An intelligent analysis system for obstetric ultrasound examination data

Abstract

The present invention relates to the field of image processing technology, and in particular to an intelligent analysis system for obstetric ultrasound examination data, comprising: obtaining an obstetric ultrasound image; obtaining the discreteness of a pixel point according to a preset sliding window of the pixel point, and then obtaining a characteristic value of the pixel point; obtaining the probability that the pixel point is speckle noise according to the difference between the characteristic value of the pixel point and the characteristic values ​​of other pixel points in the neighborhood, and then obtaining an improved metric distance, clustering the obstetric ultrasound image according to the improved metric distance to obtain a number of clustering clusters; obtaining the adjacent clustering clusters of each pixel point, and then obtaining the size of the filter window of each pixel point, and obtaining a denoised image according to the size of the filter window of each pixel point. The present invention can effectively suppress noise and retain the edges between tissue regions by adaptively adjusting the filter window size of each pixel point.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to an intelligent analysis system for obstetric ultrasound examination data. Background Art

[0002] Obstetric ultrasound examination uses ultrasound technology to detect the uterus of pregnant women and the fetus therein to evaluate the growth and development of the fetus, including the size and position of the fetus. Compared with other medical imaging diagnostic methods, ultrasound diagnosis has the advantages of being non-destructive, painless, inexpensive, and convenient, making it one of the irreplaceable diagnostic methods in all clinical departments. However, due to the inherent characteristics of acoustic imaging itself, a type of speckled noise, called speckle noise, is widely present in ultrasound images. Therefore, when using existing target detection algorithms to detect the fetus in ultrasound images, the detection results are inaccurate due to the influence of speckle noise, and the results of the evaluation of the growth and development of the fetus are inaccurate.

[0003] Gaussian filtering is a commonly used image denoising method. One of the parameters that needs to be set is the filter window size. The size of the filter window determines how large a set of neighborhood pixels needs to be considered when filtering. Traditional Gaussian filtering is to smooth the only filter window set for the global image. However, due to the different degrees of contamination of speckle noise, if the filter window is set too large, more neighborhood pixel information will be considered to smooth the noise during processing, but at the same time, the image details will be blurred, affecting the clarity of the image. If the filter window is set too small, only a small number of neighborhood pixel information will be considered. At this time, some speckle noise with a smaller degree of contamination will be smoothed, and for speckle noise with a larger degree of contamination, only some noise points may be smoothed, leaving residual noise points unsmoothed. Summary of the invention

[0004] The present invention provides an intelligent analysis system for obstetric ultrasound examination data to solve the existing problems.

[0005] The intelligent analysis system for obstetric ultrasound examination data of the present invention adopts the following technical solution:

[0006] An embodiment of the present invention provides an intelligent analysis system for obstetric ultrasound examination data, the system comprising the following modules:

[0007] An obstetric ultrasound image acquisition module, for acquiring obstetric ultrasound images;

[0008] The characteristic value calculation module obtains the preset sliding window of each pixel point; obtains the discrete degree of each pixel point according to the preset sliding window of each pixel point, and obtains the characteristic value of each pixel point according to the gray value and discrete degree of each pixel point;

[0009] The speckle noise probability acquisition module obtains the probability that each pixel is speckle noise according to the difference between the characteristic value of each pixel and the characteristic values ​​of other pixels in the neighborhood;

[0010] The distance calculation module obtains the weights of the color distance and the spatial distance between the pixels according to the probability that each pixel is speckle noise; obtains the improved metric distance according to the probability that each pixel is speckle noise and the weights of the color distance and the spatial distance between the pixels, and clusters the obstetric ultrasound images according to the improved metric distance to obtain a number of clusters;

[0011] Adaptive filter window size module, obtains the adjacent clusters of each pixel; obtains the size of the filter window of each pixel;

[0012] The image denoising module obtains a denoised image according to the filter window size of each pixel point. When the denoised image is used for target detection, the detection result obtained is more accurate.

[0013] Preferably, the step of obtaining the discreteness of each pixel point according to a preset sliding window of each pixel point, and obtaining the characteristic value of each pixel point according to the gray value and discreteness of each pixel point comprises the following specific steps:

[0014]

[0015] in, Obstetric ultrasound images The gray value of each pixel; Represents obstetric ultrasound images The sliding window centered at the pixel point Line Gray value of the pixel in the column; Represents the first The mean grayscale value of all pixels in the sliding window centered at a pixel; Represents the first The maximum gray value in the sliding window centered at the pixel point; Represents the first The minimum gray value in the sliding window centered at pixels; is the side length of the preset sliding window; Represents the first The discrete degree of the gray value distribution of the pixel in the sliding window centered on the pixel; Represents the first The feature value of each pixel.

[0016] Preferably, the method of obtaining the probability that each pixel is speckle noise according to the difference between the characteristic value of each pixel and the characteristic values ​​of other pixels in the neighborhood includes the following specific steps:

[0017]

[0018] in, Obstetric ultrasound images The feature value of each pixel; Obstetric ultrasound images The first pixel in its eight neighborhoods The characteristic value of each pixel; represents an exponential function with a natural constant as base; Representative The probability that a pixel is speckle noise.

[0019] Preferably, the step of obtaining the weights of the color distance and the spatial distance between pixels according to the probability that each pixel is speckle noise includes the following specific steps:

[0020] Obstetric Ultrasound Images Pixels and The weight of the color distance between pixels is , the first Pixels and The weight of the spatial distance between pixels is ,in as well as Obstetric ultrasound images pixels and The probability that a pixel belongs to speckle noise.

[0021] Preferably, the improved metric distance is obtained according to the probability that each pixel is speckle noise and the weights of the color distance and the spatial distance between the pixels, and the specific steps include the following:

[0022]

[0023] in, as well as Obstetric ultrasound images The horizontal and vertical coordinates of the pixels; Obstetric ultrasound images The gray value of each pixel; as well as Obstetric ultrasound images The horizontal and vertical coordinates of the pixels; Obstetric ultrasound images The gray value of each pixel; as well as Obstetric ultrasound images pixels and The probability that a pixel belongs to speckle noise; Obstetric ultrasound images Pixels and The metric distance between pixels; represents an exponential function with a natural constant as base; is the weight of spatial distance; is the weight of the color distance.

[0024] Preferably, the step of obtaining a plurality of clusters includes the following specific steps:

[0025] Obstetric ultrasound images are clustered based on the improved metric distance combined with the DBSCAN algorithm to obtain several clusters.

[0026] Preferably, the step of obtaining the adjacent clusters of each pixel point includes the following specific steps:

[0027] Other clusters that have shared boundary points with the cluster to which each pixel belongs are obtained, and the other clusters are used as adjacent clusters of each pixel.

[0028] Preferably, the obtaining of the size of each pixel filter window comprises the following specific steps:

[0029] Get the adjacent clusters of each pixel, and use the minimum value of the distance between each pixel and the centroid of all its adjacent clusters as the size of the filter window of each pixel.

[0030] Preferably, the step of obtaining a denoised image comprises the following specific steps:

[0031] According to the filter window size of each pixel point, Gaussian filtering is used to denoise the obstetric ultrasound image to obtain a denoised image.

[0032] Preferably, the obtaining of the preset sliding window of each pixel point comprises the following specific steps:

[0033] The default sliding window side length is , taking each pixel as the center pixel, set Sliding window.

[0034] The beneficial effect of the technical solution of the present invention is as follows: the present invention combines the gray value distribution characteristics of speckle noise in obstetric ultrasound images to quantify the characteristic value of each pixel point, and then obtains the probability that each pixel point is speckle noise based on the difference between the characteristic value of each pixel point and the characteristic values ​​of other pixels in the neighborhood, and obtains the improved metric distance. The obstetric ultrasound image is clustered according to the improved metric distance, and the speckle noise can be accurately clustered into various clusters. Then, according to the distance of each pixel point to its adjacent cluster clusters, the size of the filter window of each pixel point is obtained; according to the distance of the pixel point to its adjacent cluster cluster, the filter window size of the adaptive pixel point is used to denoise the obstetric ultrasound image, which can effectively suppress noise and retain the edges between tissue regions. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be 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.

[0036] Figure 1 The system block diagram of the intelligent analysis system for obstetric ultrasound examination data of the present invention. DETAILED DESCRIPTION

[0037] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation, structure, features and effects of an intelligent analysis system for obstetric ultrasound examination data proposed by the present invention are described in detail below 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.

[0038] 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.

[0039] The specific scheme of the intelligent analysis system for obstetric ultrasound examination data provided by the present invention is described in detail below with reference to the accompanying drawings.

[0040] See also Figure 1 , which shows a system block diagram of an intelligent analysis system for obstetric ultrasound examination data provided by an embodiment of the present invention, the system includes the following modules:

[0041] The obstetric ultrasound image acquisition module 101 acquires obstetric ultrasound images.

[0042] It should be noted that the ultrasound probe is placed on the abdomen of the pregnant woman using B-ultrasound technology, and the probe is moved at an appropriate position and angle to obtain the desired image, which is recorded as an obstetric ultrasound image.

[0043] At this point, obstetric ultrasound images were acquired.

[0044] The eigenvalue calculation module 102 obtains the eigenvalue of each pixel in the obstetric ultrasound image.

[0045] It should be noted that due to the inherent characteristics of acoustic imaging itself, a type of speckled noise, called speckle noise, is widely present in ultrasound images. In order to avoid the influence of speckle noise in ultrasound images on subsequent image detection results, a Gaussian filtering algorithm is used to smooth the speckle noise in the image. Traditional Gaussian filtering sets a single filter window size for the entire image for smoothing. The size of the filter window affects the subsequent filtering effect. If the window is too small, the noise in the image usually cannot be completely removed. If the window is too large, the image will be over-smoothed and the image details will be lost. Therefore, this embodiment proposes a filter window size that is adaptive to each pixel point. For the pixels in the tissue area, the filter window needs to be as small as possible to ensure that the edge information between the tissues will not be smoothed out. For speckle noise, the filter window needs to be as large as possible The degree of pollution in the current spot area allows the speckle noise to be completely smoothed out.

[0046] It should be further explained that the overall grayscale value of the speckle noise area is large and its grayscale value distribution is uneven, while the grayscale value of the tissue area is evenly distributed. Therefore, the characteristic value of each pixel is quantified according to the grayscale value distribution characteristics of the speckle noise in the image.

[0047] In the embodiment of the present invention, it is set The sliding window traverses each pixel point in the obstetric ultrasound image. In the embodiment of the present invention, it is set In other embodiments, the implementer can set The value of .

[0048] According to the gray value and discreteness of each pixel, the characteristic value of each pixel is obtained:

[0049]

[0050] In the formula, Obstetric ultrasound images The gray value of each pixel; Represents obstetric ultrasound images The sliding window centered at the pixel point Line Gray value of the pixel in the column; Represents the first The mean grayscale value of all pixels in the sliding window corresponding to the pixel; Represents the first The maximum grayscale value in the sliding window corresponding to the pixel points; Represents the first The minimum gray value in the sliding window corresponding to the pixel point; is the side length of the sliding window; Represents the first The discrete degree of the gray value distribution of the pixel point in the sliding window corresponding to the pixel point is recorded as The degree of discreteness of pixels; Represents the first The characteristic value of the pixel point; When the gray value of a pixel is large and the discrete degree of the gray value distribution of the pixel in the corresponding sliding window is large, The larger the value of is, the more uneven the grayscale value distribution in the current sliding window is. The pixel may be in the speckle noise area or on the edge of the tissue area. When the gray value of a pixel is small and the discrete degree of the gray value distribution of the pixel in the corresponding sliding window is small, The smaller the value of is, the more uniform the grayscale value distribution in the current sliding window is. pixels are in the tissue area.

[0051] At this point, the characteristic value of each pixel in the image is obtained.

[0052] The speckle noise probability acquisition module 103 acquires the probability that each pixel is speckle noise according to the characteristic value of each pixel acquired by the module 102 and the spatial distribution characteristics of each pixel.

[0053] It should be noted that the above module 102 knows the characteristic value of each pixel calculated based on the grayscale characteristics of speckle noise. The larger the value, the more likely the pixel is to be a speckle noise pixel or a tissue area edge pixel. However, since the characteristic value of the edge point between tissues in the ultrasound image may also be large, it is necessary to distinguish between speckle noise pixels and tissue edge pixels. It is known that speckle noise is a patch randomly scattered on the image and the grayscale value distribution in the area is uneven. Therefore, the characteristic value of the pixel in the local area of ​​the speckle noise pixel is large, while the characteristic value of the tissue edge pixel in the local area is small. Therefore, the sum of the differences between the characteristic values ​​of each pixel and the characteristic values ​​of other pixels in the eight-neighborhood is quantified and then combined with its characteristic value to calculate the probability of each pixel being speckle noise.

[0054] In the embodiment of the present invention, the probability that each pixel is speckle noise is obtained according to the eigenvalue of each pixel and the sum of the differences between its eigenvalue and the eigenvalues ​​of other pixels in the eight neighborhoods:

[0055]

[0056] In the formula, Obstetric ultrasound images The characteristic value of each pixel; Obstetric ultrasound images The first pixel in its eight neighborhoods The characteristic value of each pixel; Obstetric ultrasound images The sum of the differences between the eigenvalue of a pixel and the eigenvalues ​​of other pixels in its eight neighborhoods; represents an exponential function with a natural constant as base; Represents the first The probability that a pixel is speckle noise is The larger the eigenvalue of a pixel is and the smaller the sum of the differences between its eigenvalue and the eigenvalues ​​of other pixels in the eight neighborhoods is, the more likely the pixel is to be speckle noise.

[0057] So far, the probability that all pixels are speckle noise is obtained. When the probability value is closer to 1, it means that the pixel is more likely to be speckle noise.

[0058] The distance calculation module 104 improves the distance metric according to the probability that each pixel point is speckle noise, and clusters the obstetric ultrasound images using the improved distance metric in combination with the DBSCAN clustering algorithm to obtain a number of clusters.

[0059] It should be noted that in order to make the filter window of the tissue area as small as possible to maintain the integrity of details and edges, the window of speckle noise is made larger than the degree of contamination of its spots to ensure that it can be completely smoothed out later. The DBSCAN clustering algorithm is used to cluster the speckle noise areas into clusters, and the tissue areas are clustered into smaller clusters, so as to facilitate the subsequent adaptive smaller filter windows according to the cluster areas. Since the DBSCAN algorithm sets a unified neighborhood radius for the entire image, and determines the density of each pixel by judging the number of pixels in the neighborhood radius of each pixel point, the distance between the pixel point in the tissue area and other pixel points needs to be as large as possible when calculating the distance, so that the smaller the number of pixels in the neighborhood radius of the pixel point in the area, the smaller the cluster after clustering. For the pixel point of speckle noise, the distance between the pixel point and other pixel points needs to be as small as possible when calculating the distance, so that the number of pixels in the neighborhood radius of the pixel point in the area is large, and the cluster after clustering is large. In the embodiment of the present invention, the probability value of each pixel point belonging to the noise point is used as the weight of the distance metric.

[0060] It should be further explained that the distance measurement between pixels in the DBSCAN algorithm takes into account the spatial distance and color distance between pixels. If the weights of spatial distance and color distance are treated equally, that is, the weights of the two are equal, it will affect the effect of subsequent clustering. It is known that in order to make tissue areas cluster into smaller clusters, the distance between pixels in the tissue area needs to be as large as possible when calculating the distance. The smaller the probability difference between two pixels being speckle noise, it means that the two pixels are in the same tissue area or the same speckle noise area. At this time, the spatial distance should be focused on. When the probability difference between two pixels being speckle noise is greater, it means that the two pixels belong to different areas. At this time, the color distance should be focused on.

[0061] In the implementation of the present invention, the first Pixels and The weight of the color distance between pixels is , the first Pixels and The weight of the spatial distance between pixels is ,in, as well as Obstetric ultrasound images pixels and The probability that a pixel belongs to speckle noise.

[0062] In the embodiment of the present invention, the improved metric distance is obtained according to the probability that each pixel is speckle noise and the weights of the color distance and the spatial distance between the pixels:

[0063]

[0064] In the formula, as well as Obstetric ultrasound images The horizontal and vertical coordinates of the pixels; Obstetric ultrasound images The gray value of each pixel; as well as Obstetric ultrasound images The horizontal and vertical coordinates of the pixels; Obstetric ultrasound images The gray value of each pixel; as well as Obstetric ultrasound images pixels and The probability that a pixel belongs to speckle noise; Obstetric ultrasound images Pixels and The metric distance between pixels is, for the pixels in the tissue area, The larger the is large, which makes the number of pixels in the neighborhood radius of the pixel point small. For the pixel point of speckle noise, The smaller, The smaller the value is, the more pixels there are in the neighborhood radius of the pixel.

[0065] At this point, the improved metric distance between each pixel point is obtained, and then the obstetric ultrasound image is clustered in combination with the DBSCAN algorithm to obtain several clusters.

[0066] The adaptive filter window size module 105 adaptively adjusts the filter window size of each pixel point according to the clustering result.

[0067] It should be noted that in order to smooth out the speckle noise, it is necessary to ensure that the filter window of the speckle noise includes pixels in other areas. Therefore, the filter window size of each pixel should be larger than the size of the cluster where the current pixel is located.

[0068] It should be noted that when there are shared boundary points between two clusters, the two clusters are judged to be adjacent at this time. Therefore, in an embodiment of the present invention, other clusters that have shared boundary points with the cluster to which each pixel belongs are obtained, and the other clusters are used as the adjacent clusters of each pixel. The distance between each pixel and the centroid of its adjacent cluster is calculated, and the minimum distance is selected as the size of the current pixel filter window.

[0069] At this point, the filter window size is obtained for each pixel in the obstetric ultrasound image.

[0070] The image denoising module 106 performs a denoising operation on the image using Gaussian filtering according to the filter window size of each pixel point obtained by the module S105 to obtain a denoised image.

[0071] In the embodiment of the present invention, according to the filter window size of each pixel point, the image is smoothed using Gaussian filtering to obtain a denoised image. The speckle noise of the obtained denoised image can be completely smoothed and the detail information in the image is retained. When the denoised image is used for target detection, the detection result obtained is more accurate.

[0072] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent analysis system for obstetric ultrasound examination data, characterized in that: The system includes the following modules: An obstetric ultrasound image acquisition module, for acquiring obstetric ultrasound images; The characteristic value calculation module obtains the preset sliding window of each pixel point; obtains the discrete degree of each pixel point according to the preset sliding window of each pixel point, and obtains the characteristic value of each pixel point according to the gray value and discrete degree of each pixel point; The speckle noise probability acquisition module obtains the probability that each pixel is speckle noise according to the difference between the characteristic value of each pixel and the characteristic values ​​of other pixels in the neighborhood; The distance calculation module obtains the weights of the color distance and the spatial distance between the pixels according to the probability that each pixel is speckle noise; obtains the improved metric distance according to the probability that each pixel is speckle noise and the weights of the color distance and the spatial distance between the pixels, and clusters the obstetric ultrasound images according to the improved metric distance to obtain a number of clusters; Adaptive filter window size module, obtains the adjacent clusters of each pixel; obtains the size of the filter window of each pixel; Image denoising module, which obtains denoised images according to the filter window size of each pixel point and uses the denoised images for target detection; The improved metric distance is obtained according to the probability that each pixel is speckle noise and the weights of the color distance and the spatial distance between the pixels, and the specific steps include the following: in, as well as Obstetric ultrasound images The horizontal and vertical coordinates of the pixels; Obstetric ultrasound images The gray value of each pixel; as well as Obstetric ultrasound images The horizontal and vertical coordinates of the pixels; Obstetric ultrasound images The gray value of each pixel; as well as Obstetric ultrasound images pixels and The probability that a pixel belongs to speckle noise; Obstetric ultrasound images Pixels and The metric distance between pixels; represents an exponential function with a natural constant as base; is the weight of spatial distance; is the weight of color distance; The specific steps of obtaining the adjacent clusters of each pixel point are as follows: Obtain other clusters that have shared boundary points with the cluster to which each pixel belongs, and use the other clusters as adjacent clusters of each pixel; The specific steps of obtaining the size of each pixel filter window are as follows: The minimum distance between each pixel and the centroid of all its adjacent clusters is used as the size of the filter window for each pixel.

2. The intelligent analysis system for obstetric ultrasound examination data according to claim 1, characterized in that: The specific steps of obtaining the discreteness of each pixel point according to the preset sliding window of each pixel point and obtaining the characteristic value of each pixel point according to the gray value and the discreteness of each pixel point are as follows: in, Obstetric ultrasound images The gray value of each pixel; Represents obstetric ultrasound images with The sliding window centered at the pixel point Line Gray value of the pixel in the column; Represents the first The mean grayscale value of all pixels in the sliding window centered at a pixel; Represents the first The maximum gray value in the sliding window centered at the pixel point; Represents the first The minimum gray value in the sliding window centered at pixels; is the side length of the preset sliding window; Represents the first The discrete degree of the gray value distribution of the pixel in the sliding window centered on the pixel; Represents the first The feature value of each pixel.

3. The intelligent analysis system for obstetric ultrasound examination data according to claim 1, characterized in that: The specific steps of obtaining the probability that each pixel is speckle noise according to the difference between the characteristic value of each pixel and the characteristic values ​​of other pixels in the neighborhood are as follows: in, Obstetric ultrasound images The characteristic value of each pixel; Obstetric ultrasound images The first pixel in its eight neighborhoods The characteristic value of each pixel; represents an exponential function with a natural constant as base; Representative The probability that a pixel is speckle noise.

4. The intelligent analysis system for obstetric ultrasound examination data according to claim 1, characterized in that: The specific steps of obtaining a plurality of clusters are as follows: Obstetric ultrasound images are clustered based on the improved metric distance combined with the DBSCAN algorithm to obtain several clusters.

5. The intelligent analysis system for obstetric ultrasound examination data according to claim 1, characterized in that: The specific steps of obtaining the denoised image are as follows: According to the filter window size of each pixel point, Gaussian filtering is used to denoise the obstetric ultrasound image to obtain a denoised image.

6. The intelligent analysis system for obstetric ultrasound examination data according to claim 1, characterized in that: The specific steps of obtaining the preset sliding window of each pixel point are as follows: The default sliding window side length is , taking each pixel as the center pixel, set Sliding window.

Images