Intelligent processing method of fluorescence in-situ hybridization image
By intelligently processing fluorescence in situ hybridization images, including screening overlapping fusion regions, analyzing cell connectivity domain characteristics and classifying lymphoma, the problem of reduced classification accuracy of lymphoma caused by cell overlap and color overlap is solved, and higher classification accuracy is achieved.
Patent Information
- Application Number
- CN202510669396.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-23
- Publication Date
- 2025-06-20
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Cell overlap and color overlap in fluorescence in situ hybridization images result in reduced accuracy in tumor type classification of lymphoma cells.
By obtaining the fluorescent connectivity domain in the cell connectivity domain in the FISH image, the overlapping fusion region is screened based on the color distribution, the overall characteristic value is obtained, the translocation and overlapping cell connectivity domain are screened, the edge smoothness and fluorescence region distribution regularity are analyzed, the proliferating cell connectivity domain is distinguished, and lymphoma is intelligently classified based on this information.
It improves the accuracy of tumor type classification of lymphoma cells and avoids misjudgment of cell division and translocation by overlapping information.
Smart Images
Figure CN120182968A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image processing, and particularly relates to an intelligent processing method for fluorescence in situ hybridization images. Background Art
[0002] Fluorescence in situ hybridization (FISH) is a newly emerging molecular cytogenetic technology. Currently, this technology has been widely applied in many fields such as the research on the genomic structure of animals and plants, the analysis of fine chromosomal structure variations, virus infection analysis, human prenatal diagnosis, tumor genetics, and genomic evolution research. The basic principle of FISH is to use known labeled single-stranded nucleic acids as probes, and according to the principle of base complementarity, specifically bind with unknown single-stranded nucleic acids in the material to be detected to form detectable hybrid double-stranded nucleic acids. Since DNA molecules are linearly arranged along the longitudinal axis of chromosomes, probes can directly hybridize with chromosomes to localize specific genes on chromosomes. Various primary and secondary non-random clonal cytogenetic abnormalities can exist in lymphoma, including translocations, inversions, amplifications, deletions, aneuploidy, etc. These aberrations can be confirmed as useful tools for lymphoma diagnosis, so they can be detected by FISH technology.
[0003] By performing fluorescence in situ hybridization on lymphoma cells to obtain their FISH images, and through segmenting and extracting the fluorescent regions in the FISH images, the distribution of the fluorescent regions and the proliferation situation in lymphoma cells can be obtained, and then the tumor type of the current lymphoma cells can be determined; when judging the proliferation of lymphoma cells during cell division, due to cell overlap in the images, it may be misjudged as cell proliferation, and cell overlap may also cause a certain degree of color overlap of the chromosomes of two cells, affecting the recognition of tumor cells, thereby reducing the accuracy of classifying the tumor type of lymphoma cells. Summary of the Invention
[0004] To solve the above problems, the present invention provides an intelligent processing method for fluorescence in situ hybridization images, and the method includes: Obtain the FISH image of lymphoma; Obtain a number of fluorescent connected regions within a number of cell connected regions in the FISH image; screen the fluorescent connected regions based on the color distribution of the fluorescent connected regions to obtain an overlapping fusion region; obtain the overall characteristic value of the overlapping fusion region according to the gray value and the position distribution of the edge pixel points within the overlapping fusion region; screen all cell connected regions through the overall characteristic value to obtain translocation cell connected regions and overlapping cell connected regions; Obtain the edge smoothness of the overlapping cell connected region according to the position distribution of the edge pixel points of the overlapping cell connected region; obtain the fluorescence region distribution regularity of the overlapping cell connected region according to the edge smoothness and the distribution of the fluorescence connected regions within the overlapping cell connected region; distinguish all the overlapping cell connected regions according to the fluorescence region distribution regularity to obtain the proliferating cell connected regions. Intelligently classify lymphoma based on the numbers of the proliferating cell connected regions and the translocated cell connected regions.
[0005] Preferably, the method for screening the fluorescence connected regions according to the color distribution of the fluorescence connected regions to obtain the overlapping fusion regions specifically includes: For any cell connected region, use the Otsu threshold segmentation algorithm to segment the any cell connected region in the three color channels of R, G, and B to obtain three segmentation result images; the segmentation result images are binary images; if the gray values of any pixel point in the any cell connected region mapped in the three segmentation result images are all 1, then record any pixel point in the any cell connected region as a target pixel point, and regard the connected region formed by adjacent target pixel points as the overlapping fusion region.
[0006] Preferably, the method for obtaining the overall feature value of the overlapping fusion region according to the gray value and the position distribution of the edge pixel points in the overlapping fusion region specifically includes: Obtain the gray feature value of the overlapping fusion region according to the gray value in the overlapping fusion region; Let the sum of the gradient values of all the edge pixel points in the th overlapping fusion region be recorded as the edge feature value of the th overlapping fusion region; let the normalized value of the product of the gray feature value of the th overlapping fusion region and the edge feature value of the th overlapping fusion region be used as the overall feature value of the th overlapping fusion region.
[0007] Preferably, the method for obtaining the gray feature value of the overlapping fusion region according to the gray value in the overlapping fusion region specifically includes: Let the mean value of the gray values of all the pixel points in the th overlapping fusion region be recorded as the gray mean value in the th overlapping fusion region; let the inverse proportional normalized value of the absolute value of the difference between the mean value of the gray mean values of all the overlapping fusion regions and the gray mean value in the th overlapping fusion region be recorded as the gray feature value of the th overlapping fusion region.
[0008] Preferably, the method for screening all cell connected regions through the overall eigenvalue to obtain the translocation cell connected region and the overlapping cell connected region specifically includes: Preset a threshold parameter , for any cell connected region in the FISH image of lymphoma, if the overall eigenvalue of any overlapping fusion region in the any cell connected region is greater than or equal to the threshold parameter , record the any cell connected region as the translocation cell connected region; if the overall eigenvalues of all overlapping fusion regions in the any cell connected region are less than the threshold parameter , record the any cell connected region as the overlapping cell connected region.
[0009] Preferably, the method for obtaining the edge smoothness of the overlapping cell connected region according to the position distribution of the edge pixel points of the overlapping cell connected region specifically includes: Perform curve fitting on all edge pixel points of the th overlapping cell connected region by the least square method to obtain the edge curve of the th overlapping cell connected region; take the mean value of the curvatures of all edge pixel points on the edge curve of the th overlapping cell connected region as the edge smoothness of the th overlapping cell connected region.
[0010] Preferably, the method for obtaining the fluorescence region distribution regularity of the overlapping cell connected region according to the edge smoothness and the distribution of fluorescence connected regions in the overlapping cell connected region specifically includes: Obtain the splitting eigenvalue of the fluorescence connected region according to the distribution of the fluorescence connected regions in the overlapping cell connected region; Take the normalized value of the product of the edge smoothness of the th overlapping cell connected region, the sum of the number of all fluorescence connected regions and the cumulative sum of the splitting eigenvalues of all fluorescence connected regions in the th overlapping cell connected region, as the fluorescence region distribution regularity of the th overlapping cell connected region.
[0011] Preferably, the method for obtaining the splitting eigenvalue of the fluorescence connected region according to the distribution of the fluorescence connected regions in the overlapping cell connected region specifically includes: Take the minimum value of the distance between the th fluorescence connected region and all other fluorescence connected regions in the th overlapping cell connected region as the target distance value of the th fluorescence connected region; take the target distance value of the th fluorescence connected region and the The absolute value of the difference between the mean of the target distance values of all fluorescent connected regions within an overlapping cell connected region is denoted as the distance difference value of the th fluorescent connected region; Take the reciprocal of the minimum distance between the th fluorescent connected region and the edge of the th overlapping cell connected region, and denote it as the splitting weight factor of the th fluorescent connected region; Multiply the distance difference value of the th fluorescent connected region by the splitting weight factor, and denote it as the splitting eigenvalue of the th fluorescent connected region.
[0012] Preferably, the method for distinguishing all overlapping cell connected regions according to the regularity of the fluorescence region distribution to obtain the proliferating cell connected region specifically includes: If the fluorescence region distribution regularity of the th overlapping cell connected region is less than the threshold parameter , denote the th overlapping cell connected region as a proliferating cell connected region.
[0013] Preferably, the method for intelligently classifying lymphoma based on the number of proliferating cell connected regions and translocation cell connected regions specifically includes: Preset three parameters , , , where ; Among all cell connected regions in the FISH image of lymphoma, denote all cell connected regions except the translocation cell connected region and the proliferating cell connected region as normal cell connected regions; Denote the ratio of the number of normal cell connected regions in the FISH image of lymphoma to the number of all cell connected regions in the FISH image of lymphoma as the first ratio; denote the ratio of the number of translocation cell connected regions in the FISH image of lymphoma to the number of all cell connected regions in the FISH image of lymphoma as the second ratio; denote the ratio of the number of proliferating cell connected regions in the FISH image of lymphoma to the number of all cell connected regions in the FISH image of lymphoma as the third ratio; If the first ratio is greater than or equal to , and both the second ratio and the third ratio are less than , denote the lymphoma as an early-stage tumor; if the first ratio is greater than or equal to , less than , and both the second ratio and the third ratio are greater than or equal to , less than , denote the lymphoma as a mid-stage tumor; if the first ratio is less than , and both the second ratio and the third ratio are greater than or equal to , the lymphoma is recorded as a late-stage tumor.
[0014] The beneficial effects of the technical solution of the present invention are as follows: The present invention screens the fluorescence connected regions based on the color distribution of the fluorescence connected regions to obtain the overlapping fusion regions; screens all cell connected regions through the overall eigenvalue to obtain the translocation cell connected regions and the overlapping cell connected regions; obtains the fluorescence region distribution regularity of the overlapping cell connected regions according to the edge smoothness and the distribution of the fluorescence connected regions within the overlapping cell connected regions; distinguishes all overlapping cell connected regions according to the fluorescence region distribution regularity to obtain the proliferating cell connected regions; intelligently classifies lymphoma based on the number of proliferating cell connected regions and translocation cell connected regions; avoids misjudgment of cell division and cell translocation caused by overlapping information, thereby improving the accuracy of classifying lymphoma cell tumor types. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.
[0016] Figure 1 is a flowchart of the steps of an intelligent processing method for a fluorescence in situ hybridization image of the present invention; Figure 2 is a flowchart of the characteristic relationship of an intelligent processing method for a fluorescence in situ hybridization image of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0017] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will, in combination with the drawings and preferred embodiments, describe in detail the specific implementation manner, structure, characteristics and effects of an intelligent processing method for a fluorescence in situ hybridization image proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.
[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.
[0019] The following will specifically describe the specific solution of an intelligent processing method for a fluorescence in situ hybridization image provided by the present invention in conjunction with the drawings.
[0020] Please refer to Figure 1 , which shows a flowchart of the steps of an intelligent processing method for a fluorescence in situ hybridization image provided by an embodiment of the present invention. The method includes the following steps: Step S001: Obtain the FISH image of lymphoma.
[0021] Specifically, first, it is necessary to collect the FISH image of lymphoma. The specific process is as follows: Inoculate lymphoma cells into a culture medium and place them in an incubator at 37°C, providing appropriate gas conditions and ; use formaldehyde (4%) as a fixative. After washing the lymphoma cells on the culture plate with PBS, add the fixative and fix them at room temperature for 10 - 15 minutes, then wash them with PBS multiple times to remove the fixative; use a permeabilizing agent such as Triton X - 100 (0.1% - 0.5%) to treat the cell membrane so that the FISH probe can enter the cell; after permeabilization, wash with PBS to remove the permeabilizing agent; select a specific FISH probe according to the gene region to be labeled, mix the fluorescently labeled probe with the hybridization buffer, add it to the treated lymphoma cells, and then heat to an appropriate temperature (75°C) and maintain for a certain time (2 hours to overnight) to allow the probe to hybridize with the target DNA sequence; after hybridization, wash strictly with a solution such as a low - salt buffer (SSC) to remove the unbound probe; use a mounting medium with an anti - fluorescence quencher to mount the sample; use a fluorescence microscope to image the fixed lymphoma cells to obtain the FISH image of lymphoma.
[0022] So far, the FISH image of lymphoma is obtained through the above method.
[0023] Step S002: Obtain several fluorescent connected domains within several cell connected domains in the FISH image; screen the fluorescent connected domains based on the color distribution of the fluorescent connected domains to obtain the overlapping fusion region; obtain the overall characteristic value of the overlapping fusion region according to the gray - scale value and the position distribution of the edge pixel points within the overlapping fusion region; screen all cell connected domains through the overall characteristic value to obtain the translocation cell connected domain and the overlapping cell connected domain.
[0024] It should be noted that by obtaining the fluorescent regions within the cell regions in the FISH images of lymphoma, and based on the distribution of the fluorescent regions, it is possible to determine whether the corresponding genes exist in the current cell region. At the same time, when there are overlaps in the fragments on the chromosomes that need to be labeled, the color of the overlapping region will change. For example, a probe labeled with green fluorescence can label gene A, and a probe labeled with red fluorescence labels gene B. The overlapping region will appear yellow, indicating that these two genes have an intersection at the same position; therefore, there may be certain differences in the color overlap of the fluorescent regions, and then the RGB three channels can be used for analysis to screen out cells with chromosomal translocations and cells with chromosomal overlaps; at the same time, early screening of tumor cells can be carried out by analyzing the distribution of the fluorescent regions in the cells and the number of fluorescent regions.
[0025] Preferably, in some implementation manners of the embodiments of the present invention, by analyzing the FISH images of lymphoma, it can be determined that there are three distributions of the gray values in the FISH images: background region, cell region, and fluorescent region. The specific method for obtaining several fluorescent connected regions within several cell connected regions in the FISH image is as follows: Use the Otsu threshold algorithm to segment the FISH image to obtain several cell connected regions in the FISH image, and several fluorescent connected regions within each cell connected region.
[0026] Among them, the Otsu threshold algorithm is a prior art, and no further elaboration will be made here in this embodiment.
[0027] It should be noted that when fluorescent probes are used to label the chromosomal fragments of lymphoma, different probes have different fluorescence performances (different emission colors). When the chromosomal fragments at the same position are labeled by two probes, the color will change to form a new color, but its essence is the superposition of two colors. Therefore, the regions where the colors are superimposed can be separated through a single color channel; the purpose of distinguishing colors is to solve whether the color overlap is caused by color change or normal cell variation.
[0028] Preferably, in some implementation manners of the embodiments of the present invention, based on the color distribution of the fluorescent connected regions, the fluorescent connected regions are screened, and the specific method for obtaining the overlapping fusion region is as follows: For any cell connected region, use the Otsu threshold segmentation algorithm to segment the any cell connected region in the three color channels of R, G, and B to obtain three segmentation result images; the segmentation result images are binary images; if the gray value of any pixel point in the any cell connected region mapped in the three segmentation result images is 1, then the any pixel point in the any cell connected region is denoted as a target pixel point, and the connected region formed by adjacent target pixel points is used as the overlapping fusion region.
[0029] Among them, the method for obtaining the gray value of a pixel point in the segmentation result map is as follows: the gray value of a pixel point whose R color channel value is greater than or equal to the Otsu threshold is recorded as 1, and the gray value of a pixel point whose R color channel value is less than the Otsu threshold is recorded as 0, so as to obtain the segmentation result map of the R color channel. In this way, the segmentation result maps of the G and B color channels are obtained respectively.
[0030] Preferably, in some implementation manners of the embodiments of the present invention, when abnormal color expression is caused by cell overlap, it needs to pass through a certain number of cells to perform color fusion, so the fusion effect is not ideal. It is manifested that the edge of the overlapping fusion area is not clear, and the gray value distribution of the overlapping fusion color is relatively abnormal. Then, according to the gray value and the position distribution of the edge pixel points in the overlapping fusion area, the specific method for obtaining the overall feature value of each overlapping fusion area is as follows: Take the average value of the gray values of all pixel points in the th overlapping fusion area as the gray average value in the th overlapping fusion area; take the inverse proportional normalization value of the absolute value of the difference between the average value of the gray average values in all overlapping fusion areas and the gray average value in the th overlapping fusion area as the gray feature value of the th overlapping fusion area; Take the sum of the gradient values of all edge pixel points in the th overlapping fusion area as the edge feature value of the th overlapping fusion area; Take the normalized value of the product of the gray feature value of the th overlapping fusion area and the edge feature value of the th overlapping fusion area as the overall feature value of the th overlapping fusion area; In the formula, represents the overall feature value of the th overlapping fusion area; represents the average value of the gray values of all pixel points in the th overlapping fusion area; represents the average value of the gray average values in all overlapping fusion areas; represents the gradient value of the th edge pixel point in the th overlapping fusion area; represents taking the absolute value; represents the linear normalization function; Denote the exponential function with the natural constant as the base. In the embodiment, The model is used to present the inverse proportional relationship and normalization processing. is the input of the model. The implementer can select the inverse proportional function and normalization function according to the actual situation. It should be noted that differentiating the color change caused by overlapping and the color change caused by gene rearrangement can be used to distinguish subsequent cell types; since the above operation only separates the cells with color fusion regions, there may still be proliferating cells with overlapping fusion regions but no mutual interference of chromosomes. Therefore, it is necessary to subsequently judge the distribution of fluorescence connected regions in the cell connected regions.
[0031] Preferably, in some implementation manners of the embodiment of the present invention, the specific method for screening all cell connected regions through the overall eigenvalue to obtain the translocation cell connected region and the overlapping cell connected region is as follows: Preset a threshold parameter , where in this embodiment, is taken as an example for description. This embodiment does not make specific limitations, where is determined according to the specific implementation situation; For any cell connected region in the FISH image of lymphoma, if the overall eigenvalue of any overlapping fusion region in the any cell connected region is greater than or equal to the threshold parameter , record the any cell connected region as the translocation cell connected region; if the overall eigenvalue of all overlapping fusion regions in the any cell connected region is less than the threshold parameter , record the any cell connected region as the overlapping cell connected region.
[0032] So far, the translocation cell connected region and the overlapping cell connected region in the FISH image of lymphoma are obtained through the above method.
[0033] Step S003: Obtain the edge smoothness of the overlapping cell connected region according to the position distribution of the edge pixel points of the overlapping cell connected region; obtain the fluorescence region distribution regularity of the overlapping cell connected region according to the edge smoothness and the distribution of the fluorescence connected regions in the overlapping cell connected region; distinguish all overlapping cell connected regions according to the fluorescence region distribution regularity to obtain the proliferating cell connected region.
[0034] It should be noted that when determining whether the overlapping fusion region within the overlapping cell connected component is an overlap caused by chromosomal translocation or an overlap caused by cell proliferation, the distribution change of the fluorescence connected component in the overlapping cell connected component needs to be considered; during cell division, chromosomes will separate and arrange in a specific morphology, and at this time, gene or chromosome signals will form a linear distribution or an ordered arrangement, that is, the arrangement of chromosome signals presents a linear, strip-shaped or V-shaped, and these characteristics are very obvious in the image; if the overlapping fusion region is caused by chromosomal translocation, then the chromosome signals in the overlapping cell connected component are more chaotic and difficult to distinguish.
[0035] Preferably, in some implementation manners of the embodiments of the present invention, when the overlapping cell connected component is in the process of division, its cell connected component is not composed of the same cell. Therefore, by analyzing the edge situation of the overlapping cell connected component, if the overlapping fusion region inside it is caused by chromosomal translocation, its edge curvature is relatively obvious; if the overlapping fusion region inside it is caused by cell proliferation, its edge curvature is relatively smooth during cell division; then, the specific method for obtaining the edge smoothness of the overlapping cell connected component according to the position distribution of the edge pixel points of the overlapping cell connected component is as follows: Using the least squares method to perform curve fitting on the pixel coordinates of all edge pixel points of the th overlapping cell connected component, to obtain the edge curve of the th overlapping cell connected component; taking the mean value of the curvatures of all edge pixel points on the edge curve of the th overlapping cell connected component as the edge smoothness of the th overlapping cell connected component.
[0036] Among them, the least squares method is a prior art, and it will not be elaborated here in this embodiment.
[0037] It should be noted that the number of fluorescence connected components in a normal cell is generally relatively fixed; when the cell proliferates, the number of fluorescence connected components shown in the cell generally increases; the edge smoothness represents the distribution of the cell edge and only assists in judging the overlapping situation. When the chromosome or gene region is in different stages of division, its chromosome also appears relatively chaotic, but there will be paired fluorescence connected components that are relatively close to each other; and generally, chromosomes are in the cell nucleus, so if there are fluorescence connected components that are too close to the cell edge, they may be the fluorescence connected components during cell proliferation and division.
[0038] Preferably, in some implementation manners of the embodiments of the present invention, the specific method for obtaining the fluorescence region distribution regularity of the overlapping cell connected component according to the edge smoothness and the distribution of the fluorescence connected components in the overlapping cell connected component is as follows: Taking the The minimum distance between the th fluorescence connected domain in the overlapping cell connected domain and all other fluorescence connected domains is denoted as the target distance value of the th fluorescence connected domain; The absolute value of the difference between the target distance value of the th fluorescence connected domain and the average value of the target distance values of all fluorescence connected domains in the th overlapping cell connected domain is denoted as the distance difference value of the th fluorescence connected domain; The reciprocal of the minimum distance between the th fluorescence connected domain and the edge of the th overlapping cell connected domain is denoted as the splitting weight factor of the th fluorescence connected domain; The product of the distance difference value of the th fluorescence connected domain and the splitting weight factor is denoted as the splitting eigenvalue of the th fluorescence connected domain; The normalized value of the product of the edge smoothness of the th overlapping cell connected domain, the number of all fluorescence connected domains in the th overlapping cell connected domain, and the sum of the splitting eigenvalues of all fluorescence connected domains is used as the fluorescence region distribution regularity of the th overlapping cell connected domain; The specific formula is: In the formula, represents the fluorescence region distribution regularity of the th overlapping cell connected domain; represents the edge smoothness of the th overlapping cell connected domain; represents the number of all fluorescence connected domains in the th overlapping cell connected domain; represents the minimum distance between the th fluorescence connected domain and the edge of the th overlapping cell connected domain in the th overlapping cell connected domain; represents the minimum distance between the th fluorescence connected domain and all other fluorescence connected domains in the th overlapping cell connected domain; represents the average value of the target distance values of all fluorescence connected domains in the th overlapping cell connected domain; represents taking the absolute value; represents the linear normalization function.
[0039] Preferably, in some implementation manners of the embodiments of the present invention, the specific method for distinguishing all overlapping cell connected domains according to the regularity of the fluorescence region distribution and obtaining the translocated cell connected domain and the proliferating cell connected domain is as follows: If the fluorescence region distribution regularity of the th overlapping cell connected domain is greater than or equal to the threshold parameter , the th overlapping cell connected domain is denoted as the translocated cell connected domain; if the fluorescence region distribution regularity of the th overlapping cell connected domain is less than the threshold parameter
[0040] , the
[0041] th overlapping cell connected domain is denoted as the proliferating cell connected domain.
[0042] Preferably, in some implementation manners of the embodiments of the present invention, the specific method for classifying the malignancy degree of lymphoma based on the number of proliferating cell connected domains and translocated cell connected domains is as follows: Preset three parameters , , , where in this embodiment, , , is taken as an example for description, and this embodiment does not make specific limitations, where is determined according to the specific implementation situation; In all cell connected domains in the FISH image of lymphoma, all cell connected domains except the translocated cell connected domain and the proliferating cell connected domain are denoted as normal cell connected domains; The ratio of the number of normal cell connected domains in the FISH image of lymphoma to the number of all cell connected domains in the FISH image of lymphoma is denoted as the first ratio; the ratio of the number of translocated cell connected domains in the FISH image of lymphoma to the number of all cell connected domains in the FISH image of lymphoma is denoted as the second ratio; the ratio of the number of proliferating cell connected domains in the FISH image of lymphoma to the number of all cell connected domains in the FISH image of lymphoma is denoted as the third ratio; If the first ratio is greater than or equal to , and both the second ratio and the third ratio are less than , the lymphoma is denoted as an early-stage tumor, with a lower malignancy degree, slow proliferation, and few translocations; if the first ratio is greater than or equal to , less than , and both the second ratio and the third ratio are greater than or equal to , less than , the lymphoma is recorded as a medium-stage tumor, with a medium degree of tumor malignancy, certain gene translocations and active proliferation, and relatively controllable treatment effects; if the first ratio is less than , and both the second ratio and the third ratio are greater than or equal to , the lymphoma is recorded as a late-stage tumor, with a high degree of tumor malignancy, a large number of translocated cells and proliferating cells, and indicates rapid tumor growth and high malignancy, and the treatment is difficult.
[0043] Please refer to Figure 2 , which shows a characteristic relationship flowchart of an intelligent processing method for fluorescence in situ hybridization images; So far, this embodiment is completed.
[0044] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included in the protection scope of the present invention.
Claims
1. An intelligent processing method for fluorescence in situ hybridization images, characterized in that, The method includes the following steps: Obtain the FISH image of lymphoma; Obtain a number of fluorescent connected regions within a number of cell connected regions in the FISH image; screen the fluorescent connected regions based on the color distribution of the fluorescent connected regions to obtain an overlapping fusion region; obtain the overall eigenvalue of the overlapping fusion region according to the gray value and the position distribution of the edge pixel points within the overlapping fusion region; screen all cell connected regions through the overall eigenvalue to obtain translocation cell connected regions and overlapping cell connected regions; Obtain the edge smoothness of the overlapping cell connected region according to the position distribution of the edge pixel points of the overlapping cell connected region; obtain the fluorescence region distribution regularity of the overlapping cell connected region according to the edge smoothness and the distribution of the fluorescent connected regions within the overlapping cell connected region; distinguish all overlapping cell connected regions according to the fluorescence region distribution regularity to obtain proliferating cell connected regions; Intelligently classify lymphoma based on the number of proliferating cell connected regions and translocation cell connected regions.
2. The intelligent processing method for fluorescence in situ hybridization images according to claim 1, characterized in that, The screening of the fluorescent connected regions based on the color distribution of the fluorescent connected regions to obtain an overlapping fusion region includes the following specific method: For any cell connected region, use the Otsu threshold segmentation algorithm to segment the any cell connected region in the three color channels of R, G, and B to obtain three segmentation result maps; the segmentation result maps are binary images; if the gray value of any pixel point in the any cell connected region mapped in the three segmentation result maps is 1, then the any pixel point in the any cell connected region is denoted as a target pixel point, and the connected region formed by adjacent target pixel points is used as the overlapping fusion region.
3. The intelligent processing method for fluorescence in situ hybridization images according to claim 1, characterized in that, The obtaining of the overall eigenvalue of the overlapping fusion region according to the gray value and the position distribution of the edge pixel points within the overlapping fusion region includes the following specific method: Obtain the gray feature value of the overlapping fusion region according to the gray value within the overlapping fusion region; Denote the sum of the gradient values of all edge pixels in the -th overlapping fusion region as the edge feature value of the -th overlapping fusion region; take the normalized value of the product of the grayscale feature value of the -th overlapping fusion region and the edge feature value of the -th overlapping fusion region as the overall feature value of the -th overlapping fusion region.
4. The intelligent processing method for fluorescence in situ hybridization images according to claim 3, characterized in that, The obtaining of the gray feature value of the overlapping fusion region according to the gray value within the overlapping fusion region includes the following specific method: Denote the mean of the grayscale values of all pixel points within the th overlapping fusion region as the grayscale mean within the th overlapping fusion region; Denote the inverse proportional normalization value of the absolute value of the difference between the mean of the grayscale means within all overlapping fusion regions and the grayscale mean within the th overlapping fusion region as the grayscale feature value of the th overlapping fusion region.
5. The intelligent processing method for fluorescence in situ hybridization images according to claim 1, characterized in that, The screening of all cell connected regions through the overall eigenvalue to obtain translocation cell connected regions and overlapping cell connected regions includes the following specific method: Preset a threshold parameter , for any cell connected region in the FISH image of lymphoma, if the overall eigenvalue of any overlapping fusion region in the any cell connected region is greater than or equal to the threshold parameter , mark the any cell connected region as a translocation cell connected region; if the overall eigenvalue of all overlapping fusion regions in the any cell connected region is less than the threshold parameter , mark the any cell connected region as an overlapping cell connected region.
6. The intelligent processing method for fluorescence in situ hybridization images according to claim 1, characterized in that, The obtaining of the edge smoothness of the overlapping cell connected region according to the position distribution of the edge pixel points of the overlapping cell connected region includes the following specific method: By using the least squares method, curve fitting is performed on all the edge pixel points of the th overlapping cell connected domain to obtain the edge curve of the th overlapping cell connected domain; the mean value of the curvatures of all the edge pixel points on the edge curve of the th overlapping cell connected domain is taken as the edge smoothness of the th overlapping cell connected domain.
7. The intelligent processing method for fluorescence in situ hybridization images according to claim 1, characterized in that, The obtaining of the fluorescence region distribution regularity of the overlapping cell connected region according to the edge smoothness and the distribution of the fluorescent connected regions within the overlapping cell connected region includes the following specific method: Obtain the splitting eigenvalue of the fluorescent connected region according to the distribution of the fluorescent connected regions within the overlapping cell connected region; Take the normalization value of the product of the edge smoothness of the th overlapping cell connected domain, the number of all fluorescence connected domains within the th overlapping cell connected domain, and the cumulative sum of the splitting eigenvalue of all fluorescence connected domains, as the fluorescence region distribution regularity of the th overlapping cell connected domain.
8. The intelligent processing method for fluorescence in situ hybridization images according to claim 7, characterized in that, The obtaining of the splitting eigenvalue of the fluorescent connected region according to the distribution of the fluorescent connected regions within the overlapping cell connected region includes the following specific method: Denote the minimum value of the distances between the th fluorescence connected region within the th overlapping cell connected region and all other fluorescence connected regions as the target distance value of the th fluorescence connected region; Denote the absolute value of the difference between the target distance value of the th fluorescence connected region and the mean value of the target distance values of all fluorescence connected regions within the th overlapping cell connected region as the distance difference value of the th fluorescence connected region; Denote the reciprocal of the minimum distance between the th fluorescence connected domain and the edge of the th overlapping cell connected domain as the splitting weight factor of the th fluorescence connected domain; Multiply the distance difference value of the th fluorescence connected region by the splitting weight factor, and denote it as the splitting eigenvalue of the th fluorescence connected region.
9. The intelligent processing method for fluorescence in situ hybridization images according to claim 5, characterized in that, The distinguishing of all overlapping cell connected regions according to the fluorescence region distribution regularity to obtain proliferating cell connected regions includes the following specific method: If the fluorescence region distribution regularity of the th overlapping cell connected domain is less than the threshold parameter , the th overlapping cell connected domain is denoted as the proliferating cell connected domain.
10. The intelligent processing method for fluorescence in situ hybridization images according to claim 1, characterized in that,The intelligent classification of lymphoma based on the number of proliferating cell connected regions and translocation cell connected regions includes the following specific method: Preset three parameters , , , where ; Among all the cell connected regions in the FISH image of lymphoma, all the cell connected regions except the translocation cell connected region and the proliferation cell connected region are denoted as normal cell connected regions; The ratio of the number of normal cell connected regions in the FISH image of lymphoma to the number of all cell connected regions in the FISH image of lymphoma is denoted as the first ratio; the ratio of the number of translocation cell connected regions in the FISH image of lymphoma to the number of all cell connected regions in the FISH image of lymphoma is denoted as the second ratio; the ratio of the number of proliferation cell connected regions in the FISH image of lymphoma to the number of all cell connected regions in the FISH image of lymphoma is denoted as the third ratio; If the first ratio is greater than or equal to , and both the second ratio and the third ratio are less than , the lymphoma is recorded as an early-stage tumor; if the first ratio is greater than or equal to , less than , and both the second ratio and the third ratio are greater than or equal to , less than , the lymphoma is recorded as a mid-stage tumor; if the first ratio is less than , and both the second ratio and the third ratio are greater than or equal to , the lymphoma is recorded as a late-stage tumor.