Device, method and system for E7 protein cell staining screening of cervical lesions

Abstract

The present invention relates to the field of image enhancement technology, and specifically to an E7 protein cell staining screening device, method and system for cervical lesions, the method comprising: obtaining a target RGB image corresponding to a stained cervical cell sample, extracting a connected domain from an area composed of non-cytoplasmic pixels; performing a cell nucleus shape regularity feature analysis process on a candidate connected domain, screening out a target connected domain that characterizes the nucleus in cells with uneven staining; enhancing all cytoplasm pixels according to different preset enhancement parameters; obtaining the R cytoplasm region and the B cytoplasm region of the target connected domain under the preset enhancement parameters; determining the target enhancement parameters corresponding to the target connected domain; and performing E7 protein cell staining screening based on the target enhanced image. The present invention realizes adaptive enhancement of unevenly stained cells in a target RGB image, improves the image enhancement effect, and thus improves the accuracy of E7 protein cell staining screening.

Description

Technical Field

[0001] The present invention relates to the technical field of image enhancement, and in particular to an E7 protein cell staining screening device, method and system for cervical lesions. Background Art

[0002] In the staining screening of E7 protein, it is often difficult to stain all cells evenly, that is, there will inevitably be cells with uneven staining. Since the cytoplasm accounts for a large proportion of the cell, uneven staining is often more likely to occur in the cytoplasm, resulting in unclear staining of some cytoplasm. At this time, if the doctor directly observes these cells through an electron microscope, it may cause him to ignore some cells with unclear staining, which may lead to fewer E7 proteins being identified, and then resulting in poor accuracy of E7 protein cell staining screening.

[0003] In order to facilitate doctors' observation, the RGB image of the cervical cell sample can often be enhanced. At present, the method usually adopted for image enhancement is: according to the grayscale histogram of the image, the image is histogram equalized to achieve image enhancement.

[0004] However, when histogram equalization is performed on the RGB image of the cervical cell sample according to the grayscale histogram of the RGB image of the cervical cell sample to achieve image enhancement, the following technical problems often occur:

[0005] Since grayscale histogram equalization often performs statistical image enhancement according to the grayscale value distribution of the image, directly performing histogram equalization on the RGB image of the cervical cell sample based on the grayscale histogram of the RGB image of the cervical cell sample may cause the loss of detail information with fewer pixels, resulting in poor enhancement effect on the RGB image of the cervical cell sample, and further resulting in poor accuracy of E7 protein cell staining screening. Summary of the invention

[0006] In order to solve the technical problem of poor accuracy in E7 protein cell staining screening, the present invention proposes an E7 protein cell staining screening device, method and system for cervical lesions.

[0007] In a first aspect, the present invention provides a method for screening cells for E7 protein staining of cervical lesions, the method comprising:

[0008] Obtain a target RGB image corresponding to the stained cervical cell sample, screen out non-cytoplasmic pixels and cytoplasmic pixels from the target RGB image based on the color channel differences corresponding to the pixels in the target RGB image, and perform connected domain extraction on the regions formed by all non-cytoplasmic pixels to obtain candidate connected domains;

[0009] Performing a cell nucleus shape regularity feature analysis on each candidate connected domain, obtaining a cell nucleus possible factor corresponding to each candidate connected domain, and based on the cell nucleus possible factor corresponding to the candidate connected domain and the channel value corresponding to the cytoplasm pixel point, screening out the target connected domain representing the cell nucleus in the unevenly stained cell from all the candidate connected domains;

[0010] According to different preset enhancement parameters, all cytoplasm pixels are enhanced to obtain the R channel value and B channel value of each cytoplasm pixel under each preset enhancement parameter;

[0011] According to the R channel value and the B channel value of the cytoplasm pixel point under each preset enhancement parameter, the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter are obtained;

[0012] Determine the target enhancement parameter corresponding to each target connected domain according to the difference between the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter;

[0013] According to the target enhancement parameters corresponding to the target connected domain, the target enhanced image corresponding to the target RGB image is obtained, and E7 protein cell staining screening is performed based on the target enhanced image.

[0014] In combination with the first aspect, in a possible implementation manner, obtaining a color channel difference corresponding to a pixel point in a target RGB image includes:

[0015] Determine any pixel point in the target RGB image as a marked pixel point;

[0016] Determine the difference between the original R channel value and the original G channel value corresponding to the marked pixel as the RG channel difference corresponding to the marked pixel, wherein the original R channel value and the original G channel value corresponding to the marked pixel are respectively the R channel value and the G channel value before the marked pixel is enhanced;

[0017] Determine the difference between the original R channel value and the original B channel value corresponding to the marked pixel as the RB channel difference corresponding to the marked pixel, wherein the original B channel value corresponding to the marked pixel is the B channel value before the marked pixel is enhanced;

[0018] The difference between the original G channel value and the original B channel value corresponding to the marked pixel point is determined as the GB channel difference corresponding to the marked pixel point;

[0019] The color channel difference corresponding to the marked pixel point is determined according to the RG channel difference, the RB channel difference and the GB channel difference corresponding to the marked pixel point, wherein the RG channel difference, the RB channel difference and the GB channel difference are all positively correlated with the color channel difference.

[0020] In combination with the first aspect above, in a possible implementation, performing a cell nucleus shape regularity feature analysis process on each candidate connected domain to obtain a possible cell nucleus factor corresponding to each candidate connected domain includes:

[0021] Determine any candidate connected domain as a marked connected domain, and determine the center point of the marked connected domain as the marked center point;

[0022] The distance between the center point of the mark and each edge pixel point in the mark connected domain is determined as the target distance corresponding to each edge pixel point in the mark connected domain;

[0023] According to the obtained target distance corresponding to all edge pixels in each candidate connected domain, the original B channel values ​​corresponding to all pixels in each candidate connected domain and the color channel difference, the possible factor of the cell nucleus corresponding to each candidate connected domain is determined, wherein the original B channel value corresponding to the pixel is the B channel value before the pixel is enhanced.

[0024] In combination with the first aspect above, in a possible implementation, the formula corresponding to the possible factor of the cell nucleus corresponding to the candidate connected domain is:

[0025] ;in, It is j The possible factors of the cell nucleus corresponding to the candidate connected domains; j is the serial number of the candidate connected domain; is a natural exponential function; It is j The mean of the color channel differences corresponding to all pixels in the candidate connected domain; It is j The mean of the original B channel values ​​corresponding to all pixels in the candidate connected domain; It is j The mean of the target distances corresponding to all edge pixels in the candidate connected domain; It is j The variance of the target distances corresponding to all edge pixels in the candidate connected domain.

[0026] In combination with the first aspect above, in a possible implementation, the method of screening out a target connected domain representing the nucleus in the unevenly stained cell from all candidate connected domains based on the possible factors of the nucleus corresponding to the candidate connected domain and the channel values ​​corresponding to the cytoplasm pixels includes:

[0027] If the cell nucleus possibility factor corresponding to the candidate connected domain is greater than the preset cell nucleus possibility threshold, the candidate connected domain is determined as the cell nucleus connected domain;

[0028] Taking the cytoplasm pixel points adjacent to each cell nucleus connected domain as seed points, regional growth is performed based on the original B channel values ​​corresponding to the cytoplasm pixel points to obtain the original B cytoplasm region corresponding to each cell nucleus connected domain, wherein the original B channel values ​​corresponding to the cytoplasm pixel points are the B channel values ​​of the cytoplasm pixel points before enhancement;

[0029] According to the original B channel value and position corresponding to the cytoplasm pixel point, all cytoplasm pixels are clustered to obtain the initial clustering cluster;

[0030] According to the total density value and inter-class distance between each two initial clustering clusters, the clustering reference index between each two initial clustering clusters is determined, wherein the total density value and the inter-class distance are both positively correlated with the clustering reference index;

[0031] Select the two initial clusters with the largest clustering reference index as the target clusters;

[0032] The difference between the original B channel value corresponding to each cytoplasm pixel and the average original B channel value of each target cluster is determined as the B channel deviation between each cytoplasm pixel and each target cluster, and two B channel deviations corresponding to each cytoplasm pixel are obtained, wherein the average original B channel value is the mean of the original B channel values ​​corresponding to all cytoplasm pixels in the target cluster;

[0033] From all cytoplasm pixels, the cytoplasm pixels whose corresponding two B channel deviations are greater than the preset deviation threshold are selected as uneven staining points;

[0034] The cell nucleus connected domain corresponding to the original B cytoplasm region to which the unevenly stained points belonged was determined as the target connected domain.

[0035] In combination with the first aspect above, in a possible implementation, obtaining the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter according to the R channel value and the B channel value of the cytoplasm pixel point under each preset enhancement parameter includes:

[0036] Determine any target connected domain as a reference connected domain, and determine any preset enhancement parameter as a marked enhancement parameter;

[0037] Taking the cytoplasm pixel points adjacent to the reference connected domain as seed points, performing region growing based on the R channel value of the cytoplasm pixel points under the label enhancement parameters, and obtaining the R cytoplasm region of the reference connected domain under the label enhancement parameters;

[0038] Taking the cytoplasm pixel points adjacent to the reference connected domain as seed points, region growing is performed based on the B channel value of the cytoplasm pixel points under the label enhancement parameters to obtain the B cytoplasm region of the reference connected domain under the label enhancement parameters.

[0039] In combination with the first aspect above, in a possible implementation manner, determining the target enhancement parameter corresponding to each target connected domain according to the difference between the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter includes:

[0040] According to the area difference, channel value difference and position difference between the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter, the formula corresponding to the enhancement effect evaluation value of each target connected domain under each preset enhancement parameter is determined as follows:

[0041] ;

[0042] ;in, It is a The target connected domain is in h Enhancement effect evaluation value under preset enhancement parameters; a is the serial number of the target connected domain; h is the serial number of the preset enhancement parameter; is the normalization function; It is the absolute value function; It is a The target connected domain is in h The area of ​​the R cytoplasmic region under the preset enhancement parameters; It is a The target connected domain is in h The area of ​​the B cell cytoplasm under the preset enhancement parameters; Indicates a The target connected domain is in h The area difference between the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters; It is a The target connected domain is in h The distance between the center point of the R cytoplasm region and the center point of the B cytoplasm region under the preset enhancement parameters, which represents the position difference between the R cytoplasm region and the B cytoplasm region; Indicates aThe target connected domain is in h The difference in channel values ​​between the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters; It is a The target connected domain is in h The number of cytoplasmic pixels in the union of the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters; t It is a The target connected domain is in h The serial number of the cytoplasmic pixel point in the union of the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters; It is a The target connected domain is in h In the union of the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters, t Cytoplasmic pixels in the h R channel value under preset enhancement parameters; It is a The target connected domain is in h In the union of the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters, t Cytoplasmic pixels in the h B channel value under preset enhancement parameters;

[0043] Any target connected domain is determined as a pending connected domain, and a preset enhancement parameter under which the enhancement effect evaluation value of the pending connected domain is the largest is selected from all preset enhancement parameters as the target enhancement parameter corresponding to the pending connected domain.

[0044] In combination with the first aspect above, in a possible implementation, obtaining a cervical cell sample by using a cervical brush includes:

[0045] After the patient completes vaginal dilation, the doctor takes out the reagent bottle with a scraping structure of the cervical brush, unscrews the protective bristle reagent bottle cap and places it stably on the table, and pulls the bristle protection sleeve to the rear end of the rotating handle, wherein the cervical brush includes: a cervical canal brush head, a cervical wall brush head, a reagent bottle with a scraping structure, a protective bristle reagent bottle cap, a bristle protection sleeve, a rotating handle and a handheld stable sleeve;

[0046] The doctor inserts the cervical brush into the vagina, inserts the cervical canal brush head into the cervical canal, visually observes that the cervical wall brush head spreads out and covers most of the cervical wall, holds the stable sleeve with the left hand, and rotates the rotating handle with the right hand to make the cervical canal brush head and cervical wall brush head start to rotate, and takes out the cervical brush after six weeks of rotation;

[0047] The doctor pushes the bristle protection sleeve to the front end to cover the brush head composed of the cervical canal brush head and the cervical wall brush head, and clamps the bristle protection sleeve on the mouth of the reagent bottle, pushes the brush head composed of the cervical canal brush head and the cervical wall brush head into the reagent bottle, so that the cervical canal brush head and the cervical wall brush head contact with the scraping structure in the reagent bottle, rotates the rotating handle in the opposite direction for six weeks, so that the exfoliated cells on the cervical canal brush head and the cervical wall brush head dissolve in the reagent, and then pulls out the cervical brush to complete the cell extraction, and makes the extracted cells into a cervical cell sample.

[0048] In a second aspect, the present invention provides an E7 protein cell staining screening system for cervical lesions, comprising a processor and a memory, wherein the processor is used to process instructions stored in the memory to implement the E7 protein cell staining screening method for cervical lesions. Specifically, the system comprises:

[0049] An acquisition screening and extraction module is used to obtain a target RGB image corresponding to the stained cervical cell sample, screen out non-cytoplasmic pixels and cytoplasmic pixels from the target RGB image based on the color channel differences corresponding to the pixels in the target RGB image, and perform connected domain extraction on the area formed by all non-cytoplasmic pixels to obtain a candidate connected domain;

[0050] The analysis and processing screening module is used to perform a cell nucleus shape regularity feature analysis on each candidate connected domain, obtain a cell nucleus possible factor corresponding to each candidate connected domain, and screen out a target connected domain representing the cell nucleus in the unevenly stained cell from all candidate connected domains based on the cell nucleus possible factor corresponding to the candidate connected domain and the channel value corresponding to the cytoplasm pixel point;

[0051] A pixel enhancement module is used to enhance all cytoplasm pixels according to different preset enhancement parameters to obtain the R channel value and B channel value of each cytoplasm pixel under each preset enhancement parameter;

[0052] A cytoplasm region acquisition module is used to acquire the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter according to the R channel value and the B channel value of the cytoplasm pixel point under each preset enhancement parameter;

[0053] An enhancement parameter determination module, used to determine the target enhancement parameter corresponding to each target connected domain according to the difference between the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter;

[0054] The image enhancement screening module is used to obtain a target enhanced image corresponding to the target RGB image according to the target enhancement parameters corresponding to the target connected domain, and perform E7 protein cell staining screening based on the target enhanced image.

[0055] In a third aspect, a server is provided, comprising 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, so that the device executes the method in the first aspect or any possible implementation of the first aspect.

[0056] In a fourth aspect, a computer program product is provided, comprising: a computer program code, which, when executed on a computer, enables the computer to execute the method in the first aspect or any possible implementation of the first aspect.

[0057] In the fifth aspect, an E7 protein cell staining screening device for cervical lesions is provided, and the E7 protein cell staining screening device for cervical lesions stores a computer program code. When the computer program code runs on a computer, the computer executes the method in the above-mentioned first aspect or any possible implementation of the first aspect.

[0058] The present invention has the following beneficial effects:

[0059] The E7 protein cell staining screening method for cervical lesions of the present invention realizes adaptive enhancement of unevenly stained cells in the target RGB image, improves the image enhancement effect, solves the technical problem of poor accuracy of E7 protein cell staining screening, and thus improves the accuracy of E7 protein cell staining screening. Compared with directly performing histogram equalization on the RGB image of the cervical cell sample according to the grayscale histogram of the RGB image of the cervical cell sample, the present invention comprehensively considers multiple features related to uneven cytoplasm staining, such as the target connected domain that characterizes the nucleus in the unevenly stained cell, the difference between the R cytoplasm region and the B cytoplasm region of the target connected domain under each preset enhancement parameter, etc., and the greater the difference between the R cytoplasm region and the B cytoplasm region of the target connected domain under a certain preset enhancement parameter, it often means that the uneven staining details in the cytoplasm region corresponding to the target connected domain are magnified under the preset enhancement parameter, which often means that the enhancement effect of the cytoplasm region corresponding to the target connected domain under the preset enhancement parameter is better, and it is easier for doctors to observe. Therefore, based on the target enhancement parameter corresponding to the target connected domain, the target enhanced image corresponding to the target RGB image is obtained, and the adaptive enhancement of the unevenly stained cytoplasm in the target RGB image is achieved, and the detail information of the uneven cytoplasm staining is retained to a certain extent, thereby improving the image enhancement effect, and then improving the accuracy of E7 protein cell staining screening. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0061] Figure 1 This is a flow chart of the E7 protein cell staining screening method for cervical lesions of the present invention;

[0062] Figure 2 It is a schematic diagram of the composition structure of the E7 protein cell staining screening system for cervical lesions of the present invention;

[0063] Figure 3 The figure is a schematic diagram of the structure of a computer device of the present invention. DETAILED DESCRIPTION

[0064] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the technical solutions proposed by the present invention are described in detail below in conjunction 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.

[0065] 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.

[0066] refer to Figure 1 , showing the process of some embodiments of the E7 protein cell staining screening method for cervical lesions of the present invention. The E7 protein cell staining screening method for cervical lesions comprises the following steps:

[0067] Step S1, obtaining a target RGB image corresponding to the stained cervical cell sample, screening out non-cytoplasmic pixels and cytoplasmic pixels from the target RGB image based on the color channel differences corresponding to the pixels in the target RGB image, and performing connected domain extraction on the area formed by all non-cytoplasmic pixels to obtain a candidate connected domain.

[0068] Among them, the cervical cell sample can be a cell sample obtained by sampling cervical columnar epithelial tissue through a cervical brush. The cervical brush can include: a cervical canal brush head, a cervical wall brush head, a reagent bottle with a scraping structure, a reagent bottle cap for protecting brush bristles, a brush bristle protection sleeve, a rotating handle and a handheld stable sleeve. The cervical canal brush head is made of medical silicone, tilted backward at 120 degrees, inserted into the cervical canal when in use, and can be used to scrape off the exfoliated cells in the cervical canal. The cervical wall brush head is made of medical PE (Polyethylene, polyethylene), kept vertical before use, and spread out after being squeezed by the cervical wall when in use, and is used to scrape off the exfoliated cells on the cervical wall. The reagent bottle with a scraping structure is a reagent bottle for containing reagents such as cell lysis solution. After the cervical brush completes sampling, the exfoliated cells on the cervical canal brush head and the cervical wall brush head can be scraped into the reagent. The reagent bottle cap for protecting brush bristles can be used to protect the reagent in the reagent bottle and to prevent the brush bristles from moving longitudinally. The bristle protection sleeve can be used to block the lateral movement of the bristles and keep the bristles vertical so that the bristles can be smoothly inserted into the reagent bottle after sampling. The rotating handle can be used to support all components of the cervical brush and drive the brush head to rotate when sampling at the cervical os and scraping cells from the reagent bottle. The handheld stabilizing sleeve can be used to keep the cervical brush stable when sampling at the cervical os and scraping cells from the reagent bottle, and when the cervical brush rotates as a whole.

[0069] Obtaining cervical cell samples through a cervical brush may include: after the patient completes vaginal dilation, the doctor removes the reagent bottle with a scraping structure of the cervical brush, unscrews the protective bristle reagent bottle cap and places it stably on the table, and pulls the bristle protection sleeve to the rear end of the rotating handle; the doctor inserts the cervical brush into the vagina, inserts the cervical canal brush head into the cervical canal, visually observes that the cervical wall brush head spreads out and covers most of the cervical wall, holds the stable sleeve with the left hand, and rotates the rotating handle with the right hand to make the cervical canal brush head and cervical wall brush head start to rotate, and removes it after six weeks of rotation Cervical brush; the doctor pushes the bristle protection sleeve to the front end to cover the brush head composed of the cervical canal brush head and the cervical wall brush head, and clamps the bristle protection sleeve on the mouth of the reagent bottle, pushes the brush head composed of the cervical canal brush head and the cervical wall brush head into the reagent bottle, so that the cervical canal brush head and the cervical wall brush head are in full contact with the scraping structure in the reagent bottle, rotates the rotating handle in the opposite direction for six weeks, so that the exfoliated cells on the cervical canal brush head and the cervical wall brush head are dissolved in the reagent, and then pulls out the cervical brush to complete the cell extraction, and the extracted cells are made into a cervical cell sample.

[0070] The target RGB image may be an RGB (Red Green Blue, color mode) image of a stained cervical cell sample, which may be obtained through an electron microscope. The color channel difference corresponding to the pixel may characterize the difference between the R channel, the G channel, and the B channel of the pixel. Non-cytoplasmic pixels may be pixels other than cytoplasmic pixels, for example, they may be cell nucleus pixels or background pixels. Cytoplasmic pixels may be pixels within the cytoplasmic region. Cell nucleus pixels may be pixels within the cell nucleus region. Background pixels may be pixels within the background region.

[0071] It should be noted that in the staining screening of E7 protein, a blue dye is usually used to mark the DNA (DeoxyriboNucleic Acid) of the cells to mark all cells, and then the colloidal gold method is used to stain the E7 protein. E7 protein is often present in the nucleus and cytoplasm of diseased cells. When the E7 protein is stained using the colloidal gold method, a red precipitate is often formed. When screening for cervical lesions, tissue samples are usually collected from the cervix to make stained sections, and the E7 protein in the stained sections is quantitatively analyzed. If the E7 protein accounts for a large proportion of the cells, it indicates that cervical lesions may exist. Among them, the tissue sample collected from the cervix is ​​also called a cervical cell sample, which can be obtained by sampling the cervical columnar epithelial tissue using a cervical brush. The stained section can represent the stained cervical cell sample.

[0072] As an example, this step may include the following steps:

[0073] The first step is to obtain the target RGB image corresponding to the stained cervical cell sample.

[0074] For example, an RGB image of a stained cervical cell sample may be taken through an electron microscope as a target RGB image.

[0075] The second step is to obtain the color channel difference corresponding to the pixel points in the target RGB image, which may include the following sub-steps:

[0076] In the first sub-step, any pixel in the target RGB image is determined as a marked pixel.

[0077] In the second sub-step, the difference between the original R channel value and the original G channel value corresponding to the marked pixel point is determined as the RG channel difference corresponding to the marked pixel point.

[0078] The original R channel value and the original G channel value corresponding to the marked pixel point may be the R channel value and the G channel value of the marked pixel point before enhancement, respectively.

[0079] In the third sub-step, the difference between the original R channel value and the original B channel value corresponding to the marked pixel point is determined as the RB channel difference corresponding to the marked pixel point.

[0080] The original B channel value corresponding to the marked pixel point may be the B channel value of the marked pixel point before enhancement.

[0081] In a fourth sub-step, the difference between the original G channel value and the original B channel value corresponding to the marked pixel point is determined as the GB channel difference corresponding to the marked pixel point.

[0082] In a fifth sub-step, the color channel difference corresponding to the marked pixel point is determined according to the RG channel difference, the RB channel difference and the GB channel difference corresponding to the marked pixel point.

[0083] Among them, the RG channel difference, the RB channel difference and the GB channel difference can all be positively correlated with the color channel difference.

[0084] For example, the formula for determining the color channel difference corresponding to the pixel point in the target RGB image can be:

[0085] ;

[0086] ;in, is the first i The color channel difference corresponding to the pixel points, that is Normalized value. i It is the pixel number in the target RGB image. is the first i The mean of the RG channel difference, RB channel difference and GB channel difference corresponding to each pixel. It is the absolute value function. is the first i The original R channel value corresponding to the pixel. is the first i The original G channel value corresponding to the pixel. is the first i The original B channel value corresponding to the pixel. is the first i The RG channel difference corresponding to each pixel. is the first i The RB channel difference corresponding to each pixel. is the first i The GB channel difference corresponding to each pixel.

[0087] It should be noted that, generally speaking, the stained cervical cell sample mainly contains three types of components, namely: background, cell nucleus and cytoplasm. The background is a cell-free area. Since it is an immunohistochemical staining, the color of the cell-free area is often the color of the stage that appears through the glass slide of the tissue section. For the convenience of observation, the stage color is often different from the color of the stained cells. It is usually white, and its RGB value is often approximately (255, 255, 255). Due to the high concentration of DNA in the cell nucleus, the cell nucleus is often painted with a heavier blue dye, making it visually darker in color. In addition, the embodiment of the present invention is mainly used to screen diseased cells. For the diseased cell nucleus, after being painted with a darker blue, it will also form a red precipitate when the E7 protein is stained using the colloidal gold method. After these colors are mixed, they are often approximately displayed as black, and their RGB values ​​are often approximately (0, 0, 0). The RGB three-channel values ​​corresponding to colors close to white or black are often relatively close. Therefore, the pixel point with a smaller difference between the RGB three-channel values ​​is more likely to be a background pixel point or a cell nucleus pixel point. When The smaller the value, the more likely it is that the target RGB image i The smaller the difference between the RGB three-channel values ​​corresponding to the pixel points, the smaller the difference between the RGB three-channel values ​​corresponding to the pixel points. i The more pixels there are, the more likely they are background pixels or cell nucleus pixels.

[0088] The third step, based on the color channel differences corresponding to the pixels in the target RGB image, screening out the non-cytoplasmic pixels and the cytoplasmic pixels from the target RGB image may include the following sub-steps:

[0089] In the first sub-step, pixels whose corresponding color channel differences are less than a preset channel difference threshold are selected from the target RGB image as non-cytoplasmic pixels.

[0090] The preset channel difference threshold may be a preset threshold for pixel point screening, for example, the preset channel difference threshold may be 0.1.

[0091] In the second sub-step, pixel points whose corresponding color channel differences are greater than or equal to a preset channel difference threshold are selected from the target RGB image as cytoplasm pixel points.

[0092] In the fourth step, connected domains are extracted from the regions formed by all non-cytoplasmic pixels, and the connected domains obtained at this time are recorded as candidate connected domains.

[0093] Step S2, performing a cell nucleus shape regularity feature analysis on each candidate connected domain to obtain a cell nucleus possible factor corresponding to each candidate connected domain, and based on the cell nucleus possible factor corresponding to the candidate connected domain and the channel value corresponding to the cytoplasm pixel point, selecting a target connected domain representing the cell nucleus in the unevenly stained cell from all candidate connected domains.

[0094] As an example, this step may include the following steps:

[0095] In the first step, any candidate connected domain is determined as a marked connected domain, and the center point of the marked connected domain is determined as the marked center point.

[0096] Among them, the method for obtaining the center point of the marked connected domain can be: obtain the coordinates of all pixel points in the marked connected domain in the image coordinate system, and record the mean of the horizontal coordinates corresponding to all pixel points in the marked connected domain as the horizontal representative value, and record the mean of the vertical coordinates corresponding to all pixel points in the marked connected domain as the vertical representative value, and record the coordinate point with the horizontal coordinate as the horizontal representative value and the vertical coordinate as the vertical representative value as the center point of the marked connected domain.

[0097] In the second step, the distance between the center point of the mark and each edge pixel point in the connected domain of the mark is determined as the target distance corresponding to each edge pixel point in the connected domain of the mark.

[0098] In the third step, the possible factor of the cell nucleus corresponding to each candidate connected domain is determined based on the target distance corresponding to all edge pixels in each candidate connected domain, the original B channel value corresponding to all pixels in each candidate connected domain, and the color channel difference.

[0099] The original B channel value corresponding to the pixel point may be the B channel value of the pixel point before enhancement.

[0100] For example, the formula for determining the possible factors of the cell nucleus corresponding to the candidate connected domain can be:

[0101] ;in, It is j The possible factors of the cell nucleus corresponding to the candidate connected domains. j is the sequence number of the candidate connected region. is a natural exponential function. It is j The mean of the color channel differences corresponding to all pixels in the candidate connected domain. It is j The mean of the original B channel values ​​corresponding to all pixels in the candidate connected domain can also be used j The mean of the gray values ​​corresponding to all pixels in the candidate connected domain is represented. It isj The mean of the target distances corresponding to all edge pixels in the candidate connected domain. It is j The variance of the target distances corresponding to all edge pixels in the candidate connected domain.

[0102] It should be noted that when The smaller the time, the more likely it is that j The smaller the difference between the RGB three-channel values ​​corresponding to the pixels in the first candidate connected domain, the smaller the difference between the RGB three-channel values ​​corresponding to the pixels in the first candidate connected domain. j The more likely the candidate connected region is to be a cell nucleus region or a background region. The smaller the time, the more likely it is that j The closer the channel value of the pixel points in the candidate connected domain before enhancement is to 0, the higher the value of the channel value of the pixel points in the candidate connected domain before enhancement is. j The closer the color of a candidate connected region is to black, the j The more likely the candidate connected domain is to be the cell nucleus. In practice, the cell nucleus is often small in area and its shape is often close to a circle. The smaller the time, the more likely it is that j The edge pixels of the candidate connected components are j The smaller the distance between the center points of the candidate connected regions, the smaller the distance between the center points of the candidate connected regions. j The smaller the area of ​​the candidate connected domain, the j The more candidate connected regions there are, the more likely they are to be cell nucleus regions. The smaller the time, the more likely it is that j The edge pixels of the candidate connected domain are j The more uniform the distance distribution between the center points of the candidate connected regions is, the better the j The more likely the shape of the candidate connected domain is to be circular, the more likely the j The more candidate connected regions there are, the more likely they are cell nucleus regions.

[0103] In the fourth step, if the cell nucleus possibility factor corresponding to the candidate connected domain is greater than the preset cell nucleus possibility threshold, the candidate connected domain is determined as the cell nucleus connected domain.

[0104] The preset cell nucleus possibility threshold may be a preset threshold for cell nucleus screening. For example, the preset cell nucleus possibility threshold may be equal to the mean value of the cell nucleus possibility factors corresponding to all candidate connected domains.

[0105] In the fifth step, the cytoplasm pixels adjacent to each cell nucleus connected domain are used as seed points, and region growing is performed based on the original B channel values ​​corresponding to the cytoplasm pixels to obtain the original B cytoplasm region corresponding to each cell nucleus connected domain.

[0106] The original B channel value corresponding to the cytoplasm pixel point may be the B channel value of the cytoplasm pixel point before enhancement.

[0107] For example, any cell nucleus connected domain is determined as a temporary connected domain, and the cytoplasm pixel point adjacent to the temporary connected domain is used as the initial seed point. Region growing is performed based on the original B channel value corresponding to the cytoplasm pixel point, that is, the cytoplasm pixel points with similar original B channel values ​​are divided into the same region, and the region after the region to which the initial seed point belongs is grown is recorded as the original B cytoplasm region corresponding to the temporary connected domain.

[0108] In the sixth step, all cytoplasm pixels are clustered according to the original B channel values ​​and positions corresponding to the cytoplasm pixels to obtain the initial clusters.

[0109] For example, a three-dimensional coordinate system is constructed with the horizontal axis of the image coordinate system as the horizontal axis, the vertical axis of the image coordinate system as the vertical axis, and the original B channel value corresponding to the cytoplasm pixel point as the vertical axis. According to the coordinates of all cytoplasm pixel points in the three-dimensional coordinate system, all cytoplasm pixel points are clustered by the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm, and the cluster cluster obtained at this time is recorded as the initial cluster cluster.

[0110] The seventh step is to determine the clustering reference index between each two initial clusters according to the total density value and the inter-cluster distance between each two initial clusters.

[0111] Among them, the total density value between the two initial clustering clusters can represent the total density between the two initial clustering clusters, which can be represented by the total number of pixels in the two initial clustering clusters. The inter-class distance between the two initial clustering clusters can be represented by the distance between the cluster centers of the two initial clustering clusters. Both the total density value and the inter-class distance can be positively correlated with the clustering reference index.

[0112] For example, the product of the total density value and the inter-class distance between every two initial clustering clusters may be determined as the clustering reference index between every two initial clustering clusters.

[0113] In the eighth step, two initial clusters with the largest clustering reference index are selected as the target clusters.

[0114] Among them, the pixels in the target cluster with the larger average original B channel value among the two target clusters can represent relatively uniformly colored negative cytoplasm pixels. Uniformly colored negative cytoplasm pixels often appear in blue. The pixels in the target cluster with the smaller average original B channel value among the two target clusters can represent relatively uniformly colored positive cytoplasm pixels. Uniformly colored positive cytoplasm pixels often appear in red.

[0115] In the ninth step, the difference between the original B channel value corresponding to each cytoplasm pixel and the average original B channel value of each target cluster is determined as the B channel deviation between each cytoplasm pixel and each target cluster, and two B channel deviations corresponding to each cytoplasm pixel are obtained.

[0116] The average original B channel value may be the average of the original B channel values ​​corresponding to all cytoplasmic pixels in the target cluster.

[0117] In the tenth step, the cytoplasm pixels whose corresponding two B channel deviations are greater than the preset deviation threshold are selected from all the cytoplasm pixels as the uneven staining points.

[0118] The preset deviation threshold may be a preset threshold for screening unevenly stained spots. For example, the preset deviation threshold may be 30. Unevenly stained spots may represent cytoplasm that is relatively unevenly colored.

[0119] In the eleventh step, the cell nucleus connected domain corresponding to the original B cytoplasm region to which the unevenly stained points belong is determined as the target connected domain.

[0120] Step S3, according to different preset enhancement parameters, all cytoplasm pixels are enhanced to obtain the R channel value and the B channel value of each cytoplasm pixel under each preset enhancement parameter.

[0121] The preset enhancement parameters may be pre-set parameters for image enhancement.

[0122] As an example, the cytoplasm pixels can be linearly enhanced. When the cytoplasm pixels are linearly enhanced, each preset enhancement parameter may include an enhancement slope and an enhancement intercept. The R channel or the B channel can be recorded as a candidate channel, and all cytoplasm pixels can be enhanced according to different preset enhancement parameters. The formula corresponding to the candidate channel of each cytoplasm pixel under each preset enhancement parameter can be obtained as follows:

[0123] ;in, It is m Cytoplasmic pixels in the h Candidate channel values ​​under preset enhancement parameters. mis the serial number of the cytoplasmic pixel. h It is the sequence number of the preset enhancement parameter. It is h The preset enhancement parameters include an enhancement slope, and the value range of the enhancement slope can be {0.5, 0.6, 0.7, 0.8, 0.9, 1}. It is m Candidate channel values ​​for cytoplasmic pixels before enhancement. It is h The first preset enhancement parameter includes an enhancement intercept, and the value range of the enhancement intercept may be {50, 40, 30, 20, 10, 0}. For example, the first preset enhancement parameter includes an enhancement slope of 0.5, and the enhancement intercept includes 50.

[0124] It should be noted that It can be characterized in h Under the preset enhancement parameters, m The corresponding channel value obtained when the cytoplasm pixel is enhanced.

[0125] Step S4, obtaining the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter according to the R channel value and the B channel value of the cytoplasm pixel under each preset enhancement parameter.

[0126] As an example, this step may include the following steps:

[0127] In the first step, any target connected domain is determined as a reference connected domain, and any preset enhancement parameter is determined as a marked enhancement parameter.

[0128] In the second step, the cytoplasm pixel points adjacent to the above-mentioned reference connected domain are used as seed points, and region growing is performed based on the R channel value of the cytoplasm pixel points under the above-mentioned labeling enhancement parameters to obtain the R cytoplasm region of the above-mentioned reference connected domain under the above-mentioned labeling enhancement parameters.

[0129] For example, the cytoplasm pixels adjacent to the reference connected domain are taken as initial seed points, and region growing is performed based on the R channel values ​​of the cytoplasm pixels under the label enhancement parameters, that is, the cytoplasm pixels with similar R channel values ​​under the label enhancement parameters are divided into the same region, and the regions obtained after the regions to which these initial seed points belong are grown are recorded as the R cytoplasm regions of the reference connected domain under the label enhancement parameters.

[0130] In the third step, the cytoplasm pixel points adjacent to the above-mentioned reference connected domain are used as seed points, and region growing is performed based on the B channel value of the cytoplasm pixel points under the above-mentioned labeling enhancement parameters to obtain the B cytoplasm region of the above-mentioned reference connected domain under the above-mentioned labeling enhancement parameters.

[0131] Among them, the method for obtaining the B cytoplasm region of the reference connected domain under the label enhancement parameters is the same as the method for obtaining the R cytoplasm region of the reference connected domain under the label enhancement parameters, which will not be repeated here.

[0132] Step S5, determining the target enhancement parameter corresponding to each target connected domain according to the difference between the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter.

[0133] As an example, this step may include the following steps:

[0134] In the first step, according to the area difference, channel value difference and position difference between the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter, the formula corresponding to the enhancement effect evaluation value of each target connected domain under each preset enhancement parameter is determined as follows:

[0135] ;

[0136] ;in, It is a The target connected domain is in h The enhancement effect evaluation value under the preset enhancement parameters. a is the sequence number of the target connected domain. h It is the sequence number of the preset enhancement parameter. is a normalization function. It is the absolute value function. It is a The target connected domain is in h The area of ​​the R cytoplasmic region under the preset enhancement parameters. It is a The target connected domain is in h The area of ​​the B cell cytoplasmic region under the preset enhancement parameters. Indicates a The target connected domain is in h The area difference between the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters. It is a The target connected domain is in h The distance between the center point of the R cytoplasm region and the center point of the B cytoplasm region under the preset enhancement parameters indicates the position difference between the R cytoplasm region and the B cytoplasm region. Indicates a The target connected domain is in h The channel value difference between the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters. It is a The target connected domain is inh The number of cytoplasmic pixels in the union of the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters. t It is a The target connected domain is in h The sequence number of the cytoplasmic pixel point in the union of the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters. It is a The target connected domain is in h In the union of the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters, t Cytoplasmic pixels in the h R channel value under the preset enhancement parameters. It is a The target connected domain is in h In the union of the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters, t Cytoplasmic pixels in the h B channel value under the preset enhancement parameters.

[0137] It should be noted that the details of the unevenly stained cytoplasm can be amplified by amplifying the difference between the R channel and the B channel in the unevenly stained cytoplasm area. In other words, the greater the difference between the R channel and the B channel in the unevenly stained cytoplasm area after image enhancement, the better the image enhancement effect. The larger the a The target connected domain is in h The larger the area difference between the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters, the larger the area difference between the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters. a The target connected domain is in h The better the enhancement effect under the preset enhancement parameters. The larger the a The target connected domain is in h The greater the position difference between the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters, the greater the position difference between the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters. a The target connected domain is in h The better the enhancement effect under the preset enhancement parameters. The larger the a The target connected domain is in h The greater the difference in channel values ​​between the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters, the greater the difference in channel values ​​between the R cytoplasm region and the B cytoplasm region. a The target connected domain is in h The better the enhancement effect under the preset enhancement parameters. The larger the a The target connected domain is inh The better the enhancement effect under the preset enhancement parameters.

[0138] In the second step, any target connected domain is determined as a pending connected domain, and the preset enhancement parameter under which the enhancement effect evaluation value of the pending connected domain is the largest is selected from all preset enhancement parameters as the target enhancement parameter corresponding to the pending connected domain.

[0139] Step S6, obtaining a target enhanced image corresponding to the target RGB image according to the target enhancement parameters corresponding to the target connected domain, and performing E7 protein cell staining screening based on the target enhanced image.

[0140] As an example, this step may include the following steps:

[0141] The first step is to obtain the target enhanced image corresponding to the target RGB image according to the target enhancement parameters corresponding to the target connected domain.

[0142] For example, the R channel values, G channel values, and B channel values ​​corresponding to all cytoplasm pixels in the original B cytoplasm region corresponding to each target connected domain in the target RGB image can be updated to their R channel values, G channel values, and B channel values ​​under the target enhancement parameters corresponding to the corresponding target connected domain, thereby achieving enhancement of the original B cytoplasm region corresponding to each target connected domain in the target RGB image, and the final enhanced target RGB image is used as the target enhanced image.

[0143] The original B cytoplasm region can represent the cytoplasm. The method for obtaining the G channel value of the cytoplasm pixel under the target enhancement parameter can be the same as the method for obtaining the R channel value of the cytoplasm pixel under the target enhancement parameter, which will not be repeated here.

[0144] In the second step, doctors can better perform E7 protein cell staining screening by observing the target enhanced image.

[0145] refer to Figure 2 Based on the same inventive concept as the above method embodiment, the present invention provides an E7 protein cell staining screening system for cervical lesions, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the above computer program is executed by the processor, the steps of implementing the E7 protein cell staining screening method for cervical lesions may specifically include:

[0146] The acquisition screening and extraction module 201 is used to obtain the target RGB image corresponding to the stained cervical cell sample, screen out non-cytoplasmic pixels and cytoplasmic pixels from the target RGB image based on the color channel difference corresponding to the pixels in the target RGB image, and perform connected domain extraction on the area formed by all non-cytoplasmic pixels to obtain a candidate connected domain;

[0147] The analysis processing and screening module 202 is used to perform a cell nucleus shape regularity feature analysis on each candidate connected domain, obtain a cell nucleus possible factor corresponding to each candidate connected domain, and screen out a target connected domain representing the cell nucleus in the unevenly stained cell from all candidate connected domains based on the cell nucleus possible factor corresponding to the candidate connected domain and the channel value corresponding to the cytoplasm pixel point;

[0148] The pixel enhancement module 203 is used to enhance all cytoplasm pixels according to different preset enhancement parameters to obtain the R channel value and the B channel value of each cytoplasm pixel under each preset enhancement parameter;

[0149] A cytoplasm region acquisition module 204 is used to acquire the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter according to the R channel value and the B channel value of the cytoplasm pixel point under each preset enhancement parameter;

[0150] An enhancement parameter determination module 205 is used to determine a target enhancement parameter corresponding to each target connected domain according to the difference between the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter;

[0151] The image enhancement screening module 206 is used to obtain a target enhanced image corresponding to the target RGB image according to the target enhancement parameters corresponding to the target connected domain, and perform E7 protein cell staining screening based on the target enhanced image.

[0152] Figure 3 is a schematic diagram of the structure of a computer device provided by an embodiment of the present invention. Figure 3 As 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, wherein when the processor 302 executes the computer program 303, the computer device can execute any one of the E7 protein cell staining screening methods for cervical lesions introduced above.

[0153] Based on the same inventive concept as the above method embodiment, the present invention provides a server, 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, so that the device executes any one of the above-mentioned E7 protein cell staining screening methods for cervical lesions.

[0154] Based on the same inventive concept as the above-mentioned method embodiment, the present invention provides a computer program product, which includes: computer program code, when the computer program code is run on a computer, the computer executes any one of the above-mentioned E7 protein cell staining screening methods for cervical lesions.

[0155] Based on the same inventive concept as the above-mentioned method embodiments, the present invention provides an E7 protein cell staining screening device for cervical lesions, and the E7 protein cell staining screening device for cervical lesions stores computer program code. When the computer program code runs on a computer, the computer executes any one of the above-mentioned E7 protein cell staining screening methods for cervical lesions.

[0156] In summary, compared with directly performing histogram equalization on the RGB image of the cervical cell sample according to the grayscale histogram of the RGB image of the cervical cell sample, the present invention comprehensively considers multiple features related to uneven cytoplasm staining, such as the target connected domain that characterizes the nucleus in the unevenly stained cell, the difference between the R cytoplasm region and the B cytoplasm region of the target connected domain under each preset enhancement parameter, etc., and the greater the difference between the R cytoplasm region and the B cytoplasm region of the target connected domain under a certain preset enhancement parameter, it often means that the uneven staining details in the cytoplasm region corresponding to the target connected domain are magnified under the preset enhancement parameter, which often means that the enhancement effect of the cytoplasm region corresponding to the target connected domain under the preset enhancement parameter is better, and it is easier for doctors to observe. Therefore, based on the target enhancement parameter corresponding to the target connected domain, the target enhanced image corresponding to the target RGB image is obtained, and the adaptive enhancement of the unevenly stained cytoplasm in the target RGB image is achieved, and the detail information of the uneven cytoplasm staining is retained to a certain extent, thereby improving the image enhancement effect, and then improving the accuracy of E7 protein cell staining screening.

[0157] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features can be replaced by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for screening E7 protein cells for cervical lesions, characterized in that: The following steps are involved: Obtain a target RGB image corresponding to the stained cervical cell sample, screen out non-cytoplasmic pixels and cytoplasmic pixels from the target RGB image based on the color channel differences corresponding to the pixels in the target RGB image, and perform connected domain extraction on the regions formed by all non-cytoplasmic pixels to obtain candidate connected domains; Performing a cell nucleus shape regularity feature analysis on each candidate connected domain, obtaining a cell nucleus possible factor corresponding to each candidate connected domain, and based on the cell nucleus possible factor corresponding to the candidate connected domain and the channel value corresponding to the cytoplasm pixel point, screening out the target connected domain representing the cell nucleus in the unevenly stained cell from all the candidate connected domains; According to different preset enhancement parameters, all cytoplasm pixels are enhanced to obtain the R channel value and B channel value of each cytoplasm pixel under each preset enhancement parameter; According to the R channel value and the B channel value of the cytoplasm pixel point under each preset enhancement parameter, the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter are obtained; Determine the target enhancement parameter corresponding to each target connected domain according to the difference between the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter; According to the target enhancement parameters corresponding to the target connected domain, a target enhanced image corresponding to the target RGB image is obtained, and E7 protein cell staining screening is performed based on the target enhanced image; The cell nucleus shape regularity feature analysis processing is performed on each candidate connected domain to obtain the possible cell nucleus factors corresponding to each candidate connected domain, including: Determine any candidate connected domain as a marked connected domain, and determine the center point of the marked connected domain as the marked center point; The distance between the center point of the mark and each edge pixel point in the mark connected domain is determined as the target distance corresponding to each edge pixel point in the mark connected domain; According to the obtained target distance corresponding to all edge pixels in each candidate connected domain, the original B channel value corresponding to all pixels in each candidate connected domain and the color channel difference, the possible factor of the cell nucleus corresponding to each candidate connected domain is determined, wherein the original B channel value corresponding to the pixel is the B channel value before the pixel is enhanced; The formula corresponding to the possible factor of the cell nucleus corresponding to the candidate connected domain is: ;in, It is j The possible factors of the cell nucleus corresponding to the candidate connected domains; j is the serial number of the candidate connected domain; is a natural exponential function; It is j The mean of the color channel differences corresponding to all pixels in the candidate connected domain; It is j The mean of the original B channel values ​​corresponding to all pixels in the candidate connected domain; It is j The mean of the target distances corresponding to all edge pixels in the candidate connected domain; It is j The variance of the target distances corresponding to all edge pixels in the candidate connected domain.

2. A method for screening E7 protein cells for cervical lesions according to claim 1, characterized in that: Get the color channel differences corresponding to the pixels in the target RGB image, including: Determine any pixel point in the target RGB image as a marked pixel point; Determine the difference between the original R channel value and the original G channel value corresponding to the marked pixel as the RG channel difference corresponding to the marked pixel, wherein the original R channel value and the original G channel value corresponding to the marked pixel are respectively the R channel value and the G channel value before the marked pixel is enhanced; Determine the difference between the original R channel value and the original B channel value corresponding to the marked pixel as the RB channel difference corresponding to the marked pixel, wherein the original B channel value corresponding to the marked pixel is the B channel value before the marked pixel is enhanced; The difference between the original G channel value and the original B channel value corresponding to the marked pixel point is determined as the GB channel difference corresponding to the marked pixel point; The color channel difference corresponding to the marked pixel point is determined according to the RG channel difference, the RB channel difference and the GB channel difference corresponding to the marked pixel point, wherein the RG channel difference, the RB channel difference and the GB channel difference are all positively correlated with the color channel difference.

3. A method for screening E7 protein cells for cervical lesions according to claim 1, characterized in that: The method of screening out target connected domains representing the nuclei in cells with uneven staining from all candidate connected domains based on the possible factors of the nuclei corresponding to the candidate connected domains and the channel values ​​corresponding to the cytoplasm pixels includes: If the cell nucleus possibility factor corresponding to the candidate connected domain is greater than the preset cell nucleus possibility threshold, the candidate connected domain is determined as the cell nucleus connected domain; Taking the cytoplasm pixel points adjacent to each cell nucleus connected domain as seed points, regional growth is performed based on the original B channel values ​​corresponding to the cytoplasm pixel points to obtain the original B cytoplasm region corresponding to each cell nucleus connected domain, wherein the original B channel values ​​corresponding to the cytoplasm pixel points are the B channel values ​​of the cytoplasm pixel points before enhancement; According to the original B channel value and position corresponding to the cytoplasm pixel point, all cytoplasm pixels are clustered to obtain the initial clustering cluster; According to the total density value and inter-class distance between each two initial clustering clusters, the clustering reference index between each two initial clustering clusters is determined, wherein the total density value and the inter-class distance are both positively correlated with the clustering reference index; Select the two initial clusters with the largest clustering reference index as the target clusters; The difference between the original B channel value corresponding to each cytoplasm pixel and the average original B channel value of each target cluster is determined as the B channel deviation between each cytoplasm pixel and each target cluster, and two B channel deviations corresponding to each cytoplasm pixel are obtained, wherein the average original B channel value is the mean of the original B channel values ​​corresponding to all cytoplasm pixels in the target cluster; From all cytoplasm pixels, the cytoplasm pixels whose corresponding two B channel deviations are greater than the preset deviation threshold are selected as uneven staining points; The cell nucleus connected domain corresponding to the original B cytoplasm region to which the unevenly stained points belonged was determined as the target connected domain.

4. A method for E7 protein cell staining screening for cervical lesions according to claim 1, characterized in that: The step of obtaining the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter according to the R channel value and the B channel value of the cytoplasm pixel point under each preset enhancement parameter comprises: Determine any target connected domain as a reference connected domain, and determine any preset enhancement parameter as a marked enhancement parameter; Taking the cytoplasm pixel points adjacent to the reference connected domain as seed points, performing region growing based on the R channel value of the cytoplasm pixel points under the label enhancement parameters, and obtaining the R cytoplasm region of the reference connected domain under the label enhancement parameters; Taking the cytoplasm pixel points adjacent to the reference connected domain as seed points, region growing is performed based on the B channel value of the cytoplasm pixel points under the label enhancement parameters to obtain the B cytoplasm region of the reference connected domain under the label enhancement parameters.

5. The method for E7 protein cell staining screening for cervical lesions according to claim 1, characterized in that: Determining the target enhancement parameter corresponding to each target connected domain according to the difference between the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter includes: According to the area difference, channel value difference and position difference between the R cytoplasm region and the B cytoplasm region of each target connected domain under each preset enhancement parameter, the formula corresponding to the enhancement effect evaluation value of each target connected domain under each preset enhancement parameter is determined as follows: ； ;in, It is a The target connected domain is in h Enhancement effect evaluation value under preset enhancement parameters; a is the serial number of the target connected domain; h is the serial number of the preset enhancement parameter; is the normalization function; It is the absolute value function; It is a The target connected domain is in h The area of ​​the R cytoplasmic region under the preset enhancement parameters; It is a The target connected domain is in h The area of ​​the B cell cytoplasm under the preset enhancement parameters; Indicates a The target connected domain is in h The area difference between the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters; It is a The target connected domain is in h The distance between the center point of the R cytoplasm region and the center point of the B cytoplasm region under the preset enhancement parameters, which represents the position difference between the R cytoplasm region and the B cytoplasm region; Indicates a The target connected domain is in h The channel value difference between the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters; It is a The target connected domain is in h The number of cytoplasmic pixels in the union of the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters; t It is a The target connected domain is in h The serial number of the cytoplasmic pixel point in the union of the R cytoplasmic region and the B cytoplasmic region under the preset enhancement parameters; It is a The target connected domain is in h In the union of the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters, t Cytoplasmic pixels in the h R channel value under preset enhancement parameters; It is a The target connected domain is in h In the union of the R cytoplasm region and the B cytoplasm region under the preset enhancement parameters, t Cytoplasmic pixels in the h B channel value under preset enhancement parameters; Any target connected domain is determined as a pending connected domain, and a preset enhancement parameter under which the enhancement effect evaluation value of the pending connected domain is the largest is selected from all preset enhancement parameters as the target enhancement parameter corresponding to the pending connected domain.

6. A method for E7 protein cell staining screening for cervical lesions according to claim 1, characterized in that: A cervical cell sample is obtained by a cervical brush, wherein the cervical brush includes: a cervical canal brush head, a cervical wall brush head, a reagent bottle with a scraping structure, a protective brush bristle reagent bottle cap, a brush bristle protection sleeve, a rotating handle and a handheld stable sleeve.

7. An E7 protein cell staining screening system for cervical lesions, characterized in that: The method comprises a processor and a memory, wherein the processor is used to process instructions stored in the memory to implement a method for E7 protein cell staining screening for cervical lesions according to any one of claims 1 to 6.

8. An E7 protein cell staining screening device for cervical lesions, characterized in that: Computer program codes are stored, and when the computer program codes are run on a computer, the computer is enabled to execute a method for screening cells for E7 protein staining of cervical lesions as described in any one of claims 1 to 6.

Images