obtain an enhanced image of the lung CT, specifically including: Taking any pixel point in the area to be enhanced as the target pixel point, taking the pixels other than the target pixel point in the area to be enhanced as reference pixels, arranging the reference pixels in descending order of the comprehensive similarity between the target pixel point and each reference pixel point, and obtaining a preset number of reference pixels in the arrangement order as reference pixels of the target pixel point; Using the comprehensive similarity between the target pixel and each reference pixel, the grayscale value of each reference pixel is weighted and averaged to obtain the enhanced grayscale value of the target pixel; The enhanced gray value of each pixel in the area to be enhanced constitutes the enhanced image of lung CT.

[0013] Preferably, the preprocessing of the lung CT image by segmentation to obtain the lung target area and the area to be enhanced specifically includes: The maximum inter-class variance method is used to segment the lung CT images. The area with a gray value greater than the segmentation threshold is taken as the lung target area, and the area with a gray value less than or equal to the segmentation threshold is taken as the area to be enhanced.

[0014] The embodiments of the present invention have at least the following beneficial effects: The present invention first performs a preliminary regional division operation on the collected lung CT images, preliminarily screens out some relatively fuzzy partial areas that need to be enhanced, and provides a data basis for the subsequent feature analysis process. Then, in the first aspect, the correlation and similarity of the grayscale extension features of the pixels in the area to be enhanced are analyzed for the grayscale distribution of the pixels in the neighborhood and the pixel distribution in the target area of ​​the lungs, and the extension similarity between the two pixels is preliminarily constructed, which characterizes the similarity of the grayscale extension rules between the two pixels in the corresponding area. Further, on the basis of the similarity of the grayscale extension rules of the preliminary analysis, the second direction further combines the pixel position distribution outside the grayscale extension return of the pixels, comprehensively characterizes the similarity, and can more accurately reflect the feature similarity between the pixels in the area to be enhanced. Finally, the grayscale value of each pixel is adjusted in combination with the grayscale value of other pixels with relatively similar characteristics, which can effectively enhance the more fuzzy details in the image, and fully consider and analyze the characteristics of the blood vessel extension rules of the lungs, so that the image enhancement effect of the lung CT image is better. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a flowchart of the steps of a lung CT image enhancement method provided by the present invention; Figure 2 is a flowchart of the steps of a method for acquiring the degree of extension similarity between every two pixel points in the area to be enhanced provided by the present invention; Figure 3 It is a flowchart of the steps of the method for obtaining the comprehensive similarity between every two pixel points in the area to be enhanced provided by the present invention. DETAILED DESCRIPTION

[0017] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of a lung CT image enhancement method proposed by the present invention, its specific implementation, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0018] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0019] The specific scheme of the lung CT image enhancement method provided by the present invention is described in detail below with reference to the accompanying drawings.

[0020] See also Figure 1 , which shows a flowchart of a method for enhancing lung CT images provided by an embodiment of the present invention, the method comprising the following steps: Step S100, acquiring a lung CT image, and performing segmentation preprocessing on the lung CT image to obtain a lung target area and an area to be enhanced.

[0021] In the case of electromagnetic or other noise interference, the presence of noise in lung CT images will cause aliasing between the detailed information in the lungs and the background information, thereby blurring or hiding the information in the CT image and causing the loss of image detail information. Therefore, it is necessary to further enhance the more detailed blood vessels in the lung CT image.

[0022] It should be noted that in the human lungs, the main blood vessels are the pulmonary artery and pulmonary vein, bronchial artery and bronchial vein, and pulmonary capillaries. Among them, the pulmonary artery trunk starts from the right ventricle and divides into two pulmonary arteries below the aortic arch. The pulmonary artery runs along with the bronchi at all levels, and finally forms a capillary network distributed on the alveolar wall. At the same time, the thickness of the blood vessels gradually decreases from the pulmonary artery to the capillaries.

[0023] Under normal circumstances, the pulmonary arteries and bronchial arteries in lung CT images are thicker and more sparsely distributed, so they are clearer and easier to distinguish. However, the pulmonary capillaries are thinner and more densely distributed, so they are easily affected by noise and blurred, making them more difficult to distinguish.

[0024] Capillaries usually extend gradually from the pulmonary artery to the bronchial artery and then to the capillaries. Therefore, the distribution of blood vessels has a distribution pattern in the extension direction, and the overall distribution direction of capillaries in the lungs is also distributed along the extension direction of the blood vessels, that is, from the middle part of the lungs to the surrounding areas. Therefore, when enhancing the lung CT images, the lung CT images can be preliminarily divided and the parts with low contrast and unclearness can be preliminarily screened out to further analyze the extension distribution pattern of the blood vessels.

[0025] Specifically, a lung CT image is first obtained. In order to better perform the image analysis process, the lung CT image in this embodiment is a grayscale image. Then, the lung CT image is segmented and preprocessed to obtain a lung target area and a region to be enhanced. The lung target area is the clearer target part in the lung CT image, and the region to be enhanced is the more blurred part to be enhanced in the lung CT image.

[0026] More specifically, the maximum inter-class variance method is used to perform regional segmentation on the lung CT image, that is, the maximum variance method is used to obtain the segmentation threshold of the lung CT image, and the lung CT image is segmented by the segmentation threshold, and the segmentation result is preprocessed, such as morphological corrosion operation, to eliminate the local isolated segmentation area, and finally obtain the lung target area and the area to be enhanced. Specifically, the area with a gray value greater than the segmentation threshold is taken as the lung target area, and the area with a gray value less than or equal to the segmentation threshold is taken as the area to be enhanced.

[0027] Step S200, based on the grayscale distribution of each pixel in the area to be enhanced within the neighborhood range and the pixel distribution in the lung target area, the correlation of the grayscale extension trend between the pixels in the area to be enhanced is analyzed to obtain the extension similarity between every two pixels in the area to be enhanced.

[0028] The actual distribution of the area to be enhanced includes the capillary part and the background part which are relatively fuzzy and difficult to distinguish. The capillaries in the lung area have a certain extension pattern and have a certain connection relationship with the target part of the lung. Based on this, in order to better enhance the vascular features of the fuzzy part to distinguish the background and the target blood vessels, it is first necessary to analyze the distribution of the extension pattern of each pixel point based on the grayscale value in the area to be enhanced, determine the grayscale extension direction at each pixel point in the area, and further analyze the feature differences and feature associations between the pixels in the area based on the grayscale extension direction.

[0029] In this embodiment, if Figure 2 As shown, the method for acquiring the extension similarity between every two pixel points in the area to be enhanced can be implemented by steps S201 to S203.

[0030] Step S201, performing region growing on each pixel in the region to be enhanced according to the grayscale distribution of each pixel in the region to be enhanced within the neighborhood range and the pixel distribution in the lung target region, to obtain the grayscale feature region of each pixel in the region to be enhanced.

[0031] Specifically, each pixel point in the area to be enhanced is used as an initial seed point, and regional growth is performed according to the set regional growth rule to obtain the grayscale feature area corresponding to each pixel point in the area to be enhanced; the grayscale feature area includes all the seed points that have been used in the process of each pixel point growing according to the regional growth rule.

[0032] Among them, the regional growing rule is: if the difference between the grayscale values ​​of the initial seed point and the pixels in its neighborhood is less than or equal to the preset difference threshold, the pixel point corresponding to the minimum grayscale value in the neighborhood of the initial seed point is used as a new seed point for growth until there is a pixel point in the lung target area within the neighborhood of the growing seed point, or the difference between the grayscale values ​​of the growing seed point and the pixels in its neighborhood is greater than the difference threshold, then the growth is stopped.

[0033] In this embodiment, the difference threshold is set to 30, and the neighborhood range is set to 8 neighborhoods. In other embodiments, the implementer can set it according to the specific implementation scenario. More specifically, for any initial seed point, the absolute value of the grayscale value difference between the initial seed point and the pixel points in the 8 neighborhoods is calculated. When the absolute value of the grayscale value difference is less than or equal to 30, the pixel point corresponding to the minimum absolute value of the grayscale value difference in the 8 neighborhoods of the initial seed point is used as a new seed point for growth. Further, the grayscale value difference in the 8 neighborhoods of the new seed point is analyzed to see whether it meets the threshold for regional growth.

[0034] At this point, after performing region growth on each pixel point in the region to be enhanced, a corresponding grayscale feature region is obtained, which represents the distribution of the region where the grayscale is extended from the position of each pixel point.

[0035] Step S202, obtaining a gray extension path for each pixel in the area to be enhanced according to the path distribution from the corresponding gray feature area to the lung target area for each pixel in the area to be enhanced.

[0036] Considering that in the regional growth process of the pixel points in the area to be enhanced, the seed points grow within the 8-neighborhood and have many extension directions, the final growth area may not only be a linear area, but also an area with a certain area. Therefore, it is necessary to further analyze the path distribution of the pixel points in the gray feature area of ​​each pixel point in the area to be enhanced, and screen out the main extension lines to represent the gray extension direction of the pixel point.

[0037] Specifically, any pixel point in the area to be enhanced is taken as the selected pixel point, and the pixel point with the shortest distance to the lung target area in the grayscale feature area of ​​the selected pixel point is obtained as the feature pixel point. In the grayscale feature area of ​​the selected pixel point, the shortest path between the selected pixel point and the corresponding feature pixel point is taken as the grayscale extension path of the selected pixel point.

[0038] It should be noted that the method for obtaining the shortest path between two pixels is a well-known technology, such as the A* algorithm, etc., which will not be introduced here. The implementer can choose according to the specific implementation scenario. At this point, the characterization of the extension direction of each pixel in the area to be enhanced in the corresponding grayscale feature area can be obtained, that is, the corresponding grayscale extension direction. In the grayscale extension direction, the feature pixel is the end point of the path.

[0039] Step S203, obtaining the extension similarity between every two pixels in the area to be enhanced according to the grayscale distribution difference of the pixels on the grayscale extension path of every two pixels in the area to be enhanced, the chain code information difference of the pixels and the length difference of the grayscale extension path.

[0040] In the area to be enhanced, if the pixels belong to the same blood vessel part or the same background part, the grayscale extension of these pixels belonging to the same part will have certain similarities and correlations. Based on multiple feature considerations, the grayscale distribution, extension direction and displacement distribution of the shortest path in the grayscale feature area of ​​each pixel in the area to be enhanced are relatively similar. Combining the feature differences and feature similarities of each two pixels in the area to be enhanced in these aspects, the similarity and correlation of each two pixels in the extension feature rules can be quantitatively characterized.

[0041] In this embodiment, any two different pixels in the area to be enhanced are taken as an example for explanation, that is, any two different pixels in the area to be enhanced are respectively recorded as the first pixel and the second pixel. It can be understood that the 8-chain code sequence of the gray extension path of each pixel in the area to be enhanced is first obtained. The method for obtaining the 8-chain code sequence is a well-known technology and will not be introduced in detail here. The chain code value can characterize the connection direction of the corresponding edge to a certain extent.

[0042] Specifically, a grayscale difference factor is determined based on a difference between the mean values ​​of the grayscale values ​​of all pixels on the grayscale extension path of the first pixel and the second pixel; a chain code difference factor is determined based on a difference between the mean values ​​of all chain code values ​​in the 8-chain code sequence corresponding to the first pixel and the second pixel; a path difference factor is determined based on a difference in length between the grayscale extension paths of the first pixel and the second pixel; and a negative correlation process is performed on the product of the grayscale difference factor, the chain code difference factor and the path difference factor to obtain the extension similarity between the first pixel and the second pixel.

[0043] As a specific example, the x-th pixel point in the area to be enhanced is taken as the first pixel point, and the y-th pixel point in the area to be enhanced is taken as the second pixel point. The calculation formula of the extension similarity between the first pixel point and the second pixel point can be specifically expressed as: in, represents the degree of extension similarity between the first pixel and the second pixel, x represents the x-th pixel in the area to be enhanced, y represents the y-th pixel in the area to be enhanced, and ; represents the mean grayscale value of all pixels on the grayscale extension path corresponding to the grayscale feature area of ​​the first pixel, The grayscale value average of all pixels on the grayscale extension path corresponding to the grayscale feature area of ​​the second pixel; Represents the mean of all chain code values ​​in the 8-chain code sequence of the grayscale extension path corresponding to the grayscale feature area of ​​the first pixel, The mean value of all chain code values ​​in the 8-chain code sequence of the grayscale extension path corresponding to the grayscale feature area of ​​the second pixel; Represents the total number of all pixels on the grayscale extension path corresponding to the grayscale feature area of ​​the first pixel, represents the total number of all pixels on the grayscale extension path corresponding to the grayscale feature area of ​​the second pixel, Represents an exponential function with the natural constant e as the base.

[0044] It represents the grayscale difference factor, which reflects the grayscale difference between two pixels on the corresponding grayscale extension path; and Both reflect the overall displacement of the edge on the grayscale extension path corresponding to the pixel point. It represents the chain code difference factor, which reflects the difference in edge direction between two pixels on the corresponding grayscale extension path; It represents the path difference factor, which reflects the length difference between two pixels on the corresponding grayscale extension path.

[0045] The greater the difference between any two pixels in these three aspects, the smaller the similarity and correlation between the two corresponding pixels in the gray extension law, and the smaller the corresponding extension similarity value. Based on this, the negative exponential function is used for negative correlation processing. The extension similarity analyzes the feature differences of the two pixels in three aspects, and preliminarily characterizes the similarity and correlation of the two pixels in the direction of the gray extension law.

[0046] Step S300, according to the distribution of pixels in the gray extension direction of each pixel in the area to be enhanced and the position distribution between other pixels and the corresponding gray extension direction, combined with the extension similarity, the comprehensive similarity between every two pixels in the area to be enhanced is obtained.

[0047] The preliminary analysis of the extension similarity between each two pixels only considers the feature similarity of the grayscale extension path of the pixels in the area to be enhanced. It is necessary to further combine the distance distribution and quantity distribution of other pixels except the grayscale extension path to analyze whether the size of the feature expression of the current grayscale extension path for the grayscale feature area of ​​the corresponding pixel has a high representative feature, and further adjust the correlation relationship and similarity relationship between each two pixels.

[0048] Based on this, Figure 3 As shown, the method for acquiring the comprehensive similarity between every two pixels in the area to be enhanced can be implemented by step S301 and step S302.

[0049] Step S301, according to the number distribution of other pixels outside the grayscale extension path of each pixel in the area to be enhanced and the distance fluctuation between other pixels and the grayscale extension path, combined with the length distribution of the grayscale extension path, a representative feature index of each pixel in the area to be enhanced is obtained.

[0050] Under normal circumstances, the distribution of blood vessels in the image is generally manifested as a certain grayscale distribution around the main extension direction of the region, and the distribution area is relatively small. Based on this, when analyzing the other pixels in the grayscale feature area of ​​the pixel points in the area to be enhanced except for the grayscale extension path, the fewer the number of distributions, the clearer the main extension direction can express the grayscale extension law of the pixel point location, and the closer the pixel point location is to the grayscale extension path, the more there are some pixel points distributed in the part closer to the grayscale extension path, which is more consistent with the characteristic distribution phenomenon of the blood vessel part.

[0051] Based on this, combined with the features of multiple aspects, the degree of characteristic performance of the grayscale extension path of each pixel in the grayscale feature area is quantified. Specifically, in the grayscale feature area of ​​the selected pixel, the variance of the shortest distance between all other pixel points except the grayscale extension path of the selected pixel and the grayscale extension path is calculated to obtain the first coefficient; in the grayscale feature area of ​​the selected pixel, the number of all other pixel points except the grayscale extension path of the selected pixel is obtained to obtain the second coefficient; according to the first coefficient, the second coefficient and the length of the grayscale extension path of the selected pixel, the representative characteristic index of the selected pixel is obtained, the length of the grayscale extension path is positively correlated with the representative characteristic index, and the first coefficient and the second coefficient are both negatively correlated with the representative characteristic index.

[0052] As a specific example, take any pixel in the area to be enhanced as an example, and take the i-th pixel in the area to be enhanced as the selected pixel. Then the representative feature index of the i-th pixel in the area to be enhanced, i.e. the selected pixel, can be expressed as: in, represents the representative characteristic index of the i-th pixel point in the area to be enhanced, i.e., the selected pixel point, and i represents the i-th pixel point in the area to be enhanced; The variance of the shortest distances from all other pixels except the gray extension path of the selected pixel in the gray feature area of ​​the selected pixel to the gray extension path; The total number of all pixels on the grayscale extension path corresponding to the grayscale feature area of ​​the selected pixel, that is, the path length; Represents the total number of all other pixels in the grayscale feature area of ​​the selected pixel, except for the grayscale extension path of the selected pixel. Represents the normalization function.

[0053] is the first coefficient, which reflects the degree of discreteness of the distance distribution from other pixels to the current main extension direction in the gray feature area of ​​the selected gray pixel. is the second coefficient, which reflects the number distribution of other pixels. The larger the values ​​of the two, the more discrete the distribution of pixels in the non-main extension direction in the gray feature area is relative to the main extension direction, and the larger the number is. This further indicates that the current main extension direction is more difficult to represent the gray extension law of the gray feature area, that is, the corresponding representativeness is worse. At the same time, The larger the value of is, the more pixel representations there are in the main extension direction, the more representative features are likely to be expressed, and thus the larger the value of the representative feature index is.

[0054] At this point, the representative feature index characterizes whether the grayscale extension path of the pixel points in the area to be enhanced has a high degree of representativeness for the grayscale feature area, so as to facilitate the subsequent screening of highly representative pixels and weight the similarity and correlation between the pixels.

[0055] Step S302: adjusting the extended similarity between every two pixels in the region to be enhanced according to the representative characteristic index of every two pixels in the region to be enhanced, so as to obtain the comprehensive similarity between every two pixels in the region to be enhanced.

[0056] Specifically, taking any two different pixel points in the area to be enhanced as an example, the product of the representative feature index of the first pixel point, the representative feature index of the second pixel point and the extended similarity between the first pixel point and the second pixel point is taken as the comprehensive similarity between the first pixel point and the second pixel point.

[0057] More specifically, the comprehensive similarity can be expressed as: , Indicates the comprehensive similarity between the first pixel and the second pixel. Represents the representative feature index of the first pixel, Represents the representative feature index of the second pixel, represents the degree of extension similarity between the first pixel and the second pixel, x represents the x-th pixel in the area to be enhanced, y represents the y-th pixel in the area to be enhanced, and .

[0058] It can be understood that the comprehensive similarity not only takes into account the feature similarity and correlation between the two pixels, but also takes into account the feature representativeness of the pixels themselves, so that when grayscale enhancement is performed based on similarity later, the grayscale features of the pixels are more representative.

[0059] Step S400, adjusting the gray value of each pixel in the area to be enhanced according to the comprehensive similarity between each pixel and other pixels, to obtain an enhanced image of the lung CT.

[0060] The greater the similarity between each pixel and other pixels in the area to be enhanced, and the more representative the features corresponding to the two pixels are, the more likely it is that the grayscale extension paths corresponding to the two pixels are more similar, which means that it is more likely to be a more detailed part such as capillaries, that is, the more it is necessary to adjust the grayscale values ​​of these pixels.

[0061] Specifically, any pixel point in the area to be enhanced is taken as the target pixel point, and the pixel points other than the target pixel point in the area to be enhanced are taken as reference pixel points. The reference pixel points are arranged in order from large to small according to the comprehensive similarity between the target pixel point and each reference pixel point, and a preset number of reference pixel points are obtained in the arrangement order as the reference pixel points of the target pixel point.

[0062] In this embodiment, the preset number is set to 20. In other embodiments, the implementer can set it according to the specific implementation scenario. More specifically, the reference pixel points characterize the pixel points in the area to be enhanced whose grayscale extension direction distribution characteristics are most similar to the target pixel points. The target pixel points are copied using the grayscale values ​​of these pixel points, which can effectively integrate some grayscale information with representative characteristics, so that the adjusted grayscale value can better express the detail information.

[0063] Furthermore, the grayscale value of each reference pixel is weighted and averaged using the comprehensive similarity between the target pixel and each reference pixel to obtain the enhanced grayscale value of the target pixel. In this embodiment, the nth pixel in the area to be enhanced is taken as the target pixel, and the enhanced grayscale value of the target pixel can be expressed by the formula: ,in represents the enhanced gray value of the target pixel, and n represents the nth pixel in the area to be enhanced; Represents the total number of reference pixels for the target pixel, Indicates the comprehensive similarity between the target pixel and the kth reference pixel. Represents the grayscale value of the target pixel and the kth comparison pixel.

[0064] It should be noted that when the value of the enhanced grayscale value is not an integer, it is necessary to perform an upward rounding operation. This method is a well-known technology and will not be introduced in detail here. Finally, the enhanced grayscale value of each pixel in the area to be enhanced is obtained according to the same method, and then the enhanced grayscale value of each pixel in the area to be enhanced constitutes the enhanced image of the lung CT. The enhanced lung CT image can have clearer vascular details and better image enhancement effect.

[0065] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A lung CT image enhancement method, characterized in that: The method comprises the following steps: Acquire lung CT images, and perform segmentation preprocessing on the lung CT images to obtain lung target areas and areas to be enhanced; According to the grayscale distribution of each pixel in the area to be enhanced within the neighborhood and the pixel distribution in the lung target area, the correlation of the grayscale extension trend between the pixels in the area to be enhanced is analyzed to obtain the extension similarity between every two pixels in the area to be enhanced, including: According to the grayscale distribution of each pixel in the area to be enhanced within the neighborhood and the pixel distribution in the lung target area, regional growth is performed on each pixel in the area to be enhanced to obtain the grayscale feature area of ​​each pixel in the area to be enhanced; according to the path distribution from the corresponding grayscale feature area to the lung target area of ​​each pixel in the area to be enhanced, the grayscale extension path of each pixel in the area to be enhanced is obtained; according to the grayscale distribution difference of the pixels on the grayscale extension path of every two pixels in the area to be enhanced, the chain code information difference of the pixels and the length difference of the grayscale extension path, the extension similarity between every two pixels in the area to be enhanced is obtained; According to the distribution of pixels in the gray extension direction of each pixel in the area to be enhanced and the position distribution between other pixels and the corresponding gray extension direction, combined with the extension similarity, the comprehensive similarity between every two pixels in the area to be enhanced is obtained, including: According to the number distribution of other pixels outside the gray extension path of each pixel in the area to be enhanced, and the fluctuation of the distance between other pixels and the gray extension path, combined with the length distribution of the gray extension path, the representative feature index of each pixel in the area to be enhanced is obtained; according to the representative feature index of every two pixels in the area to be enhanced, the extension similarity between every two pixels is adjusted to obtain the comprehensive similarity between every two pixels in the area to be enhanced; According to the comprehensive similarity between each pixel and other pixels in the area to be enhanced, the gray value of the pixel is adjusted to obtain an enhanced image of the lung CT.

2. A lung CT image enhancement method according to claim 1, characterized in that: The method for obtaining the grayscale feature area is specifically as follows: Each pixel point in the area to be enhanced is used as an initial seed point, and region growing is performed according to a set region growing rule to obtain a grayscale feature region corresponding to each pixel point in the area to be enhanced; Among them, the regional growing rule is: if the difference between the grayscale values ​​of the initial seed point and the pixels in its neighborhood is less than or equal to the preset difference threshold, the pixel point corresponding to the minimum grayscale value in the neighborhood of the initial seed point is used as a new seed point for growth until there is a pixel point in the lung target area within the neighborhood of the growing seed point, or the difference between the grayscale values ​​of the growing seed point and the pixels in its neighborhood is greater than the difference threshold, then the growth is stopped.

3. A lung CT image enhancement method according to claim 1, characterized in that: The step of obtaining the gray extension path of each pixel point in the area to be enhanced according to the path distribution from the corresponding gray feature area to the lung target area of ​​each pixel point in the area to be enhanced specifically includes: Any pixel point in the area to be enhanced is taken as the selected pixel point, and the pixel point with the shortest distance to the lung target area in the grayscale feature area of ​​the selected pixel point is obtained as the feature pixel point. In the grayscale feature area of ​​the selected pixel point, the shortest path between the selected pixel point and the corresponding feature pixel point is taken as the grayscale extension path of the selected pixel point.

4. A lung CT image enhancement method according to claim 3, characterized in that: The method of obtaining the extension similarity between every two pixels in the area to be enhanced according to the grayscale distribution difference of the pixels on the grayscale extension path of every two pixels in the area to be enhanced, the chain code information difference of the pixels, and the length difference of the grayscale extension path specifically includes: Obtain an 8-chain code sequence of the gray extension path of each pixel in the area to be enhanced; record any two different pixels in the area to be enhanced as the first pixel and the second pixel respectively; Determine a grayscale difference factor based on a difference between grayscale value means of all pixels on a grayscale extension path of a first pixel and a second pixel; Determine a chain code difference factor based on a difference between the means of all chain code values ​​in the 8-chain code sequence corresponding to the first pixel and the second pixel; Determining a path difference factor based on a length difference between grayscale extension paths of the first pixel and the second pixel; A negative correlation process is performed on the product of the grayscale difference factor, the chain code difference factor and the path difference factor to obtain the extension similarity between the first pixel point and the second pixel point.

5. A lung CT image enhancement method according to claim 4, characterized in that: The method for obtaining the representative feature index of each pixel in the area to be enhanced is specifically as follows: In the grayscale feature area of ​​the selected pixel, the variance of the shortest distance between all other pixel points except the grayscale extension path of the selected pixel point and the grayscale extension path is calculated to obtain a first coefficient; In the grayscale feature area of ​​the selected pixel, the number of all other pixel points except the grayscale extension path of the selected pixel point is obtained to obtain a second coefficient; A representative characteristic index of the selected pixel point is obtained according to the first coefficient, the second coefficient and the length of the grayscale extension path of the selected pixel point, the length of the grayscale extension path is positively correlated with the representative characteristic index, and the first coefficient and the second coefficient are both negatively correlated with the representative characteristic index.

6. A lung CT image enhancement method according to claim 5, characterized in that: The step of adjusting the extended similarity between every two pixels in the region to be enhanced according to the representative characteristic index of every two pixels in the region to be enhanced to obtain the comprehensive similarity between every two pixels in the region to be enhanced specifically includes: The product of the representative feature index of the first pixel, the representative feature index of the second pixel, and the extended similarity between the first pixel and the second pixel is taken as the comprehensive similarity between the first pixel and the second pixel.

7. The lung CT image enhancement method according to claim 1, characterized in that: The gray value of each pixel is adjusted according to the comprehensive similarity between each pixel and other pixels in the area to be enhanced to obtain an enhanced image of the lung CT, specifically including: Taking any pixel point in the area to be enhanced as the target pixel point, taking the pixels other than the target pixel point in the area to be enhanced as reference pixels, arranging the reference pixels in descending order of the comprehensive similarity between the target pixel point and each reference pixel point, and obtaining a preset number of reference pixels in the arrangement order as reference pixels of the target pixel point; Using the comprehensive similarity between the target pixel and each reference pixel, the grayscale value of each reference pixel is weighted and averaged to obtain the enhanced grayscale value of the target pixel; The enhanced gray value of each pixel in the area to be enhanced constitutes the enhanced image of the lung CT.

8. The lung CT image enhancement method according to claim 1, characterized in that: The preprocessing of the lung CT image by segmentation to obtain the lung target area and the area to be enhanced specifically includes: The maximum inter-class variance method is used to segment the lung CT images. The area with a gray value greater than the segmentation threshold is taken as the lung target area, and the area with a gray value less than or equal to the segmentation threshold is taken as the area to be enhanced.

Images