A method for automatically scoring chromosome images and a computer program product

Abstract

The present invention discloses a method for automatically scoring chromosome images and a computer program product. The scoring method comprises the following steps: S1, for each chromosome image among a plurality of chromosome images to be sorted, the chromosome image is scored according to its chromosome characteristic parameters and image state marking conditions, and a scoring value is obtained, wherein the chromosome characteristic parameters include average chromosome length, image quality score, chromosome dispersion, average chromosome curvature and chromosome karyotype constant value. The automatic scoring method for chromosome images scores chromosome images according to chromosome dispersion, average chromosome length, image quality score, average chromosome curvature, chromosome karyotype constant value and image state marking conditions, and the scoring dimension is more comprehensive, so that the scoring result of the chromosome image can better reflect the quality of the chromosome image, and the chromosome image sorting result can more accurately reflect the quality of each chromosome image.

Description

Technical Field

[0001] The present invention relates to the field of chromosome technology, and in particular to a chromosome image automatic scoring method and a computer program product. Background Art

[0002] The existing chromosome karyotype analysis method is: medical staff manually conduct preliminary screening of a large number of chromosome images, score the screened chromosome images one by one according to their work experience, and select the chromosome images that meet the requirements according to the scoring results. This manual scoring method has low scoring efficiency and is affected by subjective factors. The evaluation is not accurate enough. Different medical staff have different work experience and their scoring standards may be different, resulting in different scoring results for the same chromosome image. To this end, some technicians in the industry have proposed a method for automatically scoring chromosome images using a computer program. This method automatically scores the chromosome images according to their image features. The image features include the number of chromosomes, image clarity, chromosome dispersion, average chromosome length, and the number of homologous chromosome pairs. The chromosome images are scored according to these image features to obtain the score value of each chromosome image, and then the chromosome images are sorted according to the score value. Although this scoring method realizes the automatic scoring of chromosome images, its scoring dimension is not comprehensive enough, resulting in the chromosome image scoring results not being able to better reflect the quality of the chromosome images, so that the chromosome image sorting results cannot accurately reflect the quality of each chromosome image. Summary of the invention

[0003] The purpose of the present invention is to provide a method for automatically scoring chromosome images and a computer program product that implements the method when executed. The scoring results obtained by the scoring method can better reflect the quality of chromosome images, so that the chromosome image sorting results can more accurately reflect the quality of each chromosome image.

[0004] In order to achieve the above object, the present invention provides a method for automatically scoring chromosome images, comprising the following steps:

[0005] S1. For each chromosome image among the multiple chromosome images to be sorted, score the chromosome image according to its chromosome characteristic parameters and image status marking to obtain a score value, wherein the chromosome characteristic parameters include average chromosome length, image quality score, chromosome dispersion, average chromosome curvature and chromosome karyotype constant value.

[0006] Furthermore, step S2 is performed after step S1, sorting each chromosome image from high to low according to the score value of each chromosome image. During the sorting process, if there are multiple chromosome images with the same score value, these multiple chromosome images are sorted according to the image update time, wherein the closer the image update time is, the higher the ranking is.

[0007] Furthermore, if the image status marking conditions of some chromosome images among the multiple chromosome images to be sorted are marked and some are unmarked, then steps S1 and S2 are directly executed for all the marked chromosome images, and for all the unmarked chromosome images, before executing step S1, the following steps S10, S11, and S12 are executed:

[0008] S10, counting the number of chromosomes in each unlabeled chromosome image, and selecting from these chromosome images all chromosome images with a chromosome number of 45, all chromosome images with a chromosome number of 46, and all chromosome images with a chromosome number of 47;

[0009] S11, sorting the three selected chromosome images with the number of chromosomes being 45, 46, and 47 respectively from large to small according to the number of images, and if there are at least two chromosome images with the same number of images, then sorting the three chromosome images in the order of 46>45>47;

[0010] S12, after the above three chromosome images are sorted, the chromosome image selected in step S10 is sorted after the three chromosome images;

[0011] After executing steps S10, S11, and S12, for each of the three types of chromosome images, each chromosome image in the type of chromosome image is scored and sorted according to the steps S1 and S2.

[0012] Furthermore, step S1 specifically scores the chromosome image according to the following formula:

[0013] score=a×avg(length)+b×quality+c×dispersity-d×avg(curvature)+v+s

[0014] Among them, avg(length) refers to the average length of chromosomes in the chromosome image, quality refers to the image quality score of the chromosome image, dispersion refers to the dispersion of chromosomes in the chromosome image, avg(curvature) refers to the average curvature of chromosomes in the chromosome image, v refers to the chromosome karyotype constant value of the chromosome image, which is determined according to the chromosome karyotype length value of the chromosome image, s refers to the image status mark constant value, which is determined according to the image status mark of the chromosome image, and a, b, c, and d are the scoring weights of the average chromosome length, image quality score, chromosome dispersion, and average chromosome curvature, respectively.

[0015] Furthermore, in step S1, the scoring weight values ​​are determined according to the respective priorities of the average chromosome length, the image quality score, the chromosome dispersion, and the average chromosome curvature.

[0016] Further:

[0017] The chromosome karyotype constant value v of the chromosome image is determined as follows:

[0018] If the chromosome karyotype value of the chromosome image is N,XX or N,XY, the chromosome karyotype constant value v = 1.5;

[0019] If the chromosome karyotype value of the chromosome image is not N,XX or N,XY and the karyotype value length is ≤ the first preset value, the chromosome karyotype constant value v=1.2;

[0020] If the chromosome karyotype value of the chromosome image is not N,XX or N,XY and the length of the karyotype value is between the first preset value and the second preset value, the chromosome karyotype constant value v=1;

[0021] If the chromosome karyotype value of the chromosome image is not N,XX or N,XY and the karyotype value length is greater than the second preset value, the chromosome karyotype constant value v=-3;

[0022] And / or determine the image status mark constant value s of the chromosome image as follows:

[0023] If the image state mark of the chromosome image is unmarked, the image state mark constant value s=0;

[0024] If the image status mark of the chromosome image is confirmation count, the image status mark constant value s=1;

[0025] If the image status mark of the chromosome image is confirmation analysis, the image status mark constant value s=3;

[0026] If the image status flag of the chromosome image is to generate a report, the image status flag constant value s=5.

[0027] Furthermore, the nucleotype value length refers to the byte length occupied by the nucleotype value, the first preset value is 17, and the second preset value is 21.

[0028] Further:

[0029] The average chromosome length of the chromosome image is calculated as follows: the average length of chromosomes 1 to 3 and all X chromosomes in the chromosome image is calculated, and the average chromosome length of the chromosome image is taken as the average chromosome length of the chromosome image;

[0030] And / or the average chromosome curvature of the chromosome image is calculated in the following manner: the curvature of all chromosomes in the chromosome image is calculated, and the average chromosome curvature is calculated accordingly.

[0031] Furthermore, in step S2, the chromosome image sorting result is output to the user.

[0032] The present invention also provides a computer program product, including a computer program, which is executed by a processor to implement the above-mentioned automatic chromosome image scoring method.

[0033] Compared with the existing automatic chromosome image scoring method, the chromosome image automatic scoring method provided by the present invention, in step S1, in addition to scoring according to the chromosome dispersion and the average chromosome length as in the existing scoring method, also scores according to the image quality score of the chromosome image, the average chromosome curvature, the chromosome karyotype constant value and the image status mark. The scoring dimension is more comprehensive, so that the chromosome image scoring result can better reflect the quality of the chromosome image, and the chromosome image sorting result can more accurately reflect the quality of each chromosome image. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a schematic diagram of the scoring and sorting process adopted by the chromosome image automatic scoring method for the labeled chromosome images.

[0035] Figure 2 This is a schematic diagram of the scoring and sorting process adopted by the chromosome image automatic scoring method for unlabeled chromosome images. DETAILED DESCRIPTION

[0036] The present invention is further described in detail below in conjunction with specific implementation methods.

[0037] The automatic chromosome image scoring method of this embodiment is programmed with computer code to form a computer program product, which is executed by a processor to implement the automatic chromosome image scoring method. The following takes the automatic chromosome image scoring method executed by a computer processor as an example to illustrate the execution process of the method.

[0038] If a user (e.g., an operator) needs to automatically sort a batch of chromosome images, the user selects the chromosome image data set to be automatically sorted on the computer. After the computer obtains the chromosome image data set selected by the user, it obtains the chromosome characteristic parameters and image status markings of each chromosome image in the data set. The chromosome characteristic parameters include the average chromosome length, image quality score, chromosome dispersion, average chromosome curvature, and chromosome karyotype constant value. The computer obtains each chromosome characteristic parameter in the following manner.

[0039] (1) Average chromosome length avg(length)

[0040] Under normal circumstances, to calculate the average length of chromosomes in a chromosome image, it is necessary to calculate the average length based on the lengths of all chromosomes in the image. Although this calculation method is the most accurate, the amount of calculation is relatively large, which affects the calculation efficiency. In this embodiment, it is considered that chromosomes 1-3 in a general chromosome image are the longest and chromosome X is the shortest. Chromosomes 1-3 and chromosome X are the most recognizable among all chromosomes in a chromosome image. Therefore, in this embodiment, the average length of chromosomes 1-3 and chromosome X in a chromosome image is used as the average length of chromosomes in the image, which greatly reduces the amount of calculation and improves the calculation efficiency. The calculation process is: the computer automatically measures the length of chromosomes 1-3 and the length of all X chromosomes in the chromosome image, and then calculates the average length of chromosomes 1-3 and all X chromosomes in the chromosome image, and uses this as the average length of chromosomes in the chromosome image.

[0041] (2) Image quality score

[0042] The computer uses a trained artificial intelligence model to automatically identify the chromosome image and obtain the image quality score of the chromosome image. The artificial intelligence model uses an existing general large model, and its training process is as follows: the operator manually analyzes the image quality of some chromosome images and scores the image quality, takes a chromosome image as input, and takes the image quality score of the chromosome image as output to form a set of training samples, and then uses many such training samples to train the existing general large model, so that it has the ability to automatically identify the image quality score of the chromosome image.

[0043] (3)Chromosome dispersity

[0044] Dispersion refers to the degree of chromosome crossing in the image. The less crossing, the higher the dispersion. The computer automatically calculates the chromosome crossing ratio in the chromosome image (the crossing ratio is equal to the ratio of the number of chromosomes with crossing to the total number of chromosomes), and then determines the dispersion of the chromosome image based on the crossing ratio. The technician can formulate the corresponding relationship between the crossing ratio and the dispersion in advance and store it in the computer. After the computer calculates the crossing ratio of the chromosome image, it can query the corresponding relationship to determine the dispersion of the chromosome image.

[0045] (4) Average chromosome curvature avg (curvature)

[0046] The computer calculates the average chromosome curvature avg (curvature) of the chromosome image according to the following formula:

[0047]

[0048] Among them, d i is the width of the ith chromosome, d med is the median width of a single image, and N is the total number of chromosomes in the chromosome image.

[0049] (5) Chromosome karyotype constant value v

[0050] The computer first identifies the karyotype value of the chromosome image, then determines the byte length occupied by the karyotype value of the chromosome image, and then queries the corresponding relationship between the karyotype value byte length and the chromosome karyotype constant value based on the byte length of the karyotype value to determine the chromosome karyotype constant value of the chromosome image.

[0051] The correspondence between the byte length of the karyotype value and the chromosome karyotype constant value is as follows:

[0052] If the chromosome karyotype value of the chromosome image is N,XX or N,XY, the chromosome karyotype constant value v = 1.5;

[0053] If the chromosome karyotype value of the chromosome image is not N,XX or N,XY and the karyotype value length is ≤17 (ie, ≤the first preset value), the chromosome karyotype constant value v=1.2;

[0054] If the chromosome karyotype value of the chromosome image is not N,XX or N,XY and the length of the karyotype value is in the interval of (17, 21] (i.e., between the first preset value and the second preset value), the chromosome karyotype constant value v=1;

[0055] If the chromosome karyotype value of the chromosome image is not N,XX or N,XY and the karyotype value length is >21 (i.e. >the second preset value), the chromosome karyotype constant value v=-3;

[0056] Wherein, N is a positive integer.

[0057] The karyotype value recognized by the ideal chromosome image should be N,XX or N,XY. If the karyotype value of the chromosome image is not N,XX or N,XY, then the shorter the karyotype value, the better the recognition effect of the chromosome image, that is, the better the quality of the chromosome image. On the contrary, the longer the karyotype value, the worse the recognition effect of the chromosome image, that is, the worse the quality of the chromosome image. Therefore, this embodiment characterizes the recognition effect of the chromosome image in the karyotype value recognition through the chromosome karyotype constant value v, which is equivalent to characterizing the quality of the chromosome image. Specifically, the maximum chromosome karyotype constant value v is assigned to the karyotype value of N,XX or N,XY. The longer the karyotype value, the smaller the corresponding chromosome karyotype constant value v. For example, if the karyotype value of the chromosome image is "46,XX", it belongs to N,XX, and its chromosome karyotype constant value v=1.5. For another example, the karyotype value of a chromosome image is "46,XX,-2,+13" (representing one less chromosome 2 and one more chromosome 13; the more chromosome errors the chromosome image identifies, the longer the karyotype value length), which does not belong to N,XX or N,XY. Therefore, the computer further determines that the byte length of the karyotype value of the chromosome image is 12, that is, the karyotype value length of the chromosome image is <17, so its chromosome karyotype constant value v=1.2.

[0058] After the computer obtains each chromosome characteristic parameter of each chromosome image in the above manner, it also obtains the image state marking situation of each chromosome image. In this embodiment, different image state marking situations are assigned different image state marking constant values ​​s. After the computer obtains the image state marking situation of the chromosome image, it determines the value of the corresponding image state marking constant value s accordingly.

[0059] The corresponding relationship between the image status mark and the image status mark constant value s is as follows:

[0060] If the image state mark of the chromosome image is unmarked, the image state mark constant value s=0;

[0061] If the image status mark of the chromosome image is confirmation count, the image status mark constant value s=1;

[0062] If the image status mark of the chromosome image is confirmation analysis, the image status mark constant value s=3;

[0063] If the image status flag of the chromosome image is to generate a report, the image status flag constant value s=5.

[0064] After obtaining the chromosome characteristic parameter values ​​of each chromosome image to be sorted and the image state mark constant value s corresponding to the image state mark, the computer scores and sorts each chromosome image accordingly. The process is as follows:

[0065] If some of the chromosome images to be sorted have image status marks marked while others are unmarked, the computer will classify them into marked chromosome images and unmarked chromosome images according to the image status marks of the chromosome images to be sorted. Figure 1 The scoring and sorting process shown in FIG. 1 is used to score and sort the labeled chromosome images. Figure 2 The scoring and ranking process shown performs scoring and ranking on unlabeled chromosome images.

[0066] Score ranking for labeled chromosome images:

[0067] The computer scores each chromosome image among the marked chromosome images to be sorted according to its various chromosome characteristic parameters and image state marking conditions to obtain a score value, and the score value is positively correlated with the numerical value of each chromosome characteristic parameter and positively correlated with the image state marking constant value s corresponding to the image state marking condition. Specifically, the computer substitutes its various chromosome characteristic parameters (average chromosome length, image quality score, chromosome dispersion, average chromosome curvature and chromosome karyotype constant value) and the image state marking constant value s corresponding to the image state marking condition into the following scoring formula to calculate the score value score of the chromosome image.

[0068] score=a×avg(length)+b×quality+c×dispersity-d×avg(curvature)+v+s

[0069] Among them, avg(length) refers to the average length of chromosomes in the chromosome image, quality refers to the image quality score of the chromosome image, dispersion refers to the dispersion of chromosomes in the chromosome image, avg(curvature) refers to the average curvature of chromosomes in the chromosome image, v refers to the chromosome karyotype constant value of the chromosome image, which is determined according to the chromosome karyotype length value of the chromosome image, s refers to the image status mark constant value, which is determined according to the image status mark of the chromosome image, and a, b, c, and d are the scoring weights of the average chromosome length, image quality score, chromosome dispersion, and average chromosome curvature, respectively.

[0070] In this embodiment, the scoring weight values ​​are determined in advance according to the respective priorities of the average length of chromosomes, the image quality score, the dispersion of chromosomes, and the average curvature of chromosomes. Specifically, the average length of chromosomes, the dispersion of chromosomes, and the average curvature of chromosomes have the same priority, while the priority of the image quality score is slightly lower than the above three. Therefore, the scoring weight values ​​of the average length of chromosomes, the dispersion of chromosomes, and the average curvature of chromosomes should be the same and higher than the scoring weight value of the image quality score. However, since the order of magnitude of the average length of chromosomes is 2 orders higher than the order of magnitude of the image quality score, the dispersion of chromosomes, and the average curvature of chromosomes, when scoring, each characteristic parameter should be at the same order of magnitude to be able to more appropriately balance the influence of each characteristic parameter. In order to make each characteristic parameter at the same order of magnitude, it is necessary to reduce the order of magnitude of the average length of chromosomes. Therefore, in this embodiment, the scoring weight values ​​of the dispersion of chromosomes and the average curvature of chromosomes are reduced by 2 orders of magnitude as the scoring weight value of the average length of chromosomes. For example, the scoring weight values ​​of the dispersion of chromosomes and the average curvature of chromosomes are both 6, the scoring weight value of the average length of chromosomes is 0.06, and the scoring weight value of the image quality score is 1. In this way, the order of magnitude of each feature parameter multiplied by the score weight value is the same. In a non-preferred embodiment, the score weight value is not downgraded by an order of magnitude, but the average chromosome length is normalized in advance so that it is at the same order of magnitude as the image quality score, the chromosome dispersion, and the average chromosome curvature. Avg (length) is specifically the average chromosome length after normalization. In this way, the score weight values ​​of the average chromosome length, the chromosome dispersion, and the average chromosome curvature are all 6, and the score weight value of the image quality score is 1.

[0071] After the computer calculates the score value of each marked chromosome image, it can sort the chromosome images from high to low according to the score value of each chromosome image. During the sorting process, if there are multiple chromosome images with the same score value, these chromosome images are sorted according to the image update time, among which the more recent the image update time, the higher the ranking.

[0072] Score ranking for unlabeled chromosome images:

[0073] Under normal circumstances, humans have 46 chromosomes, and only a small number of chromosomes increase or decrease due to factors such as gene mutations. Therefore, chromosome images with 46 chromosomes should account for the vast majority of the chromosome image data set. It is not ruled out that the number of chromosomes in some chromosome images is actually 46, but due to image quality problems, only part of the chromosomes can be identified, resulting in less than 46 chromosomes. Since the operator has not marked these chromosome images, the computer needs to preliminarily sort the chromosome images according to the number of chromosomes. Specifically: the computer first counts the number of chromosomes in each unmarked chromosome image, and selects all chromosome images with a chromosome number of 45, all chromosome images with a chromosome number of 46, and all chromosome images with a chromosome number of 47 from these chromosome images; the computer then sorts the three selected chromosome images with chromosome numbers of 45, 46, and 47 from most to least according to their image numbers, and if there are at least two chromosome images with the same image number, these three chromosome images are sorted in the order of chromosome number 46>45>47; after the three chromosome images with chromosome numbers of 45, 46, and 47 are sorted, the computer arranges the remaining chromosome images with other chromosome numbers after these three chromosome images. Normally, an ideal chromosome image has 46 chromosomes, so a chromosome image with 46 chromosomes should be ranked before chromosome images with other chromosome numbers. This embodiment takes into account that chromosome images with 46 chromosomes account for the vast majority in the chromosome image data set, so this embodiment directly sorts the images from most to least according to the number of images to rank the chromosome images with 46 chromosomes in front. For example, if there are 100 chromosome images with 45 chromosomes, 130 chromosome images with 46 chromosomes, and 120 chromosome images with 47 chromosomes, then the computer ranks the chromosome images with 46 chromosomes in front, the chromosome images with 47 chromosomes in the middle, the chromosome images with 45 chromosomes behind the chromosome images with 47 chromosomes, and then the chromosome images with other chromosome numbers behind the chromosome images with 45 chromosomes. If the number of chromosome images with 45 chromosomes, chromosome images with 46 chromosomes, and chromosome images with 47 chromosomes are all 100, then the computer will sort these three chromosome images in the order of 46>45>47, that is, the chromosome image with 46 chromosomes is placed in front, the chromosome image with 45 chromosomes is placed in the middle, the chromosome image with 47 chromosomes is placed after the chromosome image with 45 chromosomes, and then the chromosome images with other chromosome numbers are placed after the chromosome image with 47 chromosomes.

[0074] After the chromosome images are preliminarily sorted according to the number of chromosomes, the computer further sorts each of the three chromosome images with chromosome numbers of 45, 46, and 47. Specifically: (1) For each chromosome image with a chromosome number of 45, the chromosome characteristic parameters (average chromosome length, image quality score, chromosome dispersion, average chromosome curvature, and chromosome karyotype constant value) and the image state mark constant value s corresponding to the image state mark are substituted into the scoring formula above to calculate the score of the chromosome image; then, the score of each chromosome image with a chromosome number of 45 is calculated according to the score of each chromosome image. (1) Sort the chromosome images according to their score values ​​from high to low. During the sorting process, if there are multiple chromosome images with the same score value, they are sorted according to the image update time, where the image with a more recent update time is ranked higher. In this way, the chromosome images with a chromosome number of 45 are scored and sorted. (2) For each chromosome image with a chromosome number of 46, the image status mark constant value corresponding to each chromosome characteristic parameter (average chromosome length, image quality score, chromosome dispersion, average chromosome curvature and chromosome karyotype constant value) and the image status mark condition is scored. Substitute s into the scoring formula above to calculate the score of the chromosome image; then sort the chromosome images from high to low according to the scores of the chromosome images with a chromosome number of 46. During the sorting process, if there are multiple chromosome images with the same score, they are sorted according to the image update time. The closer the image update time, the higher the ranking. In this way, the scoring and sorting of the chromosome images with a chromosome number of 46 are completed; (3) For each chromosome image with a chromosome number of 47, the chromosome characteristic parameters (chromosome average length, image The quality score, chromosome dispersion, average chromosome curvature and chromosome karyotype constant value) and the image state mark constant value s corresponding to the image state mark are substituted into the above scoring formula to calculate the score of the chromosome image; then, the chromosome images are sorted from high to low according to the score values ​​of the chromosome images with a chromosome number of 47. During the sorting process, if there are multiple chromosome images with the same score value, the multiple chromosome images are sorted according to the image update time, wherein the closer the image update time is, the higher the ranking is, thus completing the scoring and sorting of the chromosome images with a chromosome number of 47. Chromosome images with other chromosome numbers, such as chromosome images with a chromosome number of 48, are also scored and sorted according to the above-mentioned scoring and sorting method.Since the chromosome images of other chromosome numbers are usually of poor quality and are usually not used by operators, it is not preferred that the chromosome images of other chromosome numbers be changed to not be scored and sorted, and be directly arranged after the chromosome images of chromosome numbers 45, 46, and 47 according to their original order.

[0075] Since the image status mark constant values ​​of unmarked chromosome images are all 0, preferably, other embodiments may rewrite the above scoring formula as follows:

[0076] score=a×avg(length)+b×quality+c×dispersity-d×avg(curvature)+v

[0077] In this way, when calculating the score of each unlabeled chromosome image, only the chromosome characteristic parameters (average chromosome length, image quality score, chromosome dispersion, average chromosome curvature and chromosome karyotype constant value) are substituted into the scoring formula to calculate the score of the chromosome image.

[0078] After the computer scores and sorts the marked chromosome images and the unmarked chromosome images respectively, it obtains the sorting results of the marked chromosome images and the sorting results of the unmarked chromosome images, that is, the batch of chromosome images to be sorted are divided into two categories of chromosome images to obtain their respective sorting results. The computer then outputs the sorting results of the marked chromosome images and the sorting results of the unmarked chromosome images together as the sorting results of the batch of chromosome images to be sorted to the user (i.e., the operator). Since if the chromosome image has been marked, it means that the user approves the chromosome image, the computer preferably sorts the marked chromosome images according to the sorting results of the marked chromosome images, and then all the unmarked chromosome images are sorted after the marked chromosome images and the unmarked chromosome images are sorted according to the sorting results of the unmarked chromosome images, so as to merge the sorting results of all the chromosome images to be sorted and output the sorting results to the user. If all the chromosome images to be sorted are marked, the computer only uses the following method: Figure 1 The sorting process shown in the figure is used to score and sort this batch of chromosome images. The output sorting results are only the sorting results of the labeled chromosome images. If all the chromosome images to be sorted are not labeled, the computer will only use the following method: Figure 2 The sorting process shown scores and sorts this batch of chromosome images, and the output sorting results are only the sorting results of the unlabeled chromosome images.

[0079] Compared with the existing automatic chromosome image scoring method, the chromosome image automatic scoring method provided by the present invention, when scoring chromosome images, in addition to scoring according to chromosome dispersion and average chromosome length as in the existing scoring method, also scores according to the image quality score of the chromosome image, average chromosome curvature, chromosome karyotype constant value and image status marking. The scoring dimensions are more comprehensive, so that the scoring results of the chromosome image can better reflect the quality of the chromosome image, and the chromosome image sorting results can more accurately reflect the quality of each chromosome image. The present invention characterizes the recognition effect of the chromosome image on the karyotype value recognition by the chromosome karyotype constant value v, and takes into account the recognition effect of the chromosome image on the karyotype value recognition when scoring, so that the scoring results of the chromosome image can better reflect the quality of the chromosome image. The present invention uses an image status mark constant value s (reflecting the image status mark status) to characterize the degree of marking made by the operator on the chromosome image, and the marking degree is positively correlated with the operator's analysis and evaluation of the chromosome image. Therefore, the image status mark constant value is used to characterize the operator's analysis and evaluation of the chromosome image. The operator's analysis and evaluation of the chromosome image is integrated into the scoring, making the chromosome image sorting more reasonable, and being able to put chromosome images with good quality that better meet the operator's needs in front, so as to recommend reliable chromosome images with good quality to the operator.

[0080] 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 scoring chromosome images, characterized in that: The following steps are involved: S1. For each chromosome image among the plurality of chromosome images to be sorted, scoring is performed according to the chromosome characteristic parameters and image status marking conditions to obtain a scoring value, wherein the chromosome characteristic parameters include an average chromosome length, an image quality score, a chromosome dispersion, an average chromosome curvature, and a chromosome karyotype constant value; Step S1 specifically scores the chromosome image according to the following formula: score=a×avg(length)+b×quality+c×dispersity-d×avg(curvature)+v+s Among them, avg(length) refers to the average length of chromosomes in the chromosome image, quality refers to the image quality score of the chromosome image, dispersion refers to the dispersion of chromosomes in the chromosome image, avg(curvature) refers to the average curvature of chromosomes in the chromosome image, v refers to the chromosome karyotype constant value of the chromosome image, which is determined according to the chromosome karyotype length value of the chromosome image, s refers to the image status mark constant value, which is determined according to the image status mark of the chromosome image, a, b, c, d are the scoring weights of the average chromosome length, image quality score, chromosome dispersion, and average chromosome curvature, respectively; The chromosome karyotype constant value v of the chromosome image is determined as follows: If the chromosome karyotype value of the chromosome image is N,XX or N,XY, the chromosome karyotype constant value v=1.5; If the chromosome karyotype value of the chromosome image is not N,XX or N,XY and the karyotype value length is ≤ the first preset value, the chromosome karyotype constant value v=1.2; If the chromosome karyotype value of the chromosome image is not N,XX or N,XY and the length of the karyotype value is between the first preset value and the second preset value, the chromosome karyotype constant value v=1; If the chromosome karyotype value of the chromosome image is not N,XX or N,XY and the karyotype value length is greater than the second preset value, the chromosome karyotype constant value v=-3; And / or determine the image status mark constant value s of the chromosome image as follows: If the image state mark of the chromosome image is unmarked, the image state mark constant value s=0; If the image status mark of the chromosome image is confirmation count, the image status mark constant value s=1; If the image status mark of the chromosome image is confirmation analysis, the image status mark constant value s=3; If the image status flag of the chromosome image is to generate a report, the image status flag constant value s=5.

2. The automatic chromosome image scoring method according to claim 1, characterized in that: The method comprises step S2 executed after step S1, sorting each chromosome image from high to low according to the score value of each chromosome image. During the sorting process, if there are multiple chromosome images with the same score value, the multiple chromosome images are sorted according to the image update time, wherein the closer the image update time is, the higher the ranking is.

3. The automatic chromosome image scoring method according to claim 2, characterized in that: If the image status marking status of some chromosome images among the multiple chromosome images to be sorted is marked and some is unmarked, then steps S1 and S2 are directly executed for all the marked chromosome images, and for all the unmarked chromosome images, before executing step S1, the following steps S10, S11, and S12 are executed: S10, counting the number of chromosomes in each unlabeled chromosome image, and selecting from these chromosome images all chromosome images with a chromosome number of 45, all chromosome images with a chromosome number of 46, and all chromosome images with a chromosome number of 47; S11, sorting the three selected chromosome images with the number of chromosomes being 45, 46, and 47 respectively from large to small according to the number of images, and if there are at least two chromosome images with the same number of images, then sorting the three chromosome images in the order of 46>45>47; S12, after the above three chromosome images are sorted, the chromosome image selected in step S10 is sorted after the three chromosome images; After executing steps S10, S11, and S12, for each of the three types of chromosome images, each chromosome image in the type of chromosome image is scored and sorted according to the steps S1 and S2.

4. The method for automatically scoring chromosome images as claimed in claim 1, wherein the step In S1, the scoring weight values ​​are determined according to the respective priorities of the average chromosome length, image quality score, chromosome dispersion, and average chromosome curvature.

5. The automatic chromosome image scoring method according to claim 1, characterized in that: The nucleotype value length refers to the byte length occupied by the nucleotype value, the first preset value is 17, and the second preset value is 21.

6. The automatic chromosome image scoring method according to claim 1, characterized in that: The average chromosome length of the chromosome image is calculated as follows: the average length of chromosomes 1 to 3 and all X chromosomes in the chromosome image is calculated, and the average chromosome length of the chromosome image is taken as the average chromosome length of the chromosome image; And / or the average chromosome curvature of the chromosome image is calculated in the following manner: the curvature of all chromosomes in the chromosome image is calculated, and the average chromosome curvature is calculated accordingly.

7. The automatic chromosome image scoring method according to any one of claims 1 to 3, characterized in that: In step S2, the chromosome image sorting result is output to the user.

8. A computer program product comprising a computer program, characterized in that The computer program is executed by a processor to implement the automatic chromosome image scoring method as described in any one of claims 1 to 7.

Images