Liver CT image enhancement method and system

Abstract

The invention relates to the technical field of image enhancement, in particular to a liver CT image enhancement method and system. The method comprises the following steps: acquiring an initial window of each connected domain; obtaining the density change degree of each initial window according to the density distribution of pixel points in each initial window on the CT image; obtaining a multi-scale image corresponding to the CT image, and obtaining the high-frequency information change degree of each initial window according to the frequency domain distribution characteristics of the initial windows corresponding to the same positions among different-scale images; obtaining a window correction coefficient of each initial window, correcting each initial window, and obtaining a correction window of each initial window; and performing adaptive histogram equalization on the images of different scales according to the correction window, and performing weighted fusion in combination with the confusion degree of the images of the corresponding scales to obtain an enhanced CT image. According to the method, the effect of influence enhancement is improved by obtaining the accurate window size during image equalization processing.

Description

Technical Field

[0001] The present invention relates to the technical field of image enhancement, and in particular to a method and system for enhancing liver CT images. Background Art

[0002] The segmentation and extraction of the liver area in abdominal CT images can be used in various fields such as medical teaching and clinical auxiliary treatment. However, the liver has the characteristics of low contrast with the surrounding adjacent organs, weak boundaries, and large differences in liver morphology between cases. Therefore, image enhancement of the liver in abdominal CT images still faces great difficulties.

[0003] Considering that the liver usually presents a relatively uniform grayscale distribution in CT images, in the prior art, adaptive histogram equalization is used, that is, based on the window size obtained empirically, the image is divided into several small areas, and the histogram is calculated and equalized for each small area to enhance the local contrast of the image. However, due to the size, shape and different pathological characteristics of the liver, there are different information detail characteristics. Using an inappropriate window size may lead to excessive or insufficient enhancement of the image, resulting in poor enhancement effect. Summary of the invention

[0004] In order to solve the technical problem that an inappropriate window size may lead to excessive or insufficient enhancement of the image, resulting in poor enhancement effect, the purpose of the present invention is to provide a liver CT image enhancement method and system, and the technical solution adopted is as follows: The present invention proposes a liver CT image enhancement method, the method comprising: Obtain CT images of the liver; The connected domain of the CT image is obtained according to the grayscale distribution of the pixels on the CT image; the initial window of each connected domain is obtained according to the morphological characteristics of each connected domain; the density change degree of each initial window is obtained according to the density distribution of the pixels in each initial window on the CT image; Obtain a multi-scale image corresponding to the CT image; obtain the degree of change of high-frequency information of each initial window according to the frequency domain distribution characteristics of each initial window at the same position between the different-scale images; obtain a window correction coefficient of each initial window according to the degree of change of density and the degree of change of high-frequency information of each initial window; correct each initial window according to the window correction coefficient to obtain a corrected window of each initial window; Adaptive histogram equalization is performed on images of different scales according to the correction window, and weighted fusion is performed in combination with the degree of confusion of the images of corresponding scales to obtain an enhanced CT image.

[0005] Furthermore, the method for obtaining the connected domain includes: CANNY edge detection is performed on the CT image to obtain the edge curves in the CT image, and the independent areas surrounded by all closed edge curves are regarded as connected domains.

[0006] Furthermore, the method for obtaining the initial window includes: Obtaining the square of the number of edge pixels in each connected domain as the number of square edge pixels; obtaining the ratio of the number of all pixels in each connected domain to the number of square edge pixels as the first morphological coefficient; obtaining the product of a preset ratio parameter and the first morphological coefficient as a shape factor; If the connected shape factor is greater than the preset factor threshold, the initial window of the corresponding connected domain is set to the preset first window. Otherwise, the initial window of the corresponding connected domain is set to the preset second window. The preset second window is larger than the preset first window. The initial window is centered on the centroid of the connected domain and constructs a square area of ​​equal width.

[0007] Furthermore, the method for obtaining the density change degree includes: Get the HU value of each pixel on the CT image; The mean value of the difference between the HU values ​​of different pixel points and the preset liver HU value was obtained as the density change degree of each initial window.

[0008] Furthermore, the method for obtaining the degree of change of the high-frequency information includes: For any scale image, obtain the frequency domain image of each initial window; The sum of the amplitudes of all high-frequency information in each frequency domain image is obtained as a first amplitude; the sum of the frequency domain amplitudes in each frequency domain image is obtained as a second amplitude; and the ratio of the first amplitude to the second amplitude is obtained as a high-frequency ratio; According to the difference in high-frequency ratios corresponding to each initial window at the same position between images of different adjacent scales, the degree of change of high-frequency information of each initial window is obtained, and the difference in high-frequency ratios is positively correlated with the degree of change of high-frequency information.

[0009] Furthermore, the method for obtaining the high-frequency information includes: For any frequency domain image, the relative distance between the pixel point with gray value not equal to 0 and the center point of the frequency domain image is obtained, and a distance histogram corresponding to the relative distance of all the pixel points with gray value not equal to 0 is constructed. The distance histogram is segmented using the Otsu method to obtain the segmented area of ​​the distance histogram. The largest relative distance mean in the segmented area is selected, and the corresponding pixel point is used as the high-frequency information in the spectrum image.

[0010] Furthermore, the method for obtaining the window correction coefficient includes: The density change degree and high-frequency information change degree of each initial window are fused to obtain the window correction coefficient corresponding to the initial window.

[0011] Furthermore, the method for obtaining the correction window includes: Obtain the difference between the positive integer 1 and the window correction coefficient as the adjustment weight; The product of the adjustment weight and the initial size of the initial window is obtained, and the sum of the product result and the initial size is calculated as the revised size to obtain the revised window.

[0012] Furthermore, the method for acquiring the multi-scale sub-image includes: Gaussian pyramid is used to obtain multi-scale images corresponding to CT images.

[0013] The present invention also proposes a liver CT image enhancement system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the system implements any one of the steps of the liver CT image enhancement method.

[0014] The present invention has the following beneficial effects: The present invention obtains the connected domain of the CT image according to the grayscale distribution of pixel points on the CT image, which is helpful to divide the image into areas with similar grayscale values ​​and analyze the local area more specifically; since the morphology and size of the liver are important factors for determining the analysis window, the initial window of each connected domain is obtained according to the morphological characteristics of each connected domain; according to the density distribution of pixel points in each initial window on the CT image, the density change degree of each initial window is obtained, which reflects the complexity of the tissue in the initial window; a multi-scale image corresponding to the CT image is obtained, which reflects the complexity of the tissue in the window or the degree of lesions; according to the frequency domain distribution characteristics of each initial window corresponding to the same position between different scale images, the high-frequency information change degree of each initial window is obtained, which reflects the edge and texture characteristics of the image in the window and helps to evaluate the complexity and detail richness of the image in the window; then the window correction coefficient of each initial window is obtained, each initial window is corrected, and the correction window of each initial window is obtained, which more accurately adapts to the characteristics of different tissue areas and optimizes the performance of histogram equalization; adaptive histogram equalization is performed on images of different scales according to the correction window, and weighted fusion is performed in combination with the chaos degree of the corresponding scale image to obtain an enhanced CT image. The present invention improves the effect of impact enhancement by obtaining an accurate window size during image equalization processing. 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 A flow chart of a liver CT image enhancement method provided by one embodiment of the present invention; Figure 2 A flow chart of a method for acquiring the degree of change of high-frequency information provided by an embodiment of 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 liver CT image enhancement method and system proposed by the present invention, its specific implementation, structure, features and effects are described in detail as follows 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 liver CT image enhancement method and system provided by the present invention is described in detail below with reference to the accompanying drawings.

[0020] See also Figure 1 , which shows a flow chart of a liver CT image enhancement method provided by an embodiment of the present invention, and the specific method includes: Step S1: Acquire a CT image of the liver.

[0021] In the embodiment of the present invention, in order to accurately analyze the liver region, it is first necessary to use a CT scanner to obtain a CT image of the liver at the cross section of the patient's liver region, and further preprocess and analyze. It should be noted that the processing method for each CT image is the same, which will not be repeated here, and only one CT image will be used as an example in the following.

[0022] Taking into account the noise introduced by the equipment, environment or scanning process, in order to facilitate the subsequent image processing process, the acquired CT image is preprocessed to enhance the image quality. It should be noted that the image preprocessing operation is a technical means well known to those skilled in the art, which can be specifically set according to the specific implementation scenario. In one embodiment of the present invention, bilateral filtering is used for preprocessing, balanced denoising and retaining edge information, which can highlight the contour and details of the image and make the image clearer. The image is then linearly normalized to unify the contrast and brightness differences between different images. The specific means are technical means well known to those skilled in the art and will not be described in detail here.

[0023] Step S2: obtaining a connected domain of the CT image according to the grayscale distribution of pixels on the CT image; obtaining an initial window of each connected domain according to the morphological characteristics of each connected domain; and obtaining a density variation degree of pixels in each initial window according to the density distribution of pixels in the CT image in each initial window.

[0024] Due to the complexity of the internal tissue structure of the human body, the grayscale distribution in CT images often shows unevenness, and there are significant differences in the grayscale values ​​between different tissues. Through connected domain analysis, different tissue structures in CT images can be identified, reducing the interference of noise on the analysis, and facilitating subsequent separate processing and analysis of each area. The connected domain of the CT image is obtained based on the grayscale distribution of pixels on the CT image.

[0025] Preferably, in one embodiment of the present invention, the method for acquiring the connected domain includes: The CANNY edge detection is performed on the CT image to obtain the edge curve in the CT image, and the area enclosed by the closed edge curve is taken as the connected domain.

[0026] It should be noted that the specific CANNY edge detection is a technical means well known to those skilled in the art and will not be elaborated here.

[0027] The morphology and size of the liver are important factors in determining the analysis window. Considering that the liver morphology is relatively regular and elliptical or circular, the morphological characteristics of the connected domains are analyzed. If the liver is large and morphologically normal, a larger window can be selected to facilitate global analysis. If the liver is small and has complex structural tissue or local lesions, a smaller window can be selected to more accurately analyze the details. The initial window of each connected domain is obtained based on the morphological characteristics of each connected domain.

[0028] Preferably, in one embodiment of the present invention, the method for obtaining the initial window includes: Obtain the square of the number of edge pixels in each connected domain as the number of square edge pixels; obtain the ratio of the number of all pixels in each connected domain to the number of square edge pixels as the first morphological coefficient; obtain the product of the preset ratio parameter and the first morphological coefficient as the shape factor; It should be noted that the number of edge pixels reflects the perimeter of the connected domain, and the number of all pixels reflects the area of ​​the connected domain. Considering that when the lung region is a standard circle, the ratio of the square of the perimeter to the area is 4π. In an embodiment of the present invention, in order to evaluate the shape factor and reflect the circular similarity of the connected domain, the preset ratio parameter is 4π; wherein, when the perimeter is constant, the larger the area corresponding to the circle, the minimum value of the ratio of the square of the perimeter to the area, that is, the maximum value of the first morphological coefficient, therefore, the larger the first morphological coefficient, the larger the shape factor, the closer the morphology of the connected domain is to the standard circle, and the more regular the corresponding liver region.

[0029] In one embodiment of the present invention, for any connected domain, the formula for obtaining the shape factor is expressed as: ; in, The shape factor representing the connected domain; Represents the number of all pixels in the connected domain, that is, the area of ​​the connected domain; Represents the number of edge pixels of the connected domain, that is, the perimeter of the connected domain; It represents the ratio of the number of all pixels in the connected domain to the number of squares of edge pixels, that is, the first morphological coefficient.

[0030] If the shape factor of the connected domain is greater than the preset factor threshold, the initial window of the corresponding connected domain is set to the preset first window. Otherwise, the initial window of the corresponding connected domain is set to the preset second window. The preset first window is larger than the preset second window. The initial window is centered on the centroid of the connected domain to construct a square area of ​​equal width.

[0031] It should be noted that, in one embodiment of the present invention, the preset factor threshold is 0.5, and the preset size of the first window is 50. 50, the default size of the second window is 30 30; In other embodiments of the present invention, the preset factor threshold, the preset first window and the preset second window can be specifically set according to specific circumstances, and are not limited or elaborated herein.

[0032] CT images reflect the degree of attenuation of radiation intensity when radiation passes through the body. Since human organs or tissues are composed of multiple material components and different densities, the density of tissues in different parts of the human body is different, and the degree of linear absorption of rays is different, resulting in different radiation reactions. Therefore, by analyzing the density in each initial window in the image, the complexity of the density distribution of pixels in the local window is determined; based on the density distribution of pixels in each initial window on the CT image, the degree of density change of each initial window is obtained.

[0033] Preferably, in one embodiment of the present invention, the method for acquiring density change information includes: Get the HU value of each pixel on the CT image; The average value of the ratio between the HU values ​​of different pixel points and the preset liver HU value was obtained and normalized as the density change degree of each initial window.

[0034] In one embodiment of the present invention, the formula for obtaining the density variation degree is expressed as: ; in, Indicates The degree of density change of the initial window; Indicates The number of pixels in the initial window; Indicates The first HU value of each pixel; Indicates the preset liver HU value; Represents the normalization function.

[0035] In the formula for density change, Indicates the calculation The first The difference between the HU value of the first pixel and the preset liver HU value reflects the The degree to which the pixel belongs to the liver area; Indicates the calculation The larger the difference mean, the greater the difference between the HU value of all pixels in the initial window and the preset liver HU value, the less it belongs to the lung area, the more complex the information distribution, and the greater the density change.

[0036] It should be noted that, in the embodiments of the present invention, the HU value is used to describe the absorption of radiation by different tissues, reflecting their density performance on the CT image. The method for obtaining the HU value is to process the CT image in a specific format through professional medical image processing software, and then convert the pixel value in the CT image into the HU value by using a window adjustment method. The specific means are technical means well known to those skilled in the art and will not be elaborated here.

[0037] It should be noted that, in order to evaluate the correlation between the initial window and the liver area, in one embodiment of the present invention, the preset liver HU value is pre-acquired by the implementer based on relevant professional data, that is, for the standard liver HU value range between 45-65, the average is calculated as the corresponding preset liver HU value.

[0038] Step S3: Obtain a multi-scale image corresponding to the CT image; obtain the degree of change of the high-frequency information of each initial window according to the frequency domain distribution characteristics corresponding to each initial window between images of different scales; obtain the window correction coefficient of each initial window according to the degree of change of density and the degree of change of high-frequency information of each initial window; correct each initial window according to the window correction coefficient to obtain a corrected window of each initial window.

[0039] Since liver CT images usually include multiple structures such as the liver itself, blood vessels, bile ducts, and surrounding soft tissues, and the contrast of different regions or structures is different, single-scale image processing will ignore more detailed features. Therefore, a multi-scale image corresponding to the CT image is obtained; it should be noted that, in one embodiment of the present invention, the method for obtaining a multi-scale image includes: using a Gaussian pyramid to obtain a multi-scale image of the CT image.

[0040] Among them, image pyramid is a technology that forms multiple resolution levels by gradually downsampling the image and applying Gaussian blur. By constructing multiple scales at different resolutions, the information characteristics of CT images at different scale dimensions can be captured.

[0041] It should be noted that, in one embodiment of the present invention, in order to make images consistent in a certain standard or feature for subsequent processing or analysis, the images at different scales are normalized using an interpolation method so that the multi-scale sub-images have a uniform size. The specific image pyramid and interpolation method are technical means well known to those skilled in the art and will not be elaborated here.

[0042] Frequency domain distribution characteristics can reflect the energy distribution and relative importance of different frequency components in the image. In liver CT images, high-frequency information corresponds to the details in the image, such as the edge information between liver tissue and other structures such as blood vessels and bile ducts. Due to the change in scale, the high-frequency components of the image will gradually be blurred, so that the low-frequency components dominate. By analyzing the frequency distribution characteristics at different scales, we can better understand and process the image data in the initial window, and then evaluate the change of high-frequency information. According to the frequency domain distribution characteristics of each initial window at the same position between images of different scales, the degree of change of high-frequency information in each initial window is obtained.

[0043] Preferably, in one embodiment of the present invention, the method for obtaining the degree of change of high-frequency information can be found in Figure 2 , which shows a flow chart of a method for obtaining the degree of change of high-frequency information, including: Step S301: For any scale image, obtain a frequency domain image of each initial window.

[0044] Frequency domain images make it easier to identify and analyze high-frequency and low-frequency components in an image.

[0045] It should be noted that, in the embodiment of the present invention, fast Fourier transform is used to convert each scale image from the spatial domain to the frequency domain. The specific fast Fourier transform is a technical means well known to those skilled in the art and will not be described in detail here.

[0046] Step S302: obtaining the sum of the amplitudes of all high-frequency information in each frequency domain image as the first amplitude; obtaining the sum of all frequency domain amplitudes in each frequency domain image as the second amplitude; obtaining the ratio of the first amplitude to the second amplitude as the high-frequency ratio.

[0047] High-frequency information reflects the parts of the frequency domain image that contain dramatically changing edges, details, and noise. The first amplitude reflects the content of high-frequency components in the initial window in the frequency domain, and the second amplitude reflects the total energy or total information of the initial window in the frequency domain. The high-frequency ratio is an indicator to measure the relative importance of high-frequency components in the window, reflecting the proportion of high-frequency information such as edges, details, and noise in the image relative to the overall information. The larger the sum of the amplitudes of the high-frequency information, the richer the edges and details in the image, or the higher the noise level, and the more targeted analysis is needed.

[0048] Preferably, in one embodiment of the present invention, the method for acquiring high-frequency information includes: For any frequency domain image, the relative distance between the pixel point with gray value not equal to 0 and the center point of the frequency domain image is obtained, and a distance histogram corresponding to the relative distance of all the pixel points with gray value not equal to 0 is constructed. The distance histogram is segmented using the Otsu method to obtain the segmented area of ​​the distance histogram. The largest relative distance mean in the segmented area is selected, and the corresponding pixel point is used as the high-frequency information in the spectrum image.

[0049] It should be noted that in some implementation methods of the present invention, the relative distance can be obtained by adopting distance measurement methods such as Euclidean distance or Manhattan distance. The specific means are technical means well known to those skilled in the art and will not be described in detail here.

[0050] Step S303: According to the difference in high-frequency ratios corresponding to each initial window at the same position between different adjacent scale images, the degree of change of high-frequency information of each initial window is obtained, and the difference in high-frequency ratios is positively correlated with the degree of change of high-frequency information.

[0051] The difference in high-frequency ratios reflects the change in the high-frequency information content of each initial window in images of different adjacent scales. The greater the difference in high-frequency ratios, the more different the characteristics of the high-frequency information between adjacent scale images, and the greater the degree of change in the high-frequency information, indicating that there is more edge information of liver tissue or other structures in the initial window.

[0052] In one embodiment of the present invention, the formula for obtaining the degree of change of high-frequency information is expressed as: ; in, Indicates The degree of change of high-frequency information in the initial window; Indicates On the scale image The sum of the amplitudes of the high-frequency information in the frequency domain image corresponding to the initial windows; Indicates On the scale image The sum of all frequency domain amplitudes in the frequency domain image corresponding to the initial window; Indicates On the scale image The sum of the amplitudes of the high-frequency information in the frequency domain image corresponding to the initial windows; Indicates On the scale image The sum of all frequency domain amplitudes in the frequency domain image corresponding to the initial window; m represents the number of scale images; Represents the normalization function.

[0053] In the formula for the degree of change of high-frequency information, Indicates the calculation On the scale image The sum of the amplitudes of the high-frequency information in the frequency domain image corresponding to the first initial window and the ratio of the sum of all frequency domain amplitudes, that is, On the scale image The high-frequency ratio corresponding to the initial window. The larger the high-frequency ratio, the larger the sum of the amplitudes of the high-frequency information, and the more high-frequency information is contained in the initial window. Indicates the calculation scale image and the The scale image The difference in high-frequency ratios corresponding to the initial windows is greater, and the more dissimilar the high-frequency ratios are, and the greater the change in the high-frequency information represented.

[0054] The degree of density change reflects the complexity of the detail information in the initial window. The greater the density change, the more complex the detail information in the initial window, and the more focus enhancement is needed. The degree of high-frequency information change reflects the degree of information change in the initial window. The higher the degree of information change, the more important the information performance is, and the more window adjustment is needed. The window correction coefficient of each initial window is obtained according to the degree of density change and the degree of high-frequency information change of each initial window.

[0055] Preferably, in one embodiment of the present invention, the method for obtaining the window correction coefficient includes: The density change degree and high-frequency information change degree of each initial window are fused to obtain the window correction coefficient corresponding to the initial window.

[0056] In some embodiments of the present invention, the density change degree and the high-frequency information change degree are fused by multiplication or addition. The specific means are well known to those skilled in the art and will not be described in detail here.

[0057] By obtaining the window correction coefficient, the content of detail information in the window can be reflected and the window size can be reselected; if the window size is set to a smaller size, although the local detail information can be better captured, the global information may be ignored, resulting in poor enhancement effect; and if the window size is set to a larger size, although the global information can be better considered, more noise and computational complexity may be introduced; each initial window is corrected according to the window correction coefficient to obtain the corrected window of each initial window.

[0058] Preferably, in one embodiment of the present invention, the method for obtaining the correction window includes: Obtain the difference between the positive integer 1 and the window correction coefficient as the adjustment weight; The product of the adjustment weight and the initial size of the initial window is obtained, and the sum of the product result and the initial size is calculated as the correction size to obtain the correction window.

[0059] In one embodiment of the present invention, the formula for correcting the size is expressed as: ; in, Indicates The correct size of the initial window; Indicates The initial size of the initial window; Indicates The window correction factor for the initial window.

[0060] In the formula for correcting the size, It represents the difference between the positive integer 1 and the window correction coefficient, that is, the adjustment weight. The larger the window correction coefficient, the more detail information contained in the window, the greater the possibility of containing the edge information of the liver tissue part, the smaller the adjustment of the initial size, and the easier it is to distinguish local details. The smaller the window correction coefficient, the less detail information contained in the window, the larger the adjustment of the initial size, and the more helpful it is to consider the global enhancement effect.

[0061] Step S4: Adaptively histogram equalize the images of different scales according to the correction window, and perform weighted fusion based on the chaos degree of the images of corresponding scales to obtain an enhanced CT image.

[0062] If the window in adaptive histogram equalization is too large or too small, the contrast of some areas of the image may be over-enhanced or weakened, and important information may be lost. By obtaining a correction window, areas with different detail features can be processed more flexibly. In the weighted fusion process, different weights are assigned according to the degree of confusion of the corresponding scale image, which can ensure that images with rich details and high contrast are more retained and emphasized in the fusion results. Therefore, adaptive histogram equalization is performed on images of different scales according to the correction window, and weighted fusion is performed in combination with the degree of confusion of the corresponding scale image to obtain an enhanced CT image.

[0063] It should be noted that, in one embodiment of the present invention, the degree of confusion is expressed by calculating information entropy, which can reflect the richness of detail information in the image. The greater the information entropy, the greater the degree of confusion and the more detail information; the smaller the information entropy, the smaller the degree of confusion and the less detail information; the specific means are technical means well known to those skilled in the art and will not be elaborated here.

[0064] In another embodiment of the present invention, a method for acquiring an enhanced CT image includes: obtaining a ratio of the degree of confusion of each scale image to the sum of the degrees of confusion of all scale images as the fusion weight of each scale image; weighting the scale image after adaptive histogram equalization based on the fusion weight of each scale image, and fusing all scale images to obtain an enhanced CT image; wherein, the greater the information entropy of the scale image, the more detail information it contains, the more details need to be retained, and the greater the fusion weight, by comprehensively considering the amount of information and detail information of images of different scales, thereby obtaining a clearer and more natural enhanced CT image, which better displays the detail information in the liver CT image.

[0065] It should be noted that the specific adaptive histogram equalization is a technical means well known to those skilled in the art and will not be described in detail here.

[0066] In summary, the present invention obtains the initial window of each connected domain according to the morphological characteristics of each connected domain in the CT image; obtains the density change degree of each initial window according to the density distribution of the pixels in each initial window on the CT image; obtains the multi-scale image corresponding to the CT image, and obtains the high-frequency information change degree of each initial window according to the frequency domain distribution characteristics of each initial window corresponding to the same position between different scale images; then obtains the window correction coefficient of each initial window, corrects each initial window, and obtains the correction window of each initial window; performs adaptive histogram equalization on different scale images according to the correction window, and performs weighted fusion in combination with the chaos degree of the corresponding scale image to obtain an enhanced CT image. The present invention improves the effect of impact enhancement by obtaining an accurate window size during image equalization processing.

[0067] The present invention also proposes a liver CT image enhancement system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any one of the steps of a liver CT image enhancement method.

[0068] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0069] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A liver CT image enhancement method, characterized in that: The method comprises: Obtain CT images of the liver; The connected domain of the CT image is obtained according to the grayscale distribution of the pixels on the CT image; the initial window of each connected domain is obtained according to the morphological characteristics of each connected domain; the density change degree of each initial window is obtained according to the density distribution of the pixels in each initial window on the CT image; Obtain a multi-scale image corresponding to the CT image; obtain the degree of change of high-frequency information of each initial window according to the frequency domain distribution characteristics of each initial window at the same position between the different-scale images; obtain a window correction coefficient of each initial window according to the degree of change of density and the degree of change of high-frequency information of each initial window; correct each initial window according to the window correction coefficient to obtain a corrected window of each initial window; Adaptive histogram equalization is performed on images of different scales according to the correction window, and weighted fusion is performed in combination with the degree of confusion of the images of corresponding scales to obtain an enhanced CT image.

2. A liver CT image enhancement method according to claim 1, characterized in that: The method for obtaining the connected domain includes: CANNY edge detection is performed on the CT image to obtain the edge curves in the CT image, and the independent areas surrounded by all closed edge curves are regarded as connected domains.

3. A liver CT image enhancement method according to claim 1, characterized in that: The method for obtaining the initial window includes: Obtaining the square of the number of edge pixels in each connected domain as the number of square edge pixels; obtaining the ratio of the number of all pixels in each connected domain to the number of square edge pixels as the first morphological coefficient; obtaining the product of a preset ratio parameter and the first morphological coefficient as a shape factor; If the shape factor of the connected domain is greater than the preset factor threshold, the initial window of the corresponding connected domain is set to the preset first window. Otherwise, the initial window of the corresponding connected domain is set to the preset second window. The preset second window is larger than the preset first window. The initial window is a square area of ​​equal width centered on the centroid of the connected domain.

4. The method for enhancing liver CT images according to claim 1, characterized in that: The method for obtaining the density variation degree includes: Get the HU value of each pixel on the CT image; The mean value of the difference between the HU values ​​of different pixel points and the preset liver HU value was obtained as the density change degree of each initial window.

5. The method for enhancing liver CT images according to claim 1, characterized in that: The method for obtaining the degree of change of the high-frequency information includes: For any scale image, obtain the frequency domain image of each initial window; The sum of the amplitudes of all high-frequency information in each frequency domain image is obtained as a first amplitude; the sum of the frequency domain amplitudes in each frequency domain image is obtained as a second amplitude; and the ratio of the first amplitude to the second amplitude is obtained as a high-frequency ratio; According to the difference in high-frequency ratios corresponding to each initial window at the same position between images of different adjacent scales, the degree of change of high-frequency information of each initial window is obtained, and the difference in high-frequency ratios is positively correlated with the degree of change of high-frequency information.

6. A liver CT image enhancement method according to claim 1, characterized in that: The method for obtaining the high-frequency information comprises: For any frequency domain image, the relative distance between the pixel point with gray value not equal to 0 and the center point of the frequency domain image is obtained, and a distance histogram corresponding to the relative distance of all the pixel points with gray value not equal to 0 is constructed. The distance histogram is segmented using the Otsu method to obtain the segmented area of ​​the distance histogram. The largest relative distance mean in the segmented area is selected, and the corresponding pixel point is used as the high-frequency information in the spectrum image.

7. The method for enhancing liver CT images according to claim 1, characterized in that: The method for obtaining the window correction coefficient includes: The density change degree and high-frequency information change degree of each initial window are fused to obtain the window correction coefficient corresponding to the initial window.

8. The method for enhancing liver CT images according to claim 1, characterized in that: The method for obtaining the correction window includes: Obtain the difference between the positive integer 1 and the window correction coefficient as the adjustment weight; The product of the adjustment weight and the initial size of the initial window is obtained, and the sum of the product result and the initial size is calculated as the revised size to obtain the revised window.

9. The method for enhancing liver CT images according to claim 1, characterized in that: The method for acquiring the multi-scale sub-image comprises: Gaussian pyramid is used to obtain multi-scale images corresponding to CT images.

10. A liver CT image enhancement system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the liver CT image enhancement method as described in any one of claims 1 to 9 are implemented.

Images