Newborn umbilical inflammation infection risk early warning method combining umbilical secretions with ultrasound

Abstract

The present application relates to the technical field of health risk assessment, in particular to a new-born umbilical inflammation infection risk early warning method combining umbilical secretions and ultrasound, which comprises: obtaining a target color reaction image corresponding to umbilical secretions of a new-born to be detected, and obtaining a transverse image, a sagittal image and a coronal image corresponding to the umbilical part of the new-born to be detected; analyzing the distribution of leukocyte released protease based on the target color reaction image; determining a blood vessel wall stratification index corresponding to each preset direction according to the thickness difference of the media region and the adventitia region in each preset direction; performing infection spread analysis on the sagittal image and the coronal image; and determining an umbilical inflammation infection risk early warning auxiliary value corresponding to the new-born to be detected according to the esterase activity distribution index, the spread infiltration scale index and all blood vessel wall stratification indexes. The present application improves the rationality of setting the umbilical inflammation infection risk early warning auxiliary value by analyzing multiple images.

Description

Technical Field

[0001] This invention relates to the field of health risk assessment technology, specifically to a method for early warning of neonatal omphalitis infection risk by combining umbilical secretions with ultrasound. Background Technology

[0002] With the development of technology, the application of risk warning auxiliary values ​​to assist doctors in early warning of infection risks is becoming increasingly widespread. For example, it can be used to assist doctors in early warning of neonatal omphalitis infection risks. Currently, the method for obtaining risk warning auxiliary values ​​to assist doctors in early warning of infection risks is usually as follows: by identifying the inflammatory areas affected by infection from the image through differences in grayscale values, and generating risk warning auxiliary values ​​based on the inflammatory areas.

[0003] However, when identifying omphalitis infection areas from images based on differences in grayscale values ​​and generating risk warning auxiliary values ​​based on these areas, the following technical problems often arise:

[0004] Since the grayscale difference between the omphalitis infection area and the normal area is not particularly obvious, the accuracy of identifying the omphalitis infection area based on the different grayscale values ​​is often poor. This leads to a poor rationality in setting the auxiliary value for omphalitis infection risk warning, which in turn cannot effectively assist doctors in making judgments. Summary of the Invention

[0005] To address the technical problem of poorly set auxiliary values ​​for neonatal omphalitis infection risk warning, this invention proposes a method for neonatal omphalitis infection risk warning that combines umbilical secretions with ultrasound.

[0006] In a first aspect, the present invention provides a method for early warning of neonatal omphalitis infection risk by combining umbilical secretions with ultrasound, the method comprising:

[0007] Acquire the target colorimetric reaction image corresponding to the umbilical secretion of the newborn to be tested, and acquire the cross-sectional image, sagittal image and coronal image corresponding to the umbilicus of the newborn to be tested;

[0008] Based on the target colorimetric reaction image, the distribution of proteases released by leukocytes was analyzed to obtain the esterase activity distribution index;

[0009] The target vessel wall region is selected from the cross-sectional image, and the media and adventitia regions are selected from the target vessel wall region.

[0010] Based on the thickness difference between the media and adventitia regions in each preset direction, the vessel wall layering index corresponding to each preset direction is determined;

[0011] Based on the vessel wall layering index corresponding to all preset directions, infection spread analysis was performed on sagittal and coronal images to obtain spread and infiltration scale indicators.

[0012] Based on esterase activity distribution indicators, spread and infiltration scale indicators, and all vascular wall layering indices, the auxiliary warning value for the risk of omphalitis infection in the newborns to be tested was determined.

[0013] In conjunction with the first aspect above, in one possible implementation, the step of analyzing the distribution of leukocyte-released proteases based on the target colorimetric reaction image to obtain esterase activity distribution indicators includes:

[0014] The white blood cell region is selected from the target colorimetric reaction image, and the target colorimetric reaction image is divided into equal parts to obtain the target sub-region;

[0015] The average gray value of all pixels in each white blood cell region is used to determine the white blood cell protease activity index for each white blood cell region.

[0016] The variance of the leukocyte protease activity index corresponding to all leukocyte regions within each target sub-region is determined as the esterase activity fluctuation factor for each target sub-region.

[0017] The information entropy of the esterase activity fluctuation factor corresponding to all target sub-regions is determined as the esterase activity distribution index.

[0018] In conjunction with the first aspect above, in one possible implementation, determining the vessel wall layering index corresponding to each preset direction based on the thickness difference between the median and adventitia regions in each preset direction includes:

[0019] Choose any preset direction as the marking direction, and draw a target ray with the center of the target blood vessel wall region as the endpoint and the marking direction as the extension direction;

[0020] The number of pixels at the intersection of the target ray and the middle film region is determined as the representative thickness of the middle film corresponding to the marking direction;

[0021] The number of pixels at the intersection of the target ray and the outer film region is determined as the representative thickness of the outer film corresponding to the marking direction;

[0022] The ratio of the representative thickness of the adventitia to the representative thickness of the media corresponding to the marked direction is determined as the vessel wall layering index corresponding to the marked direction.

[0023] In conjunction with the first aspect above, in one possible implementation, the step of performing infection spread analysis on sagittal and coronal images based on the vessel wall layering index corresponding to all preset directions to obtain a spread and infiltration scale index includes:

[0024] Based on the vessel wall layering index corresponding to all preset directions, the target vessel wall region is divided to obtain the target local region;

[0025] Infection spread analysis is performed on the region corresponding to each target local area in the sagittal image to determine the infection spread factor of each target local area in the sagittal image.

[0026] Similarly, infection spread analysis is performed based on the region in the coronal image corresponding to each target local region to determine the infection spread factor of each target local region in the coronal image;

[0027] Based on the infection spread factors of all target local areas in sagittal and coronal images, the scale index of spread and infiltration was determined.

[0028] In conjunction with the first aspect above, in one possible implementation, the step of dividing the target blood vessel wall region according to the blood vessel wall layering index corresponding to all preset directions to obtain the target local region includes:

[0029] Based on the vessel wall layering index corresponding to all preset directions, clusters are performed on all preset directions to obtain target clusters;

[0030] The consecutive preset directions in each target cluster are formed into a preset direction sequence, and the preset direction located at the endpoint of each preset direction sequence is determined as a candidate direction;

[0031] With the center of the target blood vessel wall region as the endpoint and each candidate direction as the extension direction, a reference ray is drawn for each candidate direction;

[0032] The target blood vessel wall region is segmented using each reference ray as a dividing line to obtain the target local region.

[0033] In conjunction with the first aspect above, in one possible implementation, the step of performing infection spread analysis based on the region corresponding to each target local region in the sagittal image to determine the infection spread factor of each target local region in the sagittal image includes:

[0034] The region in the sagittal image corresponding to the target blood vessel wall region is determined as the reference blood vessel wall region.

[0035] Any target local region is defined as a marked local region, and the region in the sagittal image corresponding to the marked local region is defined as a reference local region;

[0036] The intersection of the boundary of the reference blood vessel wall region and the reference local region is determined as the boundary segment to be spread;

[0037] From the sagittal image, select the propagation reference point corresponding to each pixel in the boundary segment to be propagated;

[0038] Curve fitting is performed on the propagation reference points corresponding to all pixels in the boundary segment to be propagated to obtain the target fitting curve;

[0039] The region between the boundary segment to be spread and the target fitted curve is defined as the possible infection spread zone corresponding to the marked local region;

[0040] The potential spread zone of the infection is divided to obtain the local analysis zone sequence corresponding to the marked local region;

[0041] The infection spread factor of the marked local region in the sagittal image is determined based on the grayscale difference between adjacent local analysis bands in the local analysis band sequence, the variance of the grayscale values ​​corresponding to all pixels in the infection potential spread band, and the mean of the vascular wall layering index corresponding to all preset directions of the marked local region.

[0042] In conjunction with the first aspect above, in one possible implementation, the step of selecting the propagation reference point corresponding to each pixel in the boundary segment to be propagated from the sagittal image includes:

[0043] Any pixel in the boundary segment to be spread is designated as a temporary pixel, and the area in the sagittal image other than the reference blood vessel wall area is designated as the potential area to be spread.

[0044] The intersection of the normal of the temporary pixel on the boundary segment to be spread and the possible region to be spread is determined as the target intersection line segment;

[0045] Pixels that are at a preset distance from the temporary pixel are selected from the target intersection line segment and used as the propagation reference points corresponding to the temporary pixel.

[0046] In conjunction with the first aspect above, in one possible implementation, the step of dividing the potential infection spread zone to obtain the local analysis band sequence corresponding to the marked local region includes:

[0047] Divide the line connecting each pixel in the boundary segment to be spread and its corresponding spread reference point into equal parts to obtain the sequence of equally divided points corresponding to each pixel in the boundary segment to be spread.

[0048] Curve fitting is performed on the equally divided points with the same index in the equally divided point sequence corresponding to all pixels in the boundary segment to be spread to obtain candidate curves;

[0049] Using each candidate curve as a dividing line, the potentially spreading infection zone is segmented to obtain local analysis zones;

[0050] Based on the minimum distance between the local analysis band and the reference vessel wall region, all local analysis bands are sorted in ascending order to obtain a local analysis band sequence.

[0051] In conjunction with the first aspect above, in one possible implementation, determining the spread and infiltration scale index based on the infection spread factors of all target local regions in sagittal and coronal images includes:

[0052] The sum of the infection spread factors of all target local areas in the sagittal plane image is determined as the sagittal spread factor;

[0053] The cumulative value of the infection spread factor of all target local areas under coronal images is determined as the coronal spread factor;

[0054] The product of the sagittal spread factor and the coronal spread factor is determined as the spread and infiltration scale index.

[0055] In conjunction with the first aspect above, in one possible implementation, determining the auxiliary value for early warning of omphalitis infection risk in the newborn to be tested based on esterase activity distribution indicators, spread and infiltration scale indicators, and all vascular wall layering indices includes:

[0056] The mean of all vessel wall stratification indices is determined as the representative index of vessel wall stratification.

[0057] The cumulative product of the esterase activity distribution index, the spread and infiltration scale index, and the vascular wall layering representative index was normalized to obtain an auxiliary value for early warning of omphalitis infection risk.

[0058] Secondly, the present invention provides a neonatal omphalitis infection risk early warning system combining umbilical secretions and ultrasound, the system comprising:

[0059] Multiple image acquisition modules are used to acquire the target colorimetric reaction image corresponding to the umbilical secretion of the newborn to be tested, and to acquire the cross-sectional image, sagittal image and coronal image corresponding to the umbilicus of the newborn to be tested.

[0060] The data analysis module is used to analyze the distribution of proteases released by leukocytes based on the target colorimetric reaction image, and to obtain the esterase activity distribution index.

[0061] The region filtering module is used to filter out the target vessel wall region from the cross-sectional image, and then filter out the media region and adventitia region from the target vessel wall region.

[0062] The vessel wall layering index determination module is used to determine the vessel wall layering index corresponding to each preset direction based on the thickness difference between the media and adventitia regions in each preset direction.

[0063] The infection spread analysis module is used to perform infection spread analysis on sagittal and coronal images based on the vessel wall layering index corresponding to all preset directions, and obtain the spread and infiltration scale index.

[0064] The module for determining the auxiliary value for early warning of omphalitis infection risk is used to determine the auxiliary value for early warning of omphalitis infection risk for the newborn to be tested based on the esterase activity distribution index, the spread and infiltration scale index, and all vascular wall layering indices.

[0065] Thirdly, a server is provided, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, causing the device to perform the methods of the first aspect or any possible implementation thereof.

[0066] Fourthly, a computer program product is provided, comprising: computer program code, which, when run on a computer, causes the computer to perform the methods described in the first aspect or any possible implementation thereof.

[0067] Fifthly, a computer-readable storage medium is provided that stores computer program code, which, when executed on a computer, causes the computer to perform the methods described in the first aspect or any possible implementation thereof.

[0068] The present invention has the following beneficial effects:

[0069] This invention provides a method for early warning of neonatal omphalitis infection risk by combining umbilical secretions with ultrasound. By analyzing multiple images, including target colorimetric reaction images, cross-sectional images, sagittal images, and coronal images, this method addresses the technical problem of poor rationality in setting auxiliary values ​​for omphalitis infection risk warning and improves their rationality. Specifically, this invention quantifies several indicators related to neonatal omphalitis infection risk by analyzing these images, such as esterase activity distribution indicators, vascular wall layering index, and infiltration scale indicators. This quantifies the auxiliary values ​​for omphalitis infection risk warning and improves their rationality. Attached Figure Description

[0070] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0071] Figure 1 This is a flowchart of a neonatal omphalitis infection risk warning method combining umbilical secretions and ultrasound according to the present invention.

[0072] Figure 2 This is a schematic diagram of the composition and structure of a neonatal omphalitis infection risk warning system that combines umbilical secretions and ultrasound according to the present invention.

[0073] Figure 3 This is a schematic diagram of the structure of a computer device according to the present invention. Detailed Implementation

[0074] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the specific implementation methods, structures, features, and effects of the technical solution proposed according to the present invention are described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0075] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0076] refer to Figure 1 This document illustrates the flowchart of some embodiments of a neonatal omphalitis infection risk warning method combining umbilical secretions and ultrasound according to the present invention. The neonatal omphalitis infection risk warning method combining umbilical secretions and ultrasound includes the following steps:

[0077] Step S1: Obtain the target colorimetric reaction image corresponding to the umbilical secretion of the newborn to be tested, and obtain the cross-sectional image, sagittal image and coronal image corresponding to the umbilicus of the newborn to be tested.

[0078] Neonatal omphalitis is a common local bacterial infection in newborns, often occurring before or after the umbilical cord stump falls off. Newborns to be tested may be those undergoing omphalitis infection risk assessment. Umbilical secretions can be collected by rotating a sterile swab with a depth limiter inside the newborn's umbilical fossa for 10 seconds; the secretions collected during this process are considered umbilical secretions.

[0079] In practice, the enzyme activity level can be assessed through a leukocyte esterase colorimetric reaction; the brighter the cell color, the greater the degree of inflammation. The target colorimetric reaction image can be a fluorescence image corresponding to the leukocyte esterase colorimetric reaction of umbilical secretions, which can be obtained using ELISA (Enzyme-Linked Immunosorbent Assay) fluorescence method.

[0080] The transverse image of the umbilicus of the newborn being examined can be a cross-sectional image obtained perpendicular to the umbilical vessels. The sagittal image of the umbilicus of the newborn being examined can be an image obtained by longitudinal scanning along the umbilicus to the round ligament of the liver. The coronal image of the umbilicus of the newborn being examined can be an image covering the full-thickness structure of the abdominal wall. In practice, a 15MHz high-frequency linear array probe (axial resolution 0.1mm, penetration depth 4cm) can be used to perform standardized three-section scanning to obtain sagittal, transverse, and coronal images.

[0081] It should be noted that, in this embodiment of the invention, the acquisition times of the umbilical secretions, cross-sectional images, sagittal images, and coronal images can all be the same. The image sizes of the cross-sectional images, sagittal images, and coronal images can be the same.

[0082] Step S2: Based on the target colorimetric reaction image, analyze the distribution of proteases released by leukocytes to obtain esterase activity distribution indicators.

[0083] As an example, this step may include the following steps:

[0084] The first step is to screen out the white blood cell region from the above target colorimetric reaction image and divide the above target colorimetric reaction image into equal parts to obtain the target sub-region.

[0085] The white blood cell region can represent white blood cells. The target sub-region can be an equally divided rectangular region, and its size can be preset, for example, its size can be 1mm×1mm.

[0086] It should be noted that omphalitis is essentially caused by bacterial infection. When bacteria enter the umbilicus, they often trigger an immune system response, leading to inflammation. During the inflammatory process, the umbilical secretions often contain a large number of white blood cells and release proteases such as esterases to fight the infecting bacteria.

[0087] The grayscale representation of leukocytes and the proteases they release, such as esterases, in colorimetric reaction images often differs from that of other regions. Therefore, thresholding can be used to divide the target colorimetric reaction image into two parts, and the connected regions in the part with the larger grayscale value are designated as leukocyte regions.

[0088] Alternatively, semantic segmentation or neural network techniques can be used to screen out white blood cell regions from the target colorimetric reaction image.

[0089] The second step is to determine the average gray value of all pixels within each white blood cell region as the white blood cell protease activity index for each white blood cell region.

[0090] It should be noted that in fluorescence imaging, a higher gray value generally indicates stronger leukocyte protease activity. Therefore, a higher gray value in the leukocyte region often indicates relatively stronger leukocyte protease activity.

[0091] The third step is to determine the variance of the leukocyte protease activity index corresponding to all leukocyte regions within each target sub-region as the esterase activity fluctuation factor for each target sub-region.

[0092] It should be noted that the larger the esterase activity fluctuation factor corresponding to the target sub-region, the greater the fluctuation of leukocyte protease activity in that target sub-region.

[0093] The fourth step is to determine the information entropy of the esterase activity fluctuation factor corresponding to all target sub-regions as the esterase activity distribution index.

[0094] It should be noted that when inflammation spreads, leukocyte proteases released from umbilical secretions often exhibit random infiltration. When cellular proteases within the local space show disordered distribution, it often indicates that leukocytes have breached the tissue barrier, exacerbating the impact of omphalitis infection. Esterase activity distribution indicators can be used to reflect the disordered distribution of leukocyte esterase activity in images. A higher value often means that leukocyte esterase activity is more likely to be disordered and diffuse. In this case, disordered release of leukocyte proteases may lead to immune system dysfunction and a worsening of the inflammation.

[0095] Step S3: Select the target vessel wall region from the cross-sectional image, and then select the media region and adventitia region from the target vessel wall region.

[0096] The target vessel wall region can be the area representing the umbilical vein wall in a cross-sectional image. In reality, in ultrasound images, the normal umbilical vein wall often presents a homogeneous three-layer structure, which, from the inside out, can be: a hypoechoic intima, a hyperechoic media, and a mesoechoic adventitia. The media and adventitia regions selected from the target vessel wall region can respectively represent the media and adventitia of the umbilical vein wall.

[0097] As an example, this step may include the following steps:

[0098] The first step is to select the target vessel wall region from the cross-sectional images.

[0099] For example, the Hough circle detection algorithm can be used to search for circular-shaped luminal structures within a 5mm radius of the umbilical ring center, thus forming the target vessel wall region. The umbilical ring center can often be characterized by the center point of a cross-sectional image.

[0100] The second step is to select the media and adventitia regions from the target vessel wall region.

[0101] For example, three initial cluster centers can be selected in the target blood vessel wall region using the elbow method. K-means clustering based on grayscale differences can then be performed to obtain three clusters. Each cluster can represent a membrane structure and constitute a membrane region. The cluster with the highest grayscale value among these three clusters can be identified as the middle membrane region; the cluster with the lowest grayscale value among these three clusters can be identified as the intima region; and the cluster with the middle grayscale value among these three clusters can be identified as the adventitia region.

[0102] Alternatively, neural network technology can be used to identify the target vessel wall region, media region, and adventitia region.

[0103] Step S4: Determine the vessel wall layering index corresponding to each preset direction based on the thickness difference between the media and adventitia regions in each preset direction.

[0104] The preset direction can be a pre-set direction, and the number of preset directions can also be pre-set. The angle between adjacent preset directions can be 1°. For example, the range of preset directions can be {0° direction, 1° direction, 2° direction, ..., 358° direction, 359° direction}.

[0105] As an example, this step may include the following steps:

[0106] The first step is to determine any preset direction as the marking direction, and draw a target ray with the center of the target blood vessel wall region as the endpoint and the marked direction as the extension direction.

[0107] The second step is to determine the number of pixels in the intersection of the target ray and the intermediate film region as the representative thickness of the intermediate film corresponding to the marked direction.

[0108] The third step is to determine the number of pixels in the intersection of the target ray and the outer membrane region as the representative thickness of the outer membrane corresponding to the marked direction.

[0109] The fourth step is to determine the ratio of the thickness of the adventitia to the thickness of the media corresponding to the marked direction as the vessel wall layering index corresponding to the marked direction.

[0110] It should be noted that when inflammatory bacteria infiltrate blood vessels, the proteases secreted by the bacteria often dissolve the elastic fibers of the medial layer, potentially leading to the disintegration and thinning of the medial structure. This results in a decrease in the thickness of the medial layer, which appears as hyperechoic on ultrasound. Simultaneously, inflammation of the adventitia layer, such as edema, may exacerbate the condition, potentially leading to an increase in the thickness of the adventitia layer, which appears as hypoechoic on ultrasound. Therefore, the vessel wall layering index corresponding to the marked direction can characterize the degree of damage to the umbilical vein membrane in that direction. A higher value generally indicates more severe damage to the umbilical vein membrane structure in that direction, and a correspondingly increased risk of tissue fluid exudation and infection spread to the abdominal cavity.

[0111] Step S5: Based on the vessel wall layering index corresponding to all preset directions, perform infection spread analysis on sagittal and coronal images to obtain spread infiltration scale indicators.

[0112] As an example, this step may include the following steps:

[0113] The first step, based on the vessel wall layering index corresponding to all preset directions, divides the target vessel wall region to obtain the target local region, which may include the following sub-steps:

[0114] The first sub-step involves clustering all preset directions based on the vessel wall layering index corresponding to all preset directions to obtain the target cluster.

[0115] For example, if the absolute value of the difference between the vascular wall layering indices corresponding to adjacent preset directions is less than or equal to 0.3, then the adjacent preset directions are divided into the same cluster, and each cluster after clustering is recorded as the target cluster.

[0116] For example, if the absolute value of the difference between the vascular wall layering indices corresponding to the 2° and 3° directions is greater than 0.3, the absolute value of the difference between the vascular wall layering indices corresponding to the 3° and 4° directions is less than or equal to 0.3, the absolute value of the difference between the vascular wall layering indices corresponding to the 4° and 5° directions is less than or equal to 0.3, the absolute value of the difference between the vascular wall layering indices corresponding to the 5° and 6° directions is less than or equal to 0.3, and the absolute value of the difference between the vascular wall layering indices corresponding to the 6° and 7° directions is greater than 0.3, then the 3°, 4°, 5°, and 6° directions can be considered as a target cluster.

[0117] The second sub-step involves forming a sequence of preset directions from consecutive preset directions in each target cluster, and identifying the preset directions located at the endpoints of each preset direction sequence as candidate directions.

[0118] For example, if a target cluster includes the following directions: 3°, 4°, 5° and 6°, then the preset direction sequence can be {3°, 4°, 5° and 6°}. At this time, there are two preset directions at the endpoints, namely the 3° direction and the 6° direction, and the 3° direction and the 6° direction can be recorded as candidate directions.

[0119] For example, if a target cluster only includes the 18° direction, then the preset direction sequence formed at this time can be {18° direction}. At this time, there is one preset direction at the endpoint, which is the 18° direction, and the 18° direction can be recorded as a candidate direction.

[0120] The third sub-step involves drawing a reference ray corresponding to each candidate direction, with the center of the target blood vessel wall region as the endpoint and each candidate direction as the extension direction.

[0121] The fourth sub-step involves dividing the target blood vessel wall region using each reference ray as a dividing line, and then recording the region obtained at this point as the target local region.

[0122] The second step involves performing infection spread analysis based on the region corresponding to each target local area in the sagittal image. Determining the infection spread factor for each target local area in the sagittal image may include the following sub-steps:

[0123] The first sub-step involves identifying the region in the sagittal image corresponding to the target blood vessel wall region as the reference blood vessel wall region.

[0124] The reference vessel wall region can be the same region in the sagittal image as the target vessel wall region.

[0125] For example, for cross-sectional images, a three-dimensional Cartesian coordinate system can be established with the center of the umbilical ring as the origin. Key points of the blood vessel wall identified from the cross-sectional image can be used as reference markers. Then, the coordinates of the cross-sectional blood vessel structure can be mapped to the sagittal image through an affine transformation matrix to obtain the reference blood vessel wall region. Among them, the key points of the blood vessel wall can be the intima boundary points.

[0126] The second sub-step involves identifying any target local region as a marked local region and identifying the region in the sagittal image corresponding to the marked local region as a reference local region.

[0127] The reference local region can be a region in the sagittal image that is the same as the region represented by the marked local region.

[0128] For example, the region in the sagittal image corresponding to the marked local region can be obtained by using affine transformation, and denoted as the reference local region.

[0129] The third sub-step involves determining the intersection of the boundary of the aforementioned reference blood vessel wall region and the aforementioned reference local region as the boundary segment to be propagated.

[0130] The fourth sub-step, selecting the propagation reference point corresponding to each pixel in the aforementioned boundary segment to be propagated from the sagittal image, may include the following steps:

[0131] First, any pixel in the aforementioned boundary segment to be spread is designated as a temporary pixel, and the area in the sagittal image other than the aforementioned reference blood vessel wall area is designated as the potential area to be spread.

[0132] Next, the intersection of the normal of the temporary pixel on the boundary segment to be spread and the possible region to be spread is determined as the target intersection line segment.

[0133] It should be noted that the target intersection line segment can represent the extendable line segment of a temporary pixel.

[0134] Finally, pixels that are at a preset distance from the temporary pixels are selected from the intersection line segments of the above targets and used as the propagation reference points corresponding to the temporary pixels.

[0135] The preset distance can be a pre-set distance, which can be equal to the length of 15 pixels. The propagation reference point corresponding to the temporary pixel can represent the pixel obtained after extending the temporary pixel by the preset distance.

[0136] The fifth sub-step involves performing curve fitting on the propagation reference points corresponding to all pixels in the aforementioned boundary segment to be propagated, thereby obtaining the target fitting curve.

[0137] It should be noted that the target fitting curve can characterize the spread boundary obtained after the boundary segment to be spread has spread by a preset distance.

[0138] The sixth sub-step involves identifying the region between the aforementioned boundary segment to be spread and the aforementioned target fitted curve as the potential infection spread zone corresponding to the aforementioned marked local region.

[0139] For example, the endpoints on the same side of the target fitted curve can be connected to the boundary segment to be spread. The area formed in this way is denoted as the infection potential spread zone corresponding to the marked local area.

[0140] The seventh sub-step, dividing the aforementioned potential infection spread zones to obtain the local analysis band sequences corresponding to the marked local regions, may include the following steps:

[0141] First, the line connecting each pixel in the boundary segment to be spread and its corresponding spread reference point is divided equally to obtain the sequence of equally divided points corresponding to each pixel in the boundary segment to be spread.

[0142] It should be noted that the number of equal division points in the sequence of equal division points corresponding to different pixels in the boundary segment to be spread can be the same.

[0143] For example, the line connecting the pixel and its corresponding spread reference point can be divided into three equal parts, resulting in three equal division points. Based on the minimum distance between these three equal division points and the reference blood vessel wall region, these three equal division points are arranged in ascending order to obtain the sequence of equal division points corresponding to the pixel.

[0144] Next, curve fitting is performed on the equally divided points with the same index in the equally divided point sequence corresponding to all pixels in the above-mentioned boundary segment to be spread, to obtain candidate curves.

[0145] It should be noted that the candidate curve can represent the spread boundary obtained after the boundary segment to be spread has spread a certain distance.

[0146] Then, using each candidate curve as a dividing line, the possible spread zone of the infection is segmented to obtain the local analysis zone.

[0147] Finally, based on the minimum distance between the local analysis band and the aforementioned reference vessel wall region, all local analysis bands are sorted in ascending order to obtain the local analysis band sequence.

[0148] The eighth sub-step involves determining the infection spread factor of the marked local area in the sagittal image based on the grayscale difference between adjacent local analysis bands in the aforementioned local analysis band sequence, the variance of the grayscale values ​​corresponding to all pixels in the aforementioned infection potential spread band, and the mean value of the vascular wall layering index corresponding to all preset directions of the marked local area.

[0149] Specifically, all preset directions corresponding to the marked local region can be preset directions covered by the marked local region. For example, if the two dividing boundary lines of the marked local region are reference rays corresponding to the 16° and 18° directions, then all preset directions corresponding to the marked local region can be the 16°, 17°, and 18° directions, respectively.

[0150] For example, the formula for determining the infection spread factor of a marked local area in a sagittal image can be:

[0151] ;

[0152] ;

[0153] in, B It is a marker of infection spread in local areas under sagittal images. A It is the mean of the grayscale differences between all adjacent local analysis bands in the local analysis band sequence corresponding to the local region.N It represents the number of local analysis bands in the local analysis band sequence corresponding to the local region. i It is the sequence number of the local analysis band in the local analysis band sequence corresponding to the local region. It is an absolute value function. It is the first local analysis band sequence corresponding to the marked local region. i The mean of the gray values ​​corresponding to all pixels in a local analysis band. It is the first local analysis band sequence corresponding to the marked local region. i +1 The mean of the gray values ​​corresponding to all pixels in the local analysis band. It is the variance of the gray values ​​of all pixels in the possible spread zone of the infection corresponding to the marked local area. L It is the average of the vessel wall layering indices corresponding to all preset directions of the marked local area.

[0154] It should be noted that, L It can characterize the overall degree of damage to the umbilical vein membrane in a localized area. The larger the value, the more severe the damage to the umbilical vein membrane structure, and the greater the risk of tissue fluid exudation and infection spreading to the abdominal cavity. A This can be characterized by the rate of change of the gray-scale gradient of the potentially spreading infection zone corresponding to a localized area. A higher value often indicates more rapid changes in the blood vessel walls and surrounding tissues within the local area, corresponding to an intensified inflammatory response, accelerated spread of exudate, and a rapid expansion of the infection range, thus leading to a higher risk of infection spread. A larger value often indicates more severe local damage, especially pus or necrotic tissue, within the potential zone of infection spread. This could exacerbate the inflammatory response and expand the infection's reach, thus accelerating its spread. Therefore, B It can characterize the spread of infection in a localized area under sagittal imaging.

[0155] The third step, similarly, involves performing infection spread analysis based on the corresponding region in the coronal images for each target local area, to determine the infection spread factor for each target local area under the coronal images.

[0156] It should be noted that the method for obtaining the infection spread factor of the target local area in coronal images is the same as the method for obtaining the infection spread factor of the target local area in sagittal images, and will not be repeated here.

[0157] The fourth step, determining the scale index of spread and infiltration based on the infection spread factors of all target local areas in sagittal and coronal images, may include the following sub-steps:

[0158] The first sub-step is to determine the sagittal spread factor by summing the infection spread factors of all target local areas in the sagittal plane image.

[0159] The second sub-step involves summing the infection spread factors of all target local areas under coronal images and determining them as the coronal spread factor.

[0160] The third sub-step involves multiplying the sagittal spread factor and the coronal spread factor to determine the spread and infiltration scale index.

[0161] It should be noted that the larger the spread and infiltration scale index, the greater the risk of infection spread.

[0162] Step S6: Determine the auxiliary value for early warning of omphalitis infection risk for the newborn to be tested based on the esterase activity distribution index, the spread and infiltration scale index, and all vascular wall layering indices.

[0163] As an example, this step may include the following steps:

[0164] The first step is to determine the mean of all vascular wall stratification indices as the representative index of vascular wall stratification.

[0165] It should be noted that the vessel wall layering index can characterize the overall degree of damage to the umbilical vein membrane. The higher the value, the more severe the damage to the umbilical vein membrane structure, and the greater the risk of tissue fluid leakage and infection spreading to the abdominal cavity.

[0166] The second step is to normalize the cumulative value among the esterase activity distribution index, the spread and infiltration scale index, and the vascular wall layering representative index to obtain the auxiliary value for early warning of omphalitis infection risk.

[0167] It should be noted that the esterase activity distribution index can reflect the disordered distribution of esterase activity in leukocytes in images. A higher value often indicates a more disordered and diffuse state of leukocyte esterase activity. In this case, disordered release of leukocyte proteases may lead to immune system dysfunction and a worsening of inflammation. A larger spread and infiltration scale index generally indicates a greater risk of infection spread. The vessel wall layering index can characterize the overall degree of damage to the umbilical vein membrane. Therefore, a higher auxiliary value for the omphalitis infection risk warning generally indicates a greater risk of neonatal omphalitis infection, suggesting higher infection activity and potential for worsening, thus requiring more proactive warnings.

[0168] Optionally, the auxiliary value for neonatal omphalitis infection risk warning can be used as a key spatiotemporal dynamic feature input to a pre-trained machine learning model, such as a gradient boosting decision tree, a temporal convolutional network, or a graph neural network. This model learns the complex nonlinear relationship between neonatal omphalitis infection risk through offline training and integrates the patient's basic physiological parameters (such as immune markers and clinical history) as auxiliary features to generate a neonatal omphalitis infection risk warning decision within the current time window.

[0169] refer to Figure 2 Based on the same inventive concept as the above-described method embodiments, this invention provides a neonatal omphalitis infection risk warning system combining umbilical secretions and ultrasound. The system includes 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 the steps of a neonatal omphalitis infection risk warning method combining umbilical secretions and ultrasound, specifically including:

[0170] The multi-image acquisition module 201 is used to acquire the target colorimetric reaction image corresponding to the umbilical secretion of the newborn to be tested, and to acquire the cross-sectional image, sagittal image and coronal image corresponding to the umbilical part of the newborn to be tested.

[0171] The data analysis module 202 is used to analyze the distribution of proteases released by leukocytes based on the target colorimetric reaction image, and to obtain the esterase activity distribution index.

[0172] The region filtering module 203 is used to filter out the target vessel wall region from the cross-sectional image, and to filter out the media region and adventitia region from the target vessel wall region.

[0173] The vessel wall layering index determination module 204 is used to determine the vessel wall layering index corresponding to each preset direction based on the thickness difference between the media region and the adventitia region in each preset direction.

[0174] The infection spread analysis module 205 is used to perform infection spread analysis on sagittal and coronal images based on the vessel wall layering index corresponding to all preset directions, and obtain the spread infiltration scale index.

[0175] The module 206 for determining the auxiliary value for early warning of omphalitis infection risk is used to determine the auxiliary value for early warning of omphalitis infection risk for the newborn to be tested based on the esterase activity distribution index, the spread and infiltration scale index, and all vascular wall layering indices.

[0176] Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. For example, as shown... Figure 3As shown, the computer device 300 includes: a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and running on the processor 302. When the processor 302 executes the computer program 303, the computer device can execute any of the aforementioned methods for early warning of neonatal omphalitis infection risk by combining umbilical secretions with ultrasound.

[0177] Based on the same inventive concept as the above-described method embodiments, the present invention provides a server, including a memory and a processor. The memory stores executable program code, and the processor retrieves and runs the executable program code from the memory, enabling the device to execute any of the above-described methods for early warning of neonatal omphalitis infection risk using a combination of umbilical secretions and ultrasound.

[0178] Based on the same inventive concept as the above-described method embodiments, the present invention provides a computer program product comprising: computer program code, which, when executed on a computer, causes the computer to execute any of the above-described methods for early warning of neonatal omphalitis infection risk by combining umbilical secretions with ultrasound.

[0179] Based on the same inventive concept as the above-described method embodiments, the present invention provides a computer-readable storage medium storing computer program code, which, when executed on a computer, causes the computer to perform any of the above-described methods for early warning of neonatal omphalitis infection risk by combining umbilical secretions and ultrasound.

[0180] In summary, this invention quantifies multiple indicators related to the risk of neonatal omphalitis infection by analyzing various images such as target colorimetric reaction images, cross-sectional images, sagittal images, and coronal images. These indicators include esterase activity distribution indicators, vascular wall layering index, and spread infiltration scale indicators. This enables the quantification of auxiliary values ​​for early warning of omphalitis infection risk and improves the rationality of setting auxiliary values ​​for early warning of omphalitis infection risk.

[0181] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for early warning of neonatal omphalitis infection risk by combining umbilical secretions and ultrasound, characterized in that, Includes the following steps: Acquire the target colorimetric reaction image corresponding to the umbilical secretion of the newborn to be tested, and acquire the cross-sectional image, sagittal image and coronal image corresponding to the umbilicus of the newborn to be tested; Based on the target colorimetric reaction image, the distribution of proteases released by leukocytes was analyzed to obtain the esterase activity distribution index; The process of identifying the target vessel wall region from the cross-sectional image, and then further identifying the media and adventitia regions from within the target vessel wall region, involves: selecting three initial cluster centers within the target vessel wall region using the elbow method; performing K-means clustering based on grayscale differences to obtain three clusters, each representing a membrane structure and constituting a membrane region; identifying the cluster with the highest grayscale value as the media region; identifying the cluster with the lowest grayscale value as the intima region; and identifying the cluster with the middle grayscale value as the adventitia region. Based on the thickness difference between the media and adventitia regions in each preset direction, the vessel wall layering index corresponding to each preset direction is determined, wherein the included angle between adjacent preset directions is 1°; the range of preset directions is {0° direction, 1° direction, 2° direction, ..., 358° direction, 359° direction}. Based on the vessel wall layering index corresponding to all preset directions, infection spread analysis was performed on sagittal and coronal images to obtain spread and infiltration scale indicators. Based on esterase activity distribution indicators, spread and infiltration scale indicators, and all vessel wall layering indices, the auxiliary warning value for the risk of omphalitis infection in the newborns to be tested was determined. The step of determining the vessel wall layering index corresponding to each preset direction based on the thickness difference between the medial and adventitia regions in each preset direction includes: determining any preset direction as a marking direction; drawing a target ray with the center of the target vessel wall region as the endpoint and the marking direction as the extension direction; determining the number of pixels in the intersection of the target ray and the medial region as the representative thickness of the medial region corresponding to the marking direction; determining the number of pixels in the intersection of the target ray and the adventitia region as the representative thickness of the adventitia region corresponding to the marking direction; and determining the ratio of the representative thickness of the adventitia to the representative thickness of the medial region corresponding to the marking direction as the vessel wall layering index corresponding to the marking direction.

2. The method for early warning of neonatal omphalitis infection risk by combining umbilical secretions and ultrasound according to claim 1, characterized in that, The distribution of proteases released by leukocytes is analyzed based on the target colorimetric reaction image to obtain esterase activity distribution indicators, including: The white blood cell region is selected from the target colorimetric reaction image, and the target colorimetric reaction image is divided into equal parts to obtain the target sub-region; The average gray value of all pixels in each white blood cell region is used to determine the white blood cell protease activity index for each white blood cell region. The variance of the leukocyte protease activity index corresponding to all leukocyte regions within each target sub-region is determined as the esterase activity fluctuation factor for each target sub-region. The information entropy of the esterase activity fluctuation factor corresponding to all target sub-regions is determined as the esterase activity distribution index.

3. The method for early warning of neonatal omphalitis infection risk by combining umbilical secretions and ultrasound according to claim 1, characterized in that, The infection spread analysis is performed on sagittal and coronal images based on the vessel wall layering index corresponding to all preset directions to obtain spread and infiltration scale indicators, including: Based on the vessel wall layering index corresponding to all preset directions, the target vessel wall region is divided to obtain the target local region; Infection spread analysis is performed on the region corresponding to each target local area in the sagittal image to determine the infection spread factor of each target local area in the sagittal image. Similarly, infection spread analysis is performed based on the region in the coronal image corresponding to each target local region to determine the infection spread factor of each target local region in the coronal image; Based on the infection spread factors of all target local areas in sagittal and coronal images, the scale index of spread and infiltration was determined.

4. The method for early warning of neonatal omphalitis infection risk by combining umbilical secretions and ultrasound according to claim 3, characterized in that, The step of dividing the target blood vessel wall region into target local regions based on the blood vessel wall layering index corresponding to all preset directions includes: Based on the vessel wall layering index corresponding to all preset directions, clusters are performed on all preset directions to obtain target clusters; The consecutive preset directions in each target cluster are formed into a preset direction sequence, and the preset direction located at the endpoint of each preset direction sequence is determined as a candidate direction; With the center of the target blood vessel wall region as the endpoint and each candidate direction as the extension direction, a reference ray is drawn for each candidate direction; The target blood vessel wall region is segmented using each reference ray as a dividing line to obtain the target local region.

5. The method for early warning of neonatal omphalitis infection risk by combining umbilical secretions and ultrasound according to claim 3, characterized in that, The step of performing infection spread analysis based on the region corresponding to each target local area in the sagittal image, and determining the infection spread factor of each target local area in the sagittal image, includes: The region in the sagittal image corresponding to the target blood vessel wall region is determined as the reference blood vessel wall region. Any target local region is defined as a marked local region, and the region in the sagittal image corresponding to the marked local region is defined as a reference local region; The intersection of the boundary of the reference blood vessel wall region and the reference local region is determined as the boundary segment to be spread; From the sagittal image, select the propagation reference point corresponding to each pixel in the boundary segment to be propagated; Curve fitting is performed on the propagation reference points corresponding to all pixels in the boundary segment to be propagated to obtain the target fitting curve; The region between the boundary segment to be spread and the target fitted curve is defined as the possible infection spread zone corresponding to the marked local region; The potential spread zone of the infection is divided to obtain the local analysis zone sequence corresponding to the marked local region; The infection spread factor of the marked local region in the sagittal image is determined based on the grayscale difference between adjacent local analysis bands in the local analysis band sequence, the variance of the grayscale values ​​corresponding to all pixels in the infection potential spread band, and the mean of the vascular wall layering index corresponding to all preset directions of the marked local region.

6. The method for early warning of neonatal omphalitis infection risk by combining umbilical secretions and ultrasound according to claim 5, characterized in that, The step of selecting the propagation reference point corresponding to each pixel in the boundary segment to be propagated from the sagittal image includes: Any pixel in the boundary segment to be spread is designated as a temporary pixel, and the area in the sagittal image other than the reference blood vessel wall area is designated as the potential area to be spread. The intersection of the normal of the temporary pixel on the boundary segment to be spread and the possible region to be spread is determined as the target intersection line segment; Pixels that are at a preset distance from the temporary pixel are selected from the target intersection line segment and used as the propagation reference points corresponding to the temporary pixel.

7. The method for early warning of neonatal omphalitis infection risk by combining umbilical secretions and ultrasound according to claim 5, characterized in that, The step of dividing the potentially spreading infection zone to obtain the local analysis band sequence corresponding to the marked local region includes: Divide the line connecting each pixel in the boundary segment to be spread and its corresponding spread reference point into equal parts to obtain the sequence of equally divided points corresponding to each pixel in the boundary segment to be spread. Curve fitting is performed on the equally divided points with the same index in the equally divided point sequence corresponding to all pixels in the boundary segment to be spread to obtain candidate curves; Using each candidate curve as a dividing line, the potentially spreading infection zone is segmented to obtain local analysis zones; Based on the minimum distance between the local analysis band and the reference vessel wall region, all local analysis bands are sorted in ascending order to obtain a local analysis band sequence.

8. The method for early warning of neonatal omphalitis infection risk by combining umbilical secretions and ultrasound according to claim 3, characterized in that, The method for determining the scale indicators of spread and infiltration based on the infection spread factors of all target local areas in sagittal and coronal images includes: The sum of the infection spread factors of all target local areas in the sagittal plane image is determined as the sagittal spread factor; The cumulative value of the infection spread factor of all target local areas under coronal images is determined as the coronal spread factor; The product of the sagittal spread factor and the coronal spread factor is determined as the spread and infiltration scale index.

9. The method for early warning of neonatal omphalitis infection risk by combining umbilical secretions and ultrasound according to claim 1, characterized in that, The auxiliary warning value for omphalitis infection risk of the newborn to be tested is determined based on esterase activity distribution indicators, spread and infiltration scale indicators, and all vascular wall layering indices, including: The mean of all vessel wall stratification indices is determined as the representative index of vessel wall stratification. The cumulative product of the esterase activity distribution index, the spread and infiltration scale index, and the vascular wall layering representative index was normalized to obtain an auxiliary value for early warning of omphalitis infection risk.

Images