Method for automatically generating cross chromosome image and computer program product

Abstract

The invention discloses a method for automatically generating a cross chromosome image and a computer program product. The method comprises the following steps: S3, extracting a reference chromosome from a reference chromosome image; and S4, the extracted reference chromosomes are moved to the to-be-synthesized chromosome image and are intersected and overlapped with the to-be-synthesized chromosomes in the image for synthesis, and the intersected and overlapped reference chromosomes and the to-be-synthesized chromosomes are not the same chromosome. According to the method, a reference chromosome is extracted from a reference chromosome image in the step S3, and then the reference chromosome is put into a to-be-synthesized chromosome image to be subjected to cross synthesis with the to-be-synthesized chromosome in the step S4, so that a cross chromosome image is obtained. When the method is used for many times, different chromosomes to be synthesized and reference chromosomes are selected for each time, and a large number of chromosome image samples with different chromosome crossing conditions can be generated for an artificial intelligence algorithm model to carry out chromosome crossing recognition analysis training.

Description

Technical Field

[0001] The present invention relates to the technical field of chromosome image processing, and in particular to a method and a computer program product for automatically generating a crossover chromosome image. Background Art

[0002] Chromosomes are carriers of genetic material in organisms. Abnormal numbers or structures of chromosomes can lead to genetic diseases. Chromosome karyotype analysis is of great significance in exploring the mechanisms of human genetic diseases and tumor pathology research. Chromosome karyotype analysis refers to the grouping, arrangement and pairing of chromosomes in chromosome images when cells are in metaphase of mitosis to generate a karyotype diagram. Chromosome karyotype analysis traditionally relies on manual processing by medical staff. With the development of computer automation and intelligent technology, the industry has proposed some solutions for using artificial intelligence algorithm models to perform chromosome karyotype analysis. For example, a large number of chromosome image annotated samples are used to train artificial intelligence algorithm models so that artificial intelligence algorithm models have corresponding recognition and analysis capabilities. Chromosome crossover and overlap may occur in chromosome images, and chromosome crossover and overlap may result in the failure to identify part of the information of one of the chromosomes during recognition. The artificial intelligence algorithm model is trained using chromosome image samples with chromosome crossover. The trained artificial intelligence algorithm model can usually only recognize and analyze the chromosome crossover that it has trained, and it is difficult to generalize to the crossover that has not been trained. Since there are few image samples of naturally formed chromosome crossovers in human cell images and the crossover positions are random, it is difficult for users to collect a large number of chromosome image samples with different chromosome crossover situations to train the artificial intelligence algorithm model. This has become a training bottleneck for the artificial intelligence algorithm model to perform chromosome crossover identification and analysis. Summary of the invention

[0003] The technical problem to be solved by the present invention is how to obtain a chromosome image with chromosome crossover and provide it to an artificial intelligence algorithm model for chromosome crossover recognition analysis training.

[0004] In order to solve the above technical problems, the present invention provides a method for automatically generating a crossover chromosome image, comprising the following steps:

[0005] S3, extracting a reference chromosome from the reference chromosome image;

[0006] S4. Move the extracted reference chromosome to the chromosome image to be synthesized, and cross-overlap with the chromosome to be synthesized in the image for synthesis, wherein the cross-overlapping reference chromosome and the chromosome to be synthesized are not the same chromosome.

[0007] Further: in S3, a plurality of reference chromosomes are specifically extracted; and in S4, the plurality of reference chromosomes extracted are specifically moved to the image of the chromosome to be synthesized so as to cross-overlap with the plurality of chromosomes to be synthesized in the image.

[0008] Further:

[0009] This includes the following steps performed before S4:

[0010] S1, determining a proposed crossover position of the chromosome to be synthesized according to a set of outline coordinate points of the chromosome to be synthesized in the chromosome image to be synthesized and a preset movement condition;

[0011] S2, based on the reference chromosome outline coordinate point set extracted from the reference chromosome image and the proposed crossing position, pre-calculate the overlap rate of the reference chromosome when it is moved to the chromosome image to be synthesized and crosses and overlaps with the chromosome to be synthesized;

[0012] The condition for executing S4 is that the overlap rate between each reference chromosome and the corresponding chromosome to be synthesized meets the preset overlap range. Specifically, S4 moves the extracted reference chromosome to the proposed crossover position in the image of the chromosome to be synthesized so as to crossover and overlap with the chromosome to be synthesized.

[0013] If the overlap rate between a reference chromosome and the corresponding chromosome to be synthesized does not meet the preset overlap range, the preset movement condition is adjusted and S1 and S2 are re-executed until the overlap rate between the reference chromosome and the corresponding chromosome to be synthesized meets the preset overlap range.

[0014] Furthermore, S1 specifically includes:

[0015] S11, obtaining the center coordinates of the chromosome to be synthesized according to the outline coordinate point set of the chromosome to be synthesized;

[0016] S12. Determine the proposed crossover position of the chromosome to be synthesized according to preset movement conditions, wherein the preset movement conditions include the movement distance of the center of the reference chromosome at the center position of the chromosome to be synthesized along the length direction of the chromosome to be synthesized.

[0017] Furthermore, S1 also includes:

[0018] S13, based on the outline coordinate point set of the reference chromosome and the above-mentioned quasi-crossing position, calculate whether the reference chromosome will exceed the boundary of the chromosome image to be synthesized when it moves to the quasi-crossing position. If so, adjust the preset movement condition and re-execute S12.

[0019] Further, S2 specifically calculates the ratio of the overlapping area between the reference chromosome and the chromosome to be synthesized to the reference chromosome area and / or the ratio to the chromosome area to be synthesized based on the outline coordinate point set of the reference chromosome and the above-mentioned simulated crossover position, and obtains the overlap rate based on the above-mentioned ratio.

[0020] Further:

[0021] S2 includes:

[0022] S21, converting the original contour coordinate point set of the reference chromosome into a new contour coordinate point set with the proposed crossover position as the new center position of the reference chromosome;

[0023] S22, generating a mask map of the chromosome to be synthesized and a mask map of the reference chromosome respectively according to the outline coordinate point set of the chromosome to be synthesized and the new outline coordinate point set of the reference chromosome;

[0024] S23, performing an erosion operation on the reference chromosome mask image to obtain a reference chromosome mask image after eliminating edge errors;

[0025] S24, calculating the ratio of the overlapping region between the reference chromosome and the chromosome to be synthesized to the reference chromosome region and / or the ratio to the chromosome region to be synthesized based on the reference chromosome mask map and the chromosome mask map after eliminating edge errors, and obtaining the overlap rate based on the above ratios;

[0026] S3 specifically, obtains a contour coordinate point set of the reference chromosome after eliminating edge errors based on the reference chromosome mask image after eliminating edge errors, converts the contour coordinate point set to a new contour coordinate point set at the original center position of the reference chromosome, and extracts the reference chromosome from the reference chromosome image based on the new contour coordinate point set.

[0027] Furthermore, S24 specifically obtains the chromosome overlapping region in the following manner: the coordinate points with chromosomes in the reference chromosome mask map after eliminating edge errors and in the chromosome mask map to be synthesized are recorded as 1, and the coordinate points without chromosomes are recorded as 0, and the reference chromosome mask map and the chromosome mask map to be synthesized are traversed to obtain all the coordinate points that are 1 in both the reference chromosome mask map and in the chromosome mask map to be synthesized, and the chromosome overlapping region is obtained accordingly.

[0028] Furthermore, in S4:

[0029] After the reference chromosome and the chromosome to be synthesized are cross-overlapped and synthesized, the smoothing filter method is used to eliminate the synthesis traces;

[0030] Alternatively, the synthesis may be performed using a generative adversarial network (GANs).

[0031] The present invention also provides a computer program product, comprising a computer program, which is executed by a processor to implement the method for automatically generating a crossover chromosome image as described above.

[0032] The method for automatically generating a crossover chromosome image of the present invention first extracts a reference chromosome from a reference chromosome image in step S3, and then in step S4, the reference chromosome is placed in the chromosome image to be synthesized and cross-synthesized with the chromosome to be synthesized to obtain a crossover chromosome image. The method is used multiple times, and different chromosomes to be synthesized and reference chromosomes are selected each time, so that a large number of chromosome image samples with different chromosome crossover situations can be generated for the artificial intelligence algorithm model to perform chromosome crossover recognition analysis training. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a flowchart of a first embodiment of a method for automatically generating a crossover chromosome image.

[0034] Figure 2 It is a flowchart of Embodiment 2 of the method for automatically generating a crossover chromosome image.

[0035] Figure 3 The method for automatically generating a cross chromosome image is as follows: Example 2 gives the length direction and angle α of the chromosome to be synthesized. oci Schematic diagram of geometry, where the rectangle is the minimum circumscribed rectangle of the chromosome to be synthesized, and the direction of the arrow is the length direction of the chromosome to be synthesized.

[0036] Figure 4 This is a simplified diagram of the operation process of Example 2 of the method for automatically generating a crossover chromosome image. DETAILED DESCRIPTION

[0037] The present invention is further described in detail below in conjunction with specific implementation methods.

[0038] Embodiment 1

[0039] The method for automatically generating a crossover chromosome image in this embodiment is programmed by computer code to form a computer program product, which is executed by a processor to achieve the following Figure 1 The method for automatically generating a cross-chromosome image is shown. The following takes a computer processor executing the method for automatically generating a cross-chromosome image as an example to illustrate the execution process of the method.

[0040] The user collects a large number of chromosome images in advance, annotates the chromosome contours of each chromosome image, and uses these annotated chromosome images to construct a chromosome image annotation data set, which includes multiple chromosome images, each of which is annotated with a set of contour coordinate points of each chromosome. The constructed chromosome image annotation data set is stored in the computer's storage device. If a cross-chromosome image needs to be generated, the computer obtains the chromosome image to be synthesized and the reference chromosome image from the chromosome image annotation data set. The chromosome image to be synthesized and the reference chromosome image can be the same chromosome image or two different chromosome images. The contour coordinate point sets of each chromosome in the image are annotated in the chromosome image to be synthesized and the reference chromosome image. The computer selects multiple chromosomes from the chromosome image to be synthesized as the chromosomes to be synthesized. i , select multiple chromosomes from the reference chromosome image as the above-mentioned chromosomes to be synthesized δ i The corresponding reference chromosome A chromosome to be synthesized δ i and a reference chromosome As a set of chromosomes to be crossed. If the chromosome image to be synthesized and the reference chromosome image are the same chromosome image, it is necessary to ensure that the selected chromosome to be synthesized δ i and its corresponding reference chromosome Not the same chromosome, that is, a set of chromosomes to be crossed over needs to be two different chromosomes.

[0041] S3. Extract reference chromosomes from the reference chromosome image.

[0042] The computer selects the chromosome to be synthesized. i and reference chromosome After that, multiple reference chromosomes are extracted from the reference chromosome image.

[0043] S4, moving the extracted reference chromosome to the chromosome image to be synthesized, and overlapping and synthesizing with the chromosome to be synthesized in the image, wherein the overlapping reference chromosome and the chromosome to be synthesized are i Not the same chromosome.

[0044] The computer extracts multiple reference chromosomes from the reference chromosome image After that, the extracted multiple reference chromosomes are Move to the chromosome image to be synthesized, and respectively interact with multiple chromosomes to be synthesized in the image δ i Cross-overlap synthesis, where the reference chromosome and chromosome δ to be synthesized iNot the same chromosome. In this way, the reference chromosome Put it into the chromosome image to be synthesized and the chromosome to be synthesized δ i Cross-synthesis to obtain a cross-chromosome image. This method is used multiple times, each time selecting a different chromosome image to be synthesized and a reference chromosome image and selecting a different chromosome to be synthesized and a reference chromosome from them, or each time selecting a different chromosome to be synthesized and a reference chromosome from the same chromosome image to be synthesized and a reference chromosome image, a large number of chromosome image samples with different chromosome crossover situations can be generated for the artificial intelligence algorithm model to perform chromosome crossover recognition analysis training.

[0045] Since the reference chromosome Put it into the chromosome image to be synthesized and the chromosome to be synthesized δ i Cross-synthesis will leave synthesis traces. In order to eliminate synthesis traces, in S4, and chromosome δ to be synthesized i After cross-overlap synthesis, a smoothing filter method is used to eliminate synthesis traces. For example, a Gaussian smoothing method or a wavelet filtering method is used to eliminate synthesis traces. In other embodiments, S4 can be changed to: Put it into the chromosome image to be synthesized and the chromosome to be synthesized δ i At the cross-overlapping position, the synthesis is not performed directly on the chromosome image to be synthesized, but each reference chromosome is They are placed on the newly created graph according to the cross-overlapping positions, and then the conditional adversarial generative network GANs (Generative Adversarial Networks) is used to synthesize the chromosome image to be synthesized and the new image with each reference chromosome placed, so that the synthesized edge will be smoother and more reasonable.

[0046] Embodiment 2

[0047] The method for automatically generating a crossover chromosome image in this embodiment is programmed by computer code to form a computer program product, which is executed by a processor to achieve the following Figure 2 The method for automatically generating a cross-chromosome image is shown. The following takes a computer processor executing the method for automatically generating a cross-chromosome image as an example to illustrate the execution process of the method.

[0048] The user collects a large number of chromosome images in advance, annotates the chromosome contours of each chromosome image, and uses these annotated chromosome images to construct a chromosome image annotation dataset, which includes multiple chromosome images, each of which is annotated with the contour coordinate point set of each chromosome. The constructed chromosome image annotation dataset is stored in the computer's storage device. If you need to generate a cross chromosome image, such as Figure 4 As shown, the computer obtains the chromosome image to be synthesized and the reference chromosome image from the chromosome image annotation data set. The chromosome image to be synthesized and the reference chromosome image can be the same chromosome image or two different chromosome images. The chromosome image to be synthesized and the reference chromosome image are both annotated with the contour coordinate point set of each chromosome in the image. The computer selects multiple chromosomes from the chromosome image to be synthesized as the chromosome to be synthesized δ i , select multiple chromosomes from the reference chromosome image as the above-mentioned chromosomes to be synthesized δ i The corresponding reference chromosome A chromosome to be synthesized δ i and a reference chromosome As a set of chromosomes to be crossed. If the chromosome image to be synthesized and the reference chromosome image are the same chromosome image, it is necessary to ensure that the selected chromosome to be synthesized δ i and its corresponding reference chromosome Not the same chromosome, that is, a set of chromosomes to be crossed over needs to be two different chromosomes.

[0049] S1. Determine a proposed crossover position of the chromosome to be synthesized according to a set of outline coordinate points of the chromosome to be synthesized in the chromosome image to be synthesized and preset movement conditions.

[0050] S1 specifically includes the following steps:

[0051] S11. Obtain the center coordinates of the chromosome to be synthesized according to the outline coordinate point set of the chromosome to be synthesized.

[0052] The computer based on the chromosome to be synthesized i The contour coordinate point set is obtained by using the minimum outer intercept algorithm in image processing technology to obtain the chromosome to be synthesized δ i The center coordinates of

[0053] S12. Determine the proposed crossover position of the chromosome to be synthesized according to preset movement conditions, wherein the preset movement conditions include the movement distance of the center of the reference chromosome at the center position of the chromosome to be synthesized along the length direction of the chromosome to be synthesized.

[0054] The user can set the preset movement conditions in the computer in advance, and the preset movement conditions include the reference chromosome The center is on the chromosome to be synthesized i The center position along the chromosome to be synthesized i The moving distance d in the length direction. The user, for example, pre-sets the range of the moving distance d in the computer as a set of preset moving conditions to be selected (each moving distance d value in the range is taken as a preset moving condition). The computer, for example, selects the value of the moving distance d from large to small, firstly selects a maximum moving distance d1 from the set of preset moving conditions to be selected as the preset moving condition, and then selects the maximum moving distance d1 and the chromosome to be synthesized according to the moving distance d1 and the chromosome to be synthesized δ. i The central position of the chromosome to be synthesized is used to determine the proposed crossover position. Specifically, i The center position along the chromosome to be synthesized i The position reached by moving the above-mentioned moving distance d1 in the length direction is the quasi-crossover position. i The center position coordinates (i.e. center coordinates) and the moving distance d1 to determine the coordinates of the proposed crossover position. Specifically, first calculate the distance along the chromosome to be synthesized δ i The length direction moves the above moving distance d1 on the coordinate system X-axis and Y-axis corresponding to the change Δx, Δy, and then on the chromosome to be synthesized δ i Center coordinates Adding the corresponding Δx and Δy to the base gives the coordinates P(x p ,y p ). To calculate Δx and Δy, we need to first obtain the chromosome to be synthesized δ i The angle α between the length direction and the horizontal line (X axis) oci (See Figure 3 ). Therefore, in this embodiment, the computer is based on the chromosome to be synthesized δ i The contour coordinate point set is obtained by using the minimum outer intercept algorithm in image processing technology to obtain the chromosome to be synthesized δ i Length direction (to be synthesized chromosome δ i The angle α between the minimum circumscribed rectangle of oci Preferably, the computer can be changed to obtain the chromosome to be synthesized in the above text i The center coordinates of At the same time, the above angle α is obtained oci .

[0055] The coordinates of the proposed intersection position P(x p ,y p ) is obtained according to the following formula:

[0056]

[0057] S13, based on the outline coordinate point set of the reference chromosome and the above-mentioned quasi-crossing position, calculate whether the reference chromosome will exceed the boundary of the chromosome image to be synthesized when it moves to the quasi-crossing position. If so, adjust the preset movement condition and re-execute S12.

[0058] Since the reference chromosome Translate to the center and align with the chromosome to be synthesized δ i After the center position, move along the moving distance to the proposed crossover position, referring to the chromosome The image may exceed the boundary of the chromosome to be synthesized, so the computer does not use the reference chromosome first. Move to the chromosome to be synthesized according to the proposed crossover position i Instead, we first verify the reference chromosome by simulation verification. Whether it will exceed the boundary of the chromosome image to be synthesized, the specific process is as follows: the computer The minimum outer intercept algorithm in image processing technology is used to obtain the width W, height H and center coordinates of the reference chromosome. The computer then uses the proposed cross position P(x p ,y p ) as the reference chromosome The center coordinates after the move, based on the reference chromosome The width W and height H of the reference chromosome are calculated Move to the proposed intersection position P(x p ,y p ) will exceed the boundary of the chromosome image to be synthesized. If it exceeds the boundary, the computer will reselect the value of the moving distance d from large to small, first select a second largest moving distance d2 from the set of preset moving conditions to be selected as the preset moving condition, and re-execute S12 to determine the new coordinates P(x p ,y p ) until the reference chromosome is verified Move to the quasi-crossover position without exceeding the boundary position of the chromosome image to be synthesized. If the image exceeds the boundary of the chromosome to be synthesized, S12 is not re-executed, but the reference chromosome is directly cut off. The part beyond the boundary of the chromosome image to be synthesized is the reference chromosome after removal As a reference chromosome for subsequent crossover synthesis

[0059] The computer determines the proposed crossover position of each chromosome to be synthesized according to S11 to S13 above.

[0060] S2. Based on the reference chromosome outline coordinate point set extracted from the reference chromosome image and the proposed crossing position, pre-calculate the overlap rate of the reference chromosome when it is moved to the chromosome image to be synthesized and crosses and overlaps with the chromosome to be synthesized.

[0061] If the chromosome cross-over overlap area is too small or too large, it is not conducive to the learning of the artificial intelligence algorithm model. Therefore, in order to obtain a better cross-chromosome image sample, this embodiment is designed to pre-simulate and verify the reference chromosome and chromosome δ to be synthesized i The overlap ratio of the two cross-overlapping is within an appropriate range before cross-synthesis is performed. The simulation verification process is as follows:

[0062] S2 includes:

[0063] S21, converting the original contour coordinate point set of the reference chromosome into a new contour coordinate point set with the proposed crossover position as the new center position of the reference chromosome.

[0064] Reference chromosome The contour coordinate point set in the reference chromosome image is its original contour coordinate point set, and the center position is In this embodiment, if crossover synthesis is performed, the reference chromosome Need to move to the center and the chromosome to be synthesized i Align the center position, and then follow the chromosome to be synthesized δ i The length direction of the moving distance is moved to the proposed intersection position P (x p ,y p ). That is, the reference chromosome The center position C j Finally, it moves to the quasi-crossing position P(x p ,y p ), that is, the center position coordinate becomes C j (x p ,y p ). Therefore, the computer uses the proposed crossover position as the reference chromosome New central location C j (x p ,y p ), the reference chromosome The original contour coordinate point set is converted into a chromosome with the quasi-crossover position as the reference New central location C j (x p ,y p ) under the new contour coordinate point set, that is, the reference chromosome is obtained Move to the quasi-crossing position P(x p,y p ) is the contour coordinate point set at .

[0065] S22, generating a mask map of the chromosome to be synthesized and a mask map of the reference chromosome respectively according to the outline coordinate point set of the chromosome to be synthesized and the new outline coordinate point set of the reference chromosome.

[0066] This embodiment uses a mask image to calculate the reference chromosome and chromosome δ to be synthesized i The overlap rate of the two crossovers, so the computer is based on the chromosome to be synthesized δ i The contour coordinate point set generates the chromosome mask map to be synthesized Then according to the reference chromosome The new contour coordinate point set generates the reference chromosome mask map

[0067] S23, performing an erosion operation on the reference chromosome mask image to obtain a reference chromosome mask image after eliminating edge errors.

[0068] Reference chromosomes annotated in the reference chromosome image There may be marginal errors, making the generated reference chromosome mask The reference chromosomes in the image also have edge errors. To reduce the impact of edge errors on overlap calculation, the computer performs masking on the reference chromosomes. Perform an erosion operation to obtain the reference chromosome mask after eliminating edge errors

[0069] S24. Calculate the ratio of the overlapping region between the reference chromosome and the chromosome to be synthesized to the reference chromosome region and / or the ratio to the chromosome region to be synthesized based on the reference chromosome mask image and the chromosome mask image after eliminating edge errors, and obtain the overlap rate based on the above ratios.

[0070] The computer obtains the reference chromosome mask after eliminating edge errors and the chromosome mask to be synthesized Afterwards, the overlap rate is calculated based on these two mask images. and chromosome δ to be synthesized i The ratio of the overlapping region to the region of the chromosome to be synthesized is used as the reference chromosome and chromosome δ to be synthesized i The overlap rate of the reference chromosome is obtained in the following manner in this embodiment. and chromosome δ to be synthesized i Overlapping area: Reference chromosome mask after eliminating edge errors Neutralize the chromosome mask to be synthesized The coordinate points with chromosomes are recorded as 1, and the coordinate points without chromosomes are recorded as 0. The reference chromosome mask map and the to-be-synthesized chromosome mask map are traversed to obtain all the coordinate points that are 1 in both the reference chromosome mask map and the to-be-synthesized chromosome mask map, which are expressed as Among them, W and H are the width and height of the chromosome mask to be synthesized, and the accumulated area of ​​these coordinate points is the reference chromosome and chromosome δ to be synthesized i Overlapping area. The chromosome region to be synthesized is the cumulative area of ​​all coordinate points of the chromosome mask map to be synthesized, expressed as The calculation formula is as follows:

[0071]

[0072] Other embodiments may be modified to use a reference chromosome and chromosome δ to be synthesized i The ratio of the overlapping region to the reference chromosome region is used as the reference chromosome and chromosome δ to be synthesized i The overlap ratio is calculated as follows:

[0073]

[0074] in, Reference chromosome and chromosome δ to be synthesized i Overlapping area, is the reference chromosome region, W and H are the width and height of the reference chromosome mask.

[0075] Other embodiments may also be modified to: first calculate the reference chromosome and chromosome δ to be synthesized i The ratio of the overlapping region to the region of the chromosome to be synthesized and the reference chromosome and chromosome δ to be synthesized i The ratio of the overlapping region to the reference chromosome region, and then the average of these two ratios is used as the reference chromosome and chromosome δ to be synthesized i The overlap rate is (p i +p j ) / 2.

[0076] The computer calculates each reference chromosome according to S21 to S24 above and the corresponding chromosome to be synthesized i After the overlap rate of each reference chromosome and the corresponding chromosome to be synthesized iWhether the overlap rate meets the preset overlap range. The preset overlap range is, for example, 10% to 50%. In other embodiments, the user can change the preset overlap range to other ranges according to actual needs, such as 10% to 30%, 30% to 50%, etc. and the corresponding chromosome to be synthesized i If the overlap rate of the reference chromosome meets the preset overlap range, the following S3 and S4 are executed. and the corresponding chromosome to be synthesized i If the overlap rate does not meet the preset overlap range, the preset movement condition is adjusted according to the above method, that is, a movement distance d3 is reselected from the set of preset movement conditions to be selected as the preset movement condition, and S1 and S2 above are re-executed until the reference chromosome and the corresponding chromosome to be synthesized i The overlap rate meets the preset overlap range. Unpreferably, S3 can be performed before S2, so that all reference chromosomes and the corresponding chromosome to be synthesized i If the overlap ratio meets the preset overlap range, the following S4 is executed.

[0077] S3. Extract reference chromosomes from the reference chromosome image.

[0078] In this embodiment, the computer obtains the reference chromosome according to the reference chromosome mask image after eliminating the edge error. The contour coordinate point set after eliminating edge errors is then converted to the reference chromosome Original center location A new contour coordinate point set is obtained, and multiple reference chromosomes are extracted from the reference chromosome image according to the new contour coordinate point set.

[0079] S4, moving the extracted reference chromosome to the proposed crossing position in the chromosome image to be synthesized, and crossing and overlapping with the chromosome to be synthesized in the image for synthesis, wherein the crossing and overlapping reference chromosome and the chromosome to be synthesized are not the same chromosome.

[0080] The computer extracts multiple reference chromosomes from the reference chromosome image After that, the extracted multiple reference chromosomes are Move to multiple chromosomes to be synthesized in the chromosome image to be synthesizedδ i The corresponding multiple proposed crossover positions are respectively related to the multiple chromosomes to be synthesized in the image δ i Crossover is performed for synthesis, where the crossover reference chromosome and chromosome δ to be synthesized iNot the same chromosome. In this way, the reference chromosome Put it into the chromosome image to be synthesized and the chromosome to be synthesized δ i Cross-synthesis to obtain a cross-chromosome image. This method is used multiple times, each time different chromosome images to be synthesized and reference chromosome images are selected to select different chromosomes to be synthesized and reference chromosomes, or each time different chromosomes to be synthesized and reference chromosomes are selected from the same chromosome images to be synthesized and reference chromosome images, a large number of chromosome image samples with different chromosome crossover situations can be generated for the artificial intelligence algorithm model to perform chromosome crossover recognition analysis training.

[0081] Since the reference chromosome Put it into the chromosome image to be synthesized and the chromosome to be synthesized δ i Cross-synthesis will leave synthesis traces. In order to eliminate synthesis traces, in S4, and chromosome δ to be synthesized i After cross-overlap synthesis, a smoothing filter method is used to eliminate synthesis traces. For example, a Gaussian smoothing method or a wavelet filtering method is used to eliminate synthesis traces. In other embodiments, S4 can be changed to: Put it into the chromosome image to be synthesized and the chromosome to be synthesized δ i At the cross-overlapping position, the synthesis is not performed directly on the chromosome image to be synthesized, but each reference chromosome is They are placed on the newly created graph according to the cross-overlapping positions, and then the conditional adversarial generative network GANs (Generative Adversarial Networks) is used to synthesize the chromosome image to be synthesized and the new image with each reference chromosome placed, so that the synthesized edge will be smoother and more reasonable.

[0082] The above is only an implementation method of the invention, and does not limit the scope of patent protection. Those skilled in the art can make non-substantial changes or substitutions based on the invention, which still fall within the scope of patent protection.

Claims

1. A method for automatically generating a crossover chromosome image, characterized in that: The following steps are involved: S3, extracting a reference chromosome from the reference chromosome image; S4. Move the extracted reference chromosome to the chromosome image to be synthesized, and cross-overlap with the chromosome to be synthesized in the image for synthesis, wherein the cross-overlapping reference chromosome and the chromosome to be synthesized are not the same chromosome.

2. The method for automatically generating a cross-chromosome image as described in claim 1 is characterized in that: in S3, multiple reference chromosomes are specifically extracted; S4 specifically moves the extracted multiple reference chromosomes to the chromosome image to be synthesized so as to cross-overlap with the multiple chromosomes to be synthesized in the image.

3. The method for automatically generating a crossover chromosome image as claimed in claim 1 or 2, characterized in that: This includes the following steps performed before S4: S1, determining a proposed crossover position of the chromosome to be synthesized according to a set of outline coordinate points of the chromosome to be synthesized in the chromosome image to be synthesized and a preset movement condition; S2, based on the reference chromosome outline coordinate point set extracted from the reference chromosome image and the proposed crossing position, pre-calculate the overlap rate of the reference chromosome when it is moved to the chromosome image to be synthesized and crosses and overlaps with the chromosome to be synthesized; The condition for executing S4 is that the overlap rate between each reference chromosome and the corresponding chromosome to be synthesized meets the preset overlap range. Specifically, S4 moves the extracted reference chromosome to the proposed crossover position in the image of the chromosome to be synthesized so as to crossover and overlap with the chromosome to be synthesized. If the overlap rate between a reference chromosome and the corresponding chromosome to be synthesized does not meet the preset overlap range, the preset movement condition is adjusted and S1 and S2 are re-executed until the overlap rate between the reference chromosome and the corresponding chromosome to be synthesized meets the preset overlap range.

4. The method for automatically generating a crossover chromosome image as claimed in claim 3, characterized in that: S1 specifically includes: S11, obtaining the center coordinates of the chromosome to be synthesized according to the outline coordinate point set of the chromosome to be synthesized; S12. Determine the proposed crossover position of the chromosome to be synthesized according to preset movement conditions, wherein the preset movement conditions include the movement distance of the center of the reference chromosome at the center position of the chromosome to be synthesized along the length direction of the chromosome to be synthesized.

5. The method for automatically generating a crossover chromosome image as claimed in claim 4, characterized in that: S1 also includes: S13, based on the outline coordinate point set of the reference chromosome and the above-mentioned quasi-crossing position, calculate whether the reference chromosome will exceed the boundary of the chromosome image to be synthesized when it moves to the quasi-crossing position. If so, adjust the preset movement condition and re-execute S12.

6. The method for automatically generating a crossover chromosome image as claimed in claim 4, characterized in that S2 Specifically, based on the outline coordinate point set of the reference chromosome and the above-mentioned quasi-crossing position, the ratio of the overlapping region between the reference chromosome and the chromosome to be synthesized to the reference chromosome region and / or the ratio to the chromosome region to be synthesized is calculated, and the overlap rate is obtained based on the above-mentioned ratio.

7. The method for automatically generating a crossover chromosome image as claimed in claim 6, characterized in that: S2 includes: S21, converting the original contour coordinate point set of the reference chromosome into a new contour coordinate point set with the proposed crossover position as the new center position of the reference chromosome; S22, generating a mask map of the chromosome to be synthesized and a mask map of the reference chromosome respectively according to the outline coordinate point set of the chromosome to be synthesized and the new outline coordinate point set of the reference chromosome; S23, performing an erosion operation on the reference chromosome mask image to obtain a reference chromosome mask image after eliminating edge errors; S24, calculating the ratio of the overlapping region between the reference chromosome and the chromosome to be synthesized to the reference chromosome region and / or the ratio to the chromosome region to be synthesized based on the reference chromosome mask map and the chromosome mask map after eliminating edge errors, and obtaining the overlap rate based on the above ratios; S3 specifically, obtains a contour coordinate point set of the reference chromosome after eliminating edge errors based on the reference chromosome mask image after eliminating edge errors, converts the contour coordinate point set to a new contour coordinate point set at the original center position of the reference chromosome, and extracts the reference chromosome from the reference chromosome image based on the new contour coordinate point set.

8. The method for automatically generating a crossover chromosome image as claimed in claim 7, characterized in that: S24 specifically obtains the chromosome overlapping region in the following manner: the coordinate points with chromosomes in the reference chromosome mask map after eliminating edge errors and the chromosome mask map to be synthesized are recorded as 1, and the coordinate points without chromosomes are recorded as 0, and the reference chromosome mask map and the chromosome mask map to be synthesized are traversed to obtain all the coordinate points that are 1 in the reference chromosome mask map and in the chromosome mask map to be synthesized at the same time, and the chromosome overlapping region is obtained based on this.

9. The method for automatically generating a crossover chromosome image as claimed in claim 1, wherein in S4: After the reference chromosome and the chromosome to be synthesized are cross-overlapped and synthesized, the smoothing filter method is used to eliminate the synthesis traces; Alternatively, the synthesis may be performed using a generative adversarial network (GANs).

10. A computer program product comprising a computer program, characterized in that The computer program is executed by a processor to implement the method for automatically generating a cross-chromosome image as described in any one of claims 1 to 9.

Images