Chromosome segmentation method, apparatus, device, and storage medium

Abstract

The application relates to a chromosome segmentation method and device, computer equipment and a storage medium. The method comprises the following steps: extracting a skeleton line of a chromosome contour in a preliminary segmentation image of a chromosome; determining the type of the chromosome according to the skeleton line; if the type represents that the chromosome belongs to an overlapping chromosome, outputting at least one candidate overlapping segmentation scheme corresponding to the type to an interactive interface; obtaining a target overlapping segmentation scheme specified from the candidate overlapping segmentation schemes based on the interactive interface; and executing the target overlapping segmentation scheme to segment the overlapping chromosome in the preliminary segmentation image into a single chromosome. The scheme can improve the segmentation accuracy.

Description

Technical Field

[0001] This application relates to the fields of computer technology and image processing technology, and in particular to a chromosome segmentation method, apparatus, device and storage medium. Background Technology

[0002] With the development of science and technology, various advanced analytical techniques have emerged. Among them, chromosome image segmentation plays a very important role in many fields (such as medicine or genetic research).

[0003] Traditional image segmentation methods employ relatively uniform and fixed segmentation approaches. However, chromosome images themselves are prone to many complex situations; for example, chromosomes often overlap. Using traditional, fixed, and uniform segmentation methods is simply unsuitable for the complexities of chromosome images, easily leading to missegmentation. Therefore, the segmentation accuracy of traditional methods is relatively low. Summary of the Invention

[0004] Therefore, it is necessary to provide a chromosome segmentation method, apparatus, computer equipment, and storage medium that can improve the accuracy of segmentation in response to the above-mentioned technical problems.

[0005] A chromosome segmentation method, the method comprising:

[0006] Extract the skeleton lines of the chromosome outline from the preliminary segmentation image of the chromosome;

[0007] The type of chromosome is determined based on the skeletal lines;

[0008] If the type indicates that the chromosome belongs to an overlapping chromosome, then

[0009] Output at least one candidate overlapping segmentation scheme corresponding to the type to the interactive interface;

[0010] Obtain the target overlapping segmentation scheme specified from the candidate overlapping segmentation schemes based on the interactive interface;

[0011] The target overlap segmentation scheme is executed to segment the overlapping chromosomes in the preliminary segmentation image into single chromosomes.

[0012] In one embodiment, the method further includes:

[0013] Convert the initial chromosome image to a grayscale image;

[0014] The grayscale image is subjected to adaptive histogram equalization.

[0015] The image after adaptive histogram equalization is binarized to obtain a preliminary segmentation image of the chromosome.

[0016] In one embodiment, the skeleton lines of the chromosome outline in the preliminary segmentation image of the extracted chromosomes include:

[0017] Extract chromosome outlines from the preliminary segmentation image of chromosomes;

[0018] The chromosome contours are filtered according to at least one of the following: the number of contour pixels and the roundness of the contour.

[0019] Extract the skeletal lines of the chromosome outlines that are retained after screening.

[0020] In one embodiment, the method further includes:

[0021] Calculate the average distance from each pixel on the chromosome outline to the center of the region;

[0022] Calculate the deviation between the distance from each pixel on the chromosome outline to the center of the region and the average distance;

[0023] The outline roundness of the chromosome profile is determined based on the average distance and the deviation.

[0024] In one embodiment, the method further includes:

[0025] Determine the average shortest distance from each pixel on the retained chromosome outline after filtering to the skeleton line;

[0026] Remove chromosome contours whose average distance is less than a preset distance threshold from the retained chromosome contours after screening;

[0027] The position coordinates of each pixel on the remaining chromosome outline after removal are smoothed using a sliding mean filter.

[0028] In one embodiment, determining the chromosome type based on the skeleton line includes:

[0029] Obtain the number of endpoints and intersections of the skeleton lines;

[0030] The type of chromosome is determined based on the number of endpoints and the number of intersections.

[0031] In one embodiment, the method further includes:

[0032] Extract the protrusions and indentations on the chromosome outline;

[0033] The step of outputting at least one candidate overlapping segmentation scheme corresponding to the type to the interactive interface includes:

[0034] Based on the convex points and the concave points, at least one candidate overlapping segmentation scheme corresponding to the type is generated;

[0035] Output the candidate overlapping segmentation scheme to the interactive interface.

[0036] In one embodiment, the method further includes:

[0037] If the type is S, then the chromosome is determined to be a single chromosome, and the preliminary segmentation image is output as the final chromosome segmentation result.

[0038] If the type is any one of T-type, X-type, and H-type, then the chromosome is determined to be an overlapping chromosome; wherein, T-type, X-type, and H-type are used to characterize different overlapping types of chromosomes.

[0039] A chromosome splitting device, the device comprising:

[0040] A type recognition module is used to extract the skeleton lines of chromosome outlines from the preliminary segmentation image of chromosomes; and to determine the type of chromosome based on the skeleton lines.

[0041] An interaction module is used to output at least one candidate overlapping segmentation scheme corresponding to the type to the interaction interface if the type indicates that the chromosome belongs to overlapping chromosomes; and to obtain a target overlapping segmentation scheme specified from the candidate overlapping segmentation schemes based on the interaction interface.

[0042] The segmentation module is used to execute a target overlap segmentation scheme to segment overlapping chromosomes in the preliminary segmentation image into single chromosomes.

[0043] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the chromosome segmentation methods of various embodiments of this application.

[0044] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the chromosome segmentation methods of the various embodiments of this application.

[0045] The aforementioned chromosome segmentation method, apparatus, computer equipment, and storage medium extract the skeleton lines of chromosome outlines from the preliminary segmentation image of chromosomes, identify chromosome types based on the skeleton lines, determine overlapping chromosomes according to their types, and then output corresponding overlap segmentation schemes for the user to choose from for each overlapping chromosome. Since the user-selected target overlap segmentation scheme is more accurate and targeted, automatically executing the user-selected target overlap segmentation scheme, compared to traditional image segmentation processing using fixed and uniform methods lacking specificity, can more accurately segment overlapping chromosomes in the preliminary segmentation image into single chromosomes. Attached Figure Description

[0046] Figure 1 This is a diagram illustrating the application environment of the chromosome segmentation method in one embodiment;

[0047] Figure 2 This is a flowchart illustrating a chromosome segmentation method in one embodiment;

[0048] Figure 3 This is a schematic diagram of an S-type chromosome in one embodiment;

[0049] Figure 4 This is a schematic diagram of a T-type chromosome in one embodiment;

[0050] Figure 5 This is a schematic diagram of an X-type chromosome in one embodiment;

[0051] Figure 6 This is a schematic diagram of an H-type chromosome in one embodiment;

[0052] Figure 7 This is a simplified diagram illustrating the principle of a chromosome segmentation method in one embodiment;

[0053] Figure 8 This is a structural block diagram of a chromosome segmentation device in one embodiment;

[0054] Figure 9 This is a structural block diagram of the chromosome segmentation device in another embodiment;

[0055] Figure 10 This is an internal structural diagram of a computer device in one embodiment;

[0056] Figure 11 This is a diagram of the internal structure of a computer device in another embodiment. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0058] The chromosome segmentation method provided in this application can be applied to, for example... Figure 1 In the application environment shown, computer device 102 can implement the chromosome segmentation method in various embodiments of this application through interaction with users 104, thereby accurately segmenting chromosomes. Computer device 102 can be a terminal or a server. Figure 1 This illustration uses only a terminal as an example. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be a standalone server or a server cluster composed of multiple servers. It can be understood that if computer device 102 is a terminal, then computer device 102 can directly interact with user 104 based on the interactive interface. If computer device 102 is a server, then computer device 102 can indirectly interact with user 104 based on the interactive interface provided by the terminal. Figure 1 The illustration uses computer device 102 as an example.

[0059] Computer device 102 can extract the skeleton lines of chromosome outlines from the preliminary segmentation image; based on the skeleton lines, the chromosome type is determined. If the type indicates that the chromosomes are overlapping chromosomes, computer device 102 can output at least one candidate overlapping segmentation scheme corresponding to that chromosome type to the interactive interface. User 104 (e.g., a doctor) can then select a target overlapping segmentation scheme for further segmentation from the candidate overlapping segmentation schemes based on the interactive interface. Computer device 102 can then execute the target overlapping segmentation scheme selected by user 104 to segment the overlapping chromosomes in the preliminary segmentation image into single chromosomes.

[0060] In one embodiment, such as Figure 2 As shown, a chromosome segmentation method is provided. This embodiment illustrates the application of this method to a computer device, which can be a terminal or a server. It is understood that this method can also be applied to a system including both a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0061] Step 202: Extract the skeleton lines of the chromosome outline from the preliminary segmentation image of the chromosome.

[0062] Preliminary segmentation refers to the segmentation process that only distinguishes chromosomes from the background in the image, without differentiating between chromosomes themselves. A preliminary segmented image is a segmented image where chromosomes are the foreground region, and all other parts of the chromosome image are the background region.

[0063] It is understood that the preliminary segmentation image only distinguishes chromosomes from the background region, but does not segment the chromosomes themselves. Therefore, overlapping chromosomes remain overlapping in the preliminary segmentation image and are not segmented into individual chromosomes. Thus, steps 204-210 in this embodiment are needed to segment the overlapping chromosomes in the preliminary segmentation image into individual chromosomes.

[0064] Chromosome outline refers to the external shape of a chromosome. Skeletal lines are lines located within the chromosome outline and used to represent the internal structure of the chromosome. In essence, skeletal lines provide a simple illustration of the chromosome's shape and structure.

[0065] To facilitate understanding of the relationship between the skeletal lines and chromosome outlines, we will now combine... Figure 3 Provide a illustrative illustration. (Refer to...) Figure 3 The outer areas are the chromosome outlines, and the white line 302 is used to indicate the skeletal outline.

[0066] Specifically, the computer device can first extract chromosome outlines from the preliminary segmentation image of the chromosomes, and then extract skeleton lines based on the chromosome outlines. It can be understood that the computer device can directly extract skeleton lines based on the extracted chromosome outlines, or it can preprocess the initially extracted chromosome outlines and extract skeleton lines based on the preprocessed outlines.

[0067] In one embodiment, the computer device can directly acquire a preliminary segmented image of the chromosome that has already been segmented. Alternatively, the computer device can perform preliminary segmentation on an initial chromosome image to obtain a preliminary segmented image of the chromosome.

[0068] In one embodiment, the computer device can directly perform binarization segmentation on the initial chromosome image to obtain a preliminary segmented image of the chromosome.

[0069] In one embodiment, the computer device may also preprocess the initial chromosome image and perform binarization segmentation on the preprocessed image to obtain a preliminary segmented image of the chromosome.

[0070] In one embodiment, preprocessing of the initial chromosome image by the computer device may include at least one of conversion to grayscale and adaptive histogram equalization. It should be noted that the computer device may also perform one or more other preprocessing steps on the initial chromosome image.

[0071] In one embodiment, the preprocessing of the initially extracted chromosome outline by the computer device may include at least one of the following preprocessing methods: hole filling, impurity screening and filtering, and cyst filtering.

[0072] In other embodiments, the computer device may not perform contour extraction but instead obtain the skeleton lines of the chromosome contour using other methods. For example, the computer device can use a skeleton line extraction template to directly extract the skeleton lines of the chromosome contour from the preliminary segmentation image. This application does not limit the method of skeleton line extraction in its embodiments.

[0073] Step 204: Determine the chromosome type based on the skeleton lines.

[0074] Specifically, after extracting the skeletal outline of a chromosome, the computer device can identify the chromosome type based on this outline. This can be understood as the computer device performing overall curve morphology recognition on the skeletal outline to identify the chromosome type; that is, the computer device can identify the chromosome type by combining the overall global features of the skeletal outline. The computer device can also identify the chromosome type based on local features of the skeletal outline. For example, it can identify the chromosome type based on the endpoints and / or intersections of the skeletal outline.

[0075] In one embodiment, the type of chromosome can be classified according to its morphology.

[0076] In one embodiment, the chromosome types classified according to morphology include at least one of the following: S-type, T-type, X-type, and H-type.

[0077] In other embodiments, chromosome types may not be classified according to chromosome morphology. For example, they may be classified directly according to whether they are single or multiple, meaning chromosome types include single and multiple chromosomes. Alternatively, chromosome types may be classified directly according to whether they overlap or not, meaning chromosome types include both overlapping and non-overlapping types. There is no limitation on how chromosome types are classified here.

[0078] Step 206: If the type characterization chromosome belongs to overlapping chromosomes, then output at least one candidate overlapping segmentation scheme corresponding to the type to the interactive interface.

[0079] The overlap segmentation scheme is a scheme used to segment overlapping chromosomes. That is, a scheme used to separate overlapping chromosomes from each other.

[0080] It is understandable that chromosome type can characterize whether the chromosomes segmented in the initial segmentation image are overlapping chromosomes. For example, for chromosome types classified by morphology, if the chromosome type is S, it means that the chromosome is a single chromosome and is not overlapping; if the chromosome type is X, it means that the chromosome is overlapping. Furthermore, for chromosome types directly classified by overlap or non-overlap, the overlapping chromosome type can necessarily characterize whether the chromosomes segmented in the initial segmentation image are overlapping chromosomes.

[0081] Specifically, if the type characterization chromosome belongs to overlapping chromosomes, the computer device can obtain at least one candidate overlap segmentation scheme corresponding to that chromosome type. The computer device can output the obtained candidate overlap segmentation scheme to an interactive interface for the user to view and select, thereby realizing human-computer interaction.

[0082] It is understandable that if the computer device is a terminal, it can directly display an interactive interface to show the candidate overlapping segmentation schemes, thereby enabling direct human-computer interaction with the user. If the computer device is a server, the server can send the candidate overlapping segmentation schemes to the terminal, which then provides an interactive interface to display these schemes, thus achieving indirect interaction between the server and the user.

[0083] In one embodiment, at least one candidate overlap segmentation scheme may be pre-set for each chromosome type. Thus, the computer device can acquire the pre-set candidate overlap segmentation scheme corresponding to the identified chromosome type.

[0084] In another embodiment, candidate overlapping segmentation schemes may not be pre-set. The computer device can also generate at least one candidate overlapping segmentation scheme corresponding to the chromosome type based on the extracted chromosome contour. This dynamically generates candidate overlapping segmentation schemes corresponding to the chromosome type based on the actual characteristics of the chromosome contour, thereby improving the accuracy of the overlapping segmentation scheme.

[0085] It is understandable that if the type characterization chromosome does not belong to overlapping chromosomes, that is, it belongs to a single chromosome, then the preliminary segmentation image can be directly used as the final chromosome segmentation result.

[0086] Step 208: Obtain the target overlapping segmentation scheme specified from the candidate overlapping segmentation schemes based on the interactive interface.

[0087] Among them, the target overlap segmentation scheme refers to the specified overlap segmentation scheme used to perform advanced segmentation of overlapping chromosomes in the initial segmentation image.

[0088] Specifically, after displaying candidate overlapping segmentation schemes in the interactive interface, users can select the final target overlapping segmentation scheme from them.

[0089] It is understandable that if the computer device is a terminal, the terminal can directly obtain the target overlap segmentation scheme selected by the user based on the interactive interface. If the computer device is a server, after the user selects the target overlap segmentation scheme based on the interactive interface, the terminal can notify the server of the target overlap segmentation scheme, thereby enabling the server to obtain the specified target overlap segmentation scheme.

[0090] In one embodiment, in addition to displaying candidate overlap segmentation schemes, the interactive interface can also display preliminary segmentation images of the chromosomes and the chromosome type. Users can then determine the overlap pattern of the chromosomes based on the chromosome type and the preliminary segmentation image, thereby selecting an accurate and appropriate target overlap segmentation scheme.

[0091] Users can be professionals or non-professionals with basic knowledge. Professionals are individuals with high-level expertise. Professionals may include at least one of the following: doctors and professional analysts. Non-professionals with basic knowledge do not require as high a level of expertise; they only need basic knowledge of chromosomes or genetics, such as someone who has studied genetics and can determine chromosome segmentation.

[0092] It is understood that since the solution of this application can automatically provide candidate overlapping segmentation schemes for users to choose from, users can more easily determine which overlapping segmentation scheme to adopt for overlapping segmentation based on the segmentation results in the preliminary segmentation image and the provided overlapping segmentation schemes. This reduces the requirements for users' professional skills, that is, it reduces the difficulty and improves convenience and applicability.

[0093] Step 210: Execute the target overlap segmentation scheme to segment the overlapping chromosomes in the preliminary segmentation image into single chromosomes.

[0094] Specifically, after acquiring a specified target overlap segmentation scheme through human-computer interaction, the computer device can automatically execute the segmentation processing steps or logic within that scheme to automatically segment overlapping chromosomes in the initial segmentation image into individual chromosomes. This ultimately achieves complete chromosome segmentation, yielding the final chromosome segmentation result.

[0095] The aforementioned chromosome segmentation method extracts the skeleton lines of chromosome outlines from the initial segmentation image and identifies chromosome types based on these skeleton lines. That is, it identifies the overlapping morphology of chromosomes based on their morphological characteristics. The overlapping chromosomes are then determined according to their type, and corresponding segmentation schemes are output for the user to choose from. Since the user-selected target segmentation scheme is more accurate and targeted—that is, it can select a more suitable target segmentation scheme for different overlapping morphologies of chromosomes—executing the user-selected target segmentation scheme reduces segmentation errors compared to traditional image segmentation processing that uses fixed and uniform methods lacking specificity. This results in more accurately segmenting overlapping chromosomes in the initial segmentation image into single chromosomes.

[0096] Furthermore, combining automated segmentation with interactive selection of overlapping segmentation schemes improves efficiency compared to manual chromosome segmentation by professionals.

[0097] In one embodiment, the method further includes a preliminary segmentation step of the initial chromosome image, specifically including the following steps: converting the initial chromosome image into a grayscale image; performing adaptive histogram equalization on the grayscale image; and performing binarization segmentation on the image after adaptive histogram equalization to obtain a preliminary segmented image of the chromosome.

[0098] Among them, adaptive histogram equalization is used to amplify the difference in grayscale values ​​between chromosomes and the background in a grayscale image.

[0099] Specifically, the computer device can read the initial chromosome image and convert it into a grayscale image. The computer device can then perform adaptive histogram equalization on the grayscale image to increase the difference in grayscale values ​​between the chromosomes and the background. Finally, the computer device can perform binarization segmentation on the image after adaptive histogram equalization to obtain a preliminary segmented image of the chromosomes.

[0100] In one embodiment, the computer device can perform adaptive histogram equalization on the grayscale image according to the following formula:

[0101] υ(s)=βυ w (s)+(1-β)υ b (s);

[0102] Where: 0≤β≤1 represents the weight; υ w (s) indicates histogram balance within the window, υ b v(s) represents the histogram equalization outside the window. v(s) represents the result after adaptive histogram equalization.

[0103] In one embodiment, the computer device can use the Otsu method (maximum inter-class variance method) to perform binarization segmentation on the image after adaptive histogram equalization.

[0104] In one embodiment, a computer device may perform binarization segmentation on an image after adaptive histogram equalization according to the following formula:

[0105] u = ω0 * u0 + ω1 * u1;

[0106] g = ω0 * (u0 - u) 2 +ω1*(u1-u) 2 =ω0*ω1*(u1-u0)*(u0-u1);

[0107] Where u is the total average gray level of the image, ω0 is the proportion of foreground pixels to the image, u0 is the average gray level of the foreground, ω1 is the proportion of background pixels to the image, u1 is the average gray level of the background; and g is the variance of the foreground and background images.

[0108] It is understandable that the foreground and background are segmented using a threshold T. The computer device can use an iterative method to obtain the threshold T that maximizes the variance g, which is then used as the final binarization segmentation threshold. Based on this binarization segmentation threshold, the image after adaptive histogram equalization is binarized, and the chromosome becomes the foreground image, while the rest of the image becomes the background image.

[0109] In the above embodiments, performing grayscale conversion, adaptive histogram equalization, and then binarization segmentation can improve the accuracy of the initial segmentation.

[0110] In one embodiment, step 202, extracting the skeleton lines of the chromosome outline in the preliminary segmentation image of the chromosome, includes: extracting the chromosome outline from the preliminary segmentation image of the chromosome; filtering the chromosome outline according to at least one of the number of outline pixels and the roundness of the outline; and extracting the skeleton lines of the chromosome outlines retained after filtering.

[0111] Here, the number of outline pixels refers to the number of pixels on the chromosome outline. Outline roundness refers to the roundness of the chromosome outline.

[0112] Specifically, the computer device can initially extract chromosome outlines from preliminary segmented images of chromosomes. The computer device can then filter the chromosome outlines based on at least one of the following: the number of outline pixels and the roundness of the outline, and extract the skeleton lines of the retained chromosome outlines after filtering.

[0113] It should be noted that screening can either select the chromosome outlines that need to be retained, or it can filter out (i.e., remove) the chromosome outlines that do not need to be retained, so that the remaining outlines are the chromosome outlines that need to be retained.

[0114] In one embodiment, the computer device can filter chromosome outlines according to the number of outline pixels.

[0115] In one embodiment, a computer device can filter contours whose number of contour pixels meets a preset quantity condition.

[0116] Specifically, the computer device can filter out chromosome contours whose number of outline pixels falls within a preset range and remove them. In one embodiment, the preset range can be greater than 200 and less than 10,000. Other ranges are also possible and not limited thereto. The computer device can also filter out chromosome contours whose number of outline pixels is less than a preset threshold and remove them. Alternatively, the computer device can also filter out chromosome contours whose number of outline pixels is greater than a preset threshold and retain them, while the unfiltered chromosome contours are removed. It is understood that filtering by the number of pixels is generally suitable for removing impurities.

[0117] In one embodiment, the computer device can screen chromosome outlines based on their roundness.

[0118] In one embodiment, a computer device can filter contours whose contour roundness meets a preset roundness condition.

[0119] Specifically, the computer device can calculate the roundness of chromosome outlines. The computer device can filter out chromosome outlines with a roundness less than a preset roundness threshold, removing them as impurities. For example, the preset roundness threshold can be 0.6, or it can be any other threshold; there is no limitation on this. Alternatively, the computer device can also filter out chromosome outlines with a roundness greater than the preset roundness threshold, retaining them, while the unfiltered outlines are removed as impurities. In other embodiments, the computer device can also filter out chromosome outlines with roundness within a preset roundness range, removing them as impurities. There is no limitation on this. It is understood that roundness-based filtering is generally suitable for removing chromosome cysts.

[0120] In one embodiment, the computer device can further filter chromosome contours based on both the number of contour pixels and the roundness of the contour, to select chromosome contours that meet a preset number condition for the number of pixels and a preset roundness condition for the roundness of the contour. For example, the computer device can filter out chromosome contours with a number of pixels greater than 200 but less than 10,000 and a roundness less than 0.6, thus removing them.

[0121] In the above embodiments, the initially extracted chromosome contours are filled with holes; the chromosome contours are then filtered according to at least one of the following: the number of contour pixels and the roundness of the contour, which can accurately select effective chromosome contours. Furthermore, the accuracy of skeleton line extraction can be improved based on the selected chromosome contours.

[0122] In one embodiment, the method further includes: filling holes in the chromosome contours extracted from the preliminary segmentation image. In this embodiment, the chromosome contours are filtered according to at least one of the number of contour pixels and the roundness of the contour, including: filtering the hole-filled chromosome contours according to at least one of the number of contour pixels and the roundness of the contour.

[0123] It is understandable that the initially extracted chromosome outlines may contain holes. Therefore, computer equipment can fill these holes in the chromosome outlines extracted from the initial segmentation image to repair them. Furthermore, the computer equipment can filter the repaired chromosome outlines after hole filling based on at least one of the following: the number of outline pixels and the roundness of the outline. This improves accuracy.

[0124] In one embodiment, the method further includes a step of determining the roundness of the outline, specifically including the following steps: calculating the average distance from each pixel on the chromosome outline to the center of the region; calculating the deviation between the distance from each pixel on the chromosome outline to the center of the region and the average distance; and determining the roundness of the outline on the chromosome outline based on the average distance and the deviation.

[0125] It should be noted that if the extracted chromosome contours are perforated before screening, then the chromosome contours in this embodiment are the perforated chromosome contours. If the chromosome contours extracted directly from the initial segmentation image without any other preprocessing are screened, then the chromosome contours in this embodiment are the chromosome contours initially extracted from the initial segmentation image. For ease of description, the following descriptions in this embodiment will be collectively referred to as chromosome contours.

[0126] The center of the region is the point on the chromosome.

[0127] Specifically, for each chromosome outline, the computer device can calculate the average distance from each pixel on the chromosome outline to the center of the region. The computer device can also calculate the distance from each individual pixel on the chromosome outline to the center of the region. Then, the computer device can determine the deviation between the individual distance from the center of the region and the average distance. Furthermore, the computer device can determine the roundness of the outline based on the average distance and the deviation.

[0128] In one embodiment, the computer device can calculate the contour roundness according to the following formula:

[0129]

[0130]

[0131]

[0132]

[0133] Where Distance is the average distance from a pixel on the chromosome outline to the center of the region; F is the number of pixels on the chromosome outline; p is the center of the region (i.e., the mass point). i Let Σ be the i-th pixel on the chromosome outline; Sigma is the deviation between the distance from each pixel on the chromosome outline to the center of the region and the average distance Distance; Sides is the outline roundness; and Roundness is the constraint relationship between the average distance Distance and the deviation Sigma.

[0134] In the above embodiments, calculating the average distance and deviation of each pixel on the chromosome outline from the center of the region can accurately and conveniently determine the roundness of the outline.

[0135] In one embodiment, the method further includes: determining the average shortest distance from each pixel on the retained chromosome contours after filtering to the skeleton line; removing chromosome contours from the retained chromosome contours whose average shortest distance is less than a preset distance threshold; and smoothing the position coordinates of each pixel on the remaining chromosome contours after removal according to a sliding mean filter.

[0136] Specifically, after extracting the skeleton lines of the chromosome contour, the computer device can determine the average shortest distance from each pixel on the retained chromosome contour after filtering to the skeleton lines. The computer device can remove chromosome contours with an average shortest distance less than a preset distance threshold, thus further optimizing the chromosome contours. The computer device can then smooth the position coordinates of each pixel on the remaining chromosome contours after removal using a sliding mean filter, resulting in a smooth contour.

[0137] In one embodiment, the average shortest distance D from each pixel on the chromosome outline to the skeleton line can be calculated using the following formula:

[0138]

[0139] Where x1 and y1, and x2 and y2 are respectively the pixels on the chromosome outline and the pixels on the skeleton line that are closest to the pixels on the chromosome outline.

[0140] In one embodiment, the computer device can smooth the position coordinates of each pixel on the remaining chromosome outline after removal according to the following formula:

[0141] g(i) = ∑f(i) / m;

[0142] Where m is the number of pixels in the smoothed set, f(i) represents the position coordinates of the i-th pixel on the contour before smoothing, and g(i) represents the position coordinates of the pixel after smoothing. It can be understood that the position coordinates can include both X and Y coordinates.

[0143] In the above embodiments, impurities are removed from the chromosome contour based on the average shortest distance from each pixel to the skeleton line. A sliding mean filter is then used to smooth the position coordinates of each pixel on the remaining chromosome contour after removal, resulting in a smooth, noise-reduced contour. Furthermore, subsequent processing based on this smoothed, noise-reduced contour improves processing accuracy. For example, the smoothed contour allows for more accurate extraction of convex and concave points, leading to the generation of more accurate candidate overlapping segmentation schemes. Consequently, after specifying a target overlapping segmentation scheme, segmentation can be performed more accurately.

[0144] In one embodiment, step 204, determining the chromosome type based on the skeleton line, includes: obtaining the number of endpoints and intersections of the skeleton line; and determining the chromosome type based on the number of endpoints and intersections.

[0145] In this context, the endpoint is the point on the skeleton line. The number of intersections refers to the points where the skeleton lines intersect. The number of endpoints refers to the number of endpoints on the skeleton line. The number of intersections refers to the number of points where the skeleton lines intersect.

[0146] Specifically, the computer equipment can identify the endpoints and intersections of the skeleton lines, and determine the type of chromosome in the preliminary segmented image based on the number of endpoints and intersections of the skeleton lines.

[0147] In one embodiment, if the number of endpoints is 2 and the number of intersections is 0, then the chromosome is determined to be S-type.

[0148] If the number of endpoints is 3 and the number of intersections is 1, then the chromosome is determined to be T-type. If the number of endpoints is 4 and the number of intersections is 1, then the chromosome is determined to be X-type. If the number of endpoints is 4 and the number of intersections is 2, then the chromosome is determined to be H-type.

[0149] In the above embodiments, the type of chromosome can be quickly and accurately identified based on local features such as the number of endpoints and intersections on the skeleton line, thereby saving system resources.

[0150] In one embodiment, the method further includes: extracting convex and concave points on the chromosome outline. In this embodiment, step 208, outputting at least one candidate overlapping segmentation scheme corresponding to the type to the interactive interface, includes: generating at least one candidate overlapping segmentation scheme corresponding to the type based on the convex and concave points; and outputting the candidate overlapping segmentation scheme to the interactive interface.

[0151] In this context, a concave point refers to a recessed location on the chromosome outline near the intersection of skeletal lines. A convex point can include the endpoints of skeletal lines. It can be understood that the endpoints of skeletal lines lie on the chromosome outline.

[0152] Specifically, the computer device can combine skeletal lines to identify convex and concave points on the chromosome outline, extracting these points. Based on these points, the computer device can generate at least one candidate overlapping segmentation scheme corresponding to the chromosome type. That is, by analyzing the features of the chromosome outline, at least one candidate overlapping segmentation scheme corresponding to the chromosome type is dynamically generated. The computer device can then output the generated candidate overlapping segmentation scheme to an interactive interface.

[0153] It is understood that the chromosome contours in this application embodiment can be the chromosome contours initially extracted from a chromosome image, or the contours retained after filtering the initially extracted chromosome contours, or the contours obtained after removing chromosome contours whose average shortest distance to the skeleton line is less than a preset distance threshold, and then smoothing the position coordinates using a sliding mean filter. No limitation is made in this regard. It should be noted that the contours after position coordinate smoothing, combined with the skeleton line, effectively remove impurities, and thus the extraction of convex and concave points based on the contours after position coordinate smoothing is more accurate, thereby generating more accurate candidate overlapping segmentation schemes.

[0154] In the above embodiments, at least one candidate overlapping segmentation scheme corresponding to the type is dynamically generated based on the convex and concave points. That is, by combining the features on the contour, candidate overlapping segmentation schemes are dynamically generated, which improves the accuracy of the generated candidate overlapping segmentation schemes. Furthermore, after specifying the target overlapping segmentation scheme, more accurate segmentation can be performed according to the target overlapping segmentation scheme.

[0155] In one embodiment, the method further includes: if the chromosome type is S, then the chromosome is determined to be a single chromosome, and the preliminary segmentation image is output as the final chromosome segmentation result; if the type is any one of T, X, and H, then the chromosome is determined to be an overlapping chromosome; wherein, T, X, and H are used to characterize different overlapping types of chromosomes.

[0156] Specifically, if the computer device identifies the chromosome type as S, it can determine that the chromosome is a single chromosome and not an overlapping chromosome. Therefore, the computer device can directly output the preliminary segmentation image as the final chromosome segmentation result without executing subsequent steps 208 and subsequent steps. If the computer device identifies the chromosome type as T, X, or H, it can determine that the chromosome is an overlapping chromosome, and thus execute steps 208 and subsequent steps.

[0157] It is understandable that overlapping chromosomes can have different overlap types or overlap patterns. T-type, X-type, and H-type are used to characterize different overlap types of chromosomes.

[0158] The computer device can output corresponding candidate overlapping segmentation schemes to the interactive interface for different types. Each chromosome type corresponds to at least one candidate overlapping segmentation scheme.

[0159] To facilitate understanding, we will now combine... Figures 3-6 Provide a illustrative explanation.

[0160] Reference Figure 3 The chromosome is S-shaped, meaning it is a single chromosome, and no additional segmentation is required. Therefore, the preliminary segmentation image can be output as the final chromosome segmentation result.

[0161] Reference Figure 4 The chromosome is T-shaped, which is an overlapping chromosome and needs to be further segmented. Therefore, two overlapping segmentation schemes are displayed for this T-shaped chromosome so that users (e.g., doctors) can choose from them.

[0162] Reference Figure 5 The chromosome is X-shaped, which is an overlapping chromosome and needs to be further segmented. Therefore, three overlapping segmentation schemes are displayed for users (e.g., doctors) to choose from for this X-shaped chromosome.

[0163] Reference Figure 6 The chromosome is of type H, which is an overlapping chromosome and needs to be further segmented. Therefore, three overlapping segmentation schemes are displayed for users (e.g., doctors) to choose from for this type H chromosome.

[0164] from Figures 4-6 It can be intuitively observed that although T-type to H-type chromosomes are all overlapping chromosomes, the type or manner of overlap varies greatly. Automatic segmentation might be inaccurate in such cases. Therefore, an overlap segmentation scheme can be provided for users to choose from, and automatic segmentation can then be performed based on the user-selected scheme to improve accuracy. Furthermore, even within the same chromosome type, morphological differences can be significant. For example, with the X-type chromosome... Figure 5Chromosomes overlapping in an X-shape are generally straight. However, in some cases, X-shaped overlapping chromosomes exhibit greater curvature. Therefore, the difference between X-shaped overlapping chromosomes with greater curvature and those with straighter curvature is significant, making them unsuitable for the same overlap segmentation scheme. Thus, multiple overlap segmentation schemes can be provided for the same type of chromosome, allowing users to select the most accurate scheme. Automatic segmentation based on the user-selected scheme can accurately separate overlapping chromosomes in the initial segmentation image into individual chromosomes. This improves accuracy compared to overlapping segmentation using a preset scheme.

[0165] Figure 7 This is a simplified diagram illustrating the chromosome segmentation principle in one embodiment. See also... Figure 7 The computer device can read chromosome images and preprocess them. Preprocessing may include grayscale conversion, histogram equalization, binarization, and opening / closing operations. Then, the computer device can filter contours. Specifically, it can extract skeleton lines, measure the distance from the contour to the skeleton lines, and filter out contours with excessively small distances. It can also remove or fill in hole contours. Furthermore, the computer device can obtain the convex and concave points, as well as the endpoints and intersections of the skeleton lines, from the retained contours after filtering. Based on the endpoints and intersections, it can identify the chromosome type and generate corresponding candidate overlap segmentation schemes for each type, outputting them for user interaction. Finally, for different types, based on the selected target overlap segmentation scheme, it can separate single chromosomes and overlapping chromosomes (i.e., intersecting chromosomes).

[0166] It should be understood that although the steps in the flowcharts of the various embodiments of this application are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts of the various embodiments of this application may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of steps or stages in other steps.

[0167] like Figure 8 As shown, in one embodiment, a chromosome segmentation device is provided, the device comprising: a type identification module 802, an interaction module 804, and a segmentation module 806; wherein:

[0168] The type recognition module 802 is used to extract the skeleton lines of the chromosome outline in the preliminary segmentation image of the chromosome; and to determine the type of the chromosome based on the skeleton lines.

[0169] The interaction module 804 is used to output at least one candidate overlapping segmentation scheme corresponding to the type to the interaction interface if the type indicates that the chromosome belongs to overlapping chromosomes; and to obtain the target overlapping segmentation scheme specified from the candidate overlapping segmentation schemes based on the interaction interface.

[0170] The segmentation module 806 is used to execute a target overlap segmentation scheme to segment overlapping chromosomes in the preliminary segmentation image into single chromosomes.

[0171] In one embodiment, the segmentation module 806 is further configured to convert the initial chromosome image into a grayscale image; perform adaptive histogram equalization processing on the grayscale image; and perform binarization segmentation processing on the image after adaptive histogram equalization processing to obtain a preliminary segmented image of the chromosome.

[0172] In one embodiment, the type recognition module 802 is further configured to extract chromosome contours from the preliminary segmentation image of the chromosomes; filter the chromosome contours according to at least one of the number of contour pixels and the roundness of the contours; and extract the skeleton lines of the chromosome contours retained after filtering.

[0173] In one embodiment, the type recognition module 802 is further configured to calculate the average distance from each pixel on the chromosome outline to the center of the region; calculate the deviation between the distance from each pixel on the chromosome outline to the center of the region and the average distance; and determine the outline roundness of the chromosome outline based on the average distance and the deviation.

[0174] like Figure 9 As shown, in one embodiment, the device further includes:

[0175] The coordinate smoothing module 803 is used to determine the average shortest distance from each pixel on the retained chromosome contour after filtering to the skeleton line; remove chromosome contours whose average distance is less than a preset distance threshold from the retained chromosome contours after filtering; and smooth the position coordinates of each pixel on the remaining chromosome contours after removal according to the sliding mean filter.

[0176] In one embodiment, the type identification module 802 is further configured to obtain the number of endpoints and intersections of the skeleton lines; and determine the type of the chromosome based on the number of endpoints and intersections.

[0177] In one embodiment, the interaction module 804 is further configured to extract convex and concave points on the chromosome outline; generate at least one candidate overlapping segmentation scheme corresponding to the type based on the convex and concave points; and output the candidate overlapping segmentation scheme to the interaction interface.

[0178] In one embodiment, the interaction module 804 is further configured to determine that the chromosome belongs to a single chromosome if the type is S-type, and output the preliminary segmentation image as the final chromosome segmentation result; and to determine that the chromosome belongs to an overlapping chromosome if the type is any one of T-type, X-type, and H-type; wherein, T-type, X-type, and H-type are used to characterize different overlapping types of chromosomes.

[0179] The aforementioned chromosome segmentation device performs preliminary segmentation on a chromosome image, extracts the skeleton lines of the chromosome outlines from the preliminary segmentation image, identifies the chromosome type based on the skeleton lines, determines overlapping chromosomes according to their type, and then outputs corresponding overlap segmentation schemes for the user to choose from for each overlapping chromosome. Since the user-selected target overlap segmentation scheme is more accurate and targeted, automatically executing the user-selected target overlap segmentation scheme, compared to traditional image segmentation processing using fixed and uniform methods lacking specificity, can more accurately segment overlapping chromosomes in the preliminary segmentation image into single chromosomes.

[0180] Specific limitations regarding the chromosome segmentation device can be found in the limitations of the chromosome segmentation method described above, and will not be repeated here. Each module in the aforementioned chromosome segmentation device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of the processor in a computer device, or stored in software in the memory of a computer device, so that the processor can call and execute the operations corresponding to each module.

[0181] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 10 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When executed by the processor, the computer program implements a chromosome segmentation method.

[0182] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 11 As shown. The computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a chromosome segmentation method. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0183] Those skilled in the art will understand that Figure 10 The structure shown in Figure 11 is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0184] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0185] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0186] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0187] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0188] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A chromosome segmentation method, characterized by, The method comprises: extracting a chromosome contour from a preliminary segmentation image of a chromosome, the preliminary segmentation image being a segmentation image in which the chromosome is a foreground region and other than the chromosome is a background region; screening the chromosome contour according to a contour pixel point number and a contour roundness, the contour pixel point number being a number of pixel points on the chromosome contour, and the contour roundness being a roundness of the chromosome contour, wherein the contour roundness calculation step comprises: calculating an average distance of each pixel point on the chromosome contour to a region center, the region center being a mass point of the chromosome, calculating a deviation between the distance of each pixel point on the chromosome contour to the region center and the average distance, and determining the contour roundness of the chromosome contour according to the average distance and the deviation; extracting a skeleton line of the chromosome contour retained after screening; determining an average shortest distance of each pixel point on the chromosome contour retained after screening to the skeleton line, removing the chromosome contour with an average distance less than a preset distance threshold from the chromosome contour retained after screening, and performing smoothing processing on position coordinates of each pixel point on the chromosome contour remaining after removal according to a sliding mean filter to obtain a smoothed chromosome contour; determining a type of the chromosome according to a number of end points on the skeleton line and a number of intersection points of the skeleton line, the number of end points being a number of end points on the skeleton line, and the number of intersection points being a number of intersection points on the skeleton line; if the number of end points is 2 and the number of intersection points is 0, judging that the type of the chromosome is S type; if the number of end points is 3 and the number of intersection points is 1, judging that the type of the chromosome is T type; if the number of end points is 4 and the number of intersection points is 1, judging that the type of the chromosome is X type; and if the number of end points is 4 and the number of intersection points is 2, judging that the type of the chromosome is H type; in a case where the type of the chromosome is S type, outputting the preliminary segmentation image as a final chromosome segmentation result to an interactive interface, and further displaying the preliminary segmentation image of the chromosome and the type of the chromosome in the interactive interface; in a case where the type of the chromosome is T type, outputting a plurality of candidate overlapping segmentation schemes dynamically generated for the T type to the interactive interface for user selection; in a case where the type of the chromosome is X type, outputting a plurality of candidate overlapping segmentation schemes dynamically generated for the X type to the interactive interface for user selection; in a case where the type of the chromosome is H type, outputting a plurality of candidate overlapping segmentation schemes dynamically generated for the H type to the interactive interface for user selection; obtaining a target overlapping segmentation scheme specified by a user from the plurality of candidate overlapping segmentation schemes based on the interactive interface; performing the target overlapping segmentation scheme specified by the user to segment the overlapping chromosomes in the preliminary segmentation image into single chromosomes.

2. The method of claim 1, wherein, The method further comprises: converting an initial chromosome image into a grayscale image; performing adaptive histogram equalization processing on the grayscale image; performing binary segmentation processing on the image after adaptive histogram equalization processing to obtain a preliminary segmentation image of the chromosome.

3. The method of claim 1, wherein, The method further comprises: extracting convex points and concave points on the chromosome contour; generating at least one candidate overlapping segmentation scheme corresponding to the type according to the convex points and the concave points.

4. A chromosome segmentation apparatus characterized by comprising: The device comprises: A type identification module is configured to extract a chromosome contour from a preliminary segmentation image of a chromosome, the preliminary segmentation image being a segmentation image in which the chromosome is a foreground region and other regions are background regions; filter the chromosome contour according to the number of contour pixels and the contour roundness, the number of contour pixels being the number of pixels on the chromosome contour, and the contour roundness being the roundness of the chromosome contour, wherein the calculation of the contour roundness comprises: calculating the average distance of each pixel on the chromosome contour to the region center, the region center being the centroid of the chromosome, calculating the deviation between the distance of each pixel on the chromosome contour to the region center and the average distance, and determining the contour roundness of the chromosome contour according to the average distance and the deviation; and extracting the skeleton line of the chromosome contour retained after the filtering. A coordinate smoothing module is configured to determine the average shortest distance of each pixel on the chromosome contour retained after the filtering to the skeleton line, remove the chromosome contour with an average distance less than a preset distance threshold from the chromosome contour retained after the filtering, and perform smoothing processing on the position coordinates of each pixel on the chromosome contour retained after the removal according to a sliding mean filter to obtain a smoothed chromosome contour. The type identification module is further configured to determine the type of the chromosome according to the number of endpoints and the number of intersection points of the skeleton line of the chromosome contour retained after the filtering, the number of endpoints being the number of endpoints on the skeleton line, and the number of intersection points being the number of intersection points on the skeleton line. The type identification module is further configured to determine that the type of the chromosome is S-type if the number of endpoints is 2 and the number of intersection points is 0, determine that the type of the chromosome is T-type if the number of endpoints is 3 and the number of intersection points is 1, determine that the type of the chromosome is X-type if the number of endpoints is 4 and the number of intersection points is 1, and determine that the type of the chromosome is H-type if the number of endpoints is 4 and the number of intersection points is 2. An interaction module is configured to output the preliminary segmentation image as the final chromosome segmentation result to an interactive interface if the type of the chromosome is S-type, the interactive interface further displaying the preliminary segmentation image of the chromosome and the type of the chromosome, output a plurality of candidate overlapping segmentation schemes dynamically generated for the T-type to the interactive interface for user selection if the type of the chromosome is T-type, output a plurality of candidate overlapping segmentation schemes dynamically generated for the X-type to the interactive interface for user selection if the type of the chromosome is X-type, output a plurality of candidate overlapping segmentation schemes dynamically generated for the H-type to the interactive interface for user selection if the type of the chromosome is H-type, and obtain a target overlapping segmentation scheme specified by a user from the plurality of candidate overlapping segmentation schemes based on the interactive interface. The segmentation module is configured to perform the target overlap segmentation scheme specified by the user to segment the overlapped chromosomes in the preliminary segmentation image into single chromosomes.

5. The apparatus of claim 4, wherein, The segmentation module is further configured to convert an initial chromosome image into a grayscale image, perform adaptive histogram equalization processing on the grayscale image, and perform binary segmentation processing on the image processed by the adaptive histogram equalization to obtain a preliminary segmentation image of the chromosomes.

6. The apparatus of claim 4, wherein, The interaction module is further configured to extract convex points and concave points on the chromosome contour, and generate at least one candidate overlap segmentation scheme corresponding to the type according to the convex points and the concave points. 7.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-6 when the computer program is executed by the processor. The processor, when executing the computer program, implements the steps of the method in any one of claims 1 to 3.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the method in any one of claims 1 to 3.

Images