Chromosome contour optimization method and device and storage medium

Abstract

The invention relates to the technical field of chromosome image processing, in particular to a chromosome contour optimization method and device and a storage medium, and the method comprises the steps: intercepting a single chromosome image with an overlapping region from a segmented image; carrying out image background impurity removal and binaryzation to obtain a first chromosome binaryzation image; reconstructing and optimizing the first chromosome binary image to obtain a first mask image; obtaining a chromosome contour mask pattern based on the first mask pattern and the first chromosome binary pattern; and obtaining a target single chromosome image based on the chromosome contour mask pattern. The Gaussian mixture model is introduced to carry out contour optimization on the binarized image, the skeleton of the chromosome is accurately extracted, the chromosome reconstructed based on skeleton information is fused with the original contour image to obtain the target single chromosome image, the rest chromosome parts of the overlapping part and background interference are removed, and the image quality is improved. And the lines at the edges of the chromosomes are smoother and more natural, so that the final effect better conforms to the intuitive feeling of human eye recognition.

Description

Technical Field

[0001] The present invention relates to the technical field of chromosome image processing, and in particular to a chromosome outline optimization method, device and storage medium. Background Art

[0002] At present, the segmentation of chromosome images is mainly based on large AI models. However, the single chromosomes segmented by the AI ​​model are often surrounded by background or parts of other chromosomes, resulting in low-precision images produced by the AI ​​model segmentation, which cannot meet the needs of subsequent research and processing. As a result, in the actual diagnosis and report generation process, manual reprocessing is usually required to perform fine corrections to the chromosome outline to delete excess background and impurities and supplement the chromosome parts lost by AI segmentation, resulting in low efficiency of subsequent processing operations.

[0003] Based on the problems in the prior art, the present invention provides a chromosome profile optimization method, device and storage medium. Summary of the Invention

[0004] The purpose of the present invention is to provide a chromosome outline optimization method, device and storage medium to solve the technical problem in the prior art that the segmented chromosome image has low accuracy and cannot meet the subsequent use requirements.

[0005] The technical solution of the present invention is: a chromosome profile optimization method, comprising: Based on the segmented image, a single chromosome image with overlapping areas is intercepted; the background of the single chromosome image is removed and binarized to obtain a first chromosome binarized image; the first chromosome binarized image is reconstructed and optimized to remove redundant contour areas of the overlapping areas to obtain a first chromosome mask image; the first mask image is fused with the first chromosome binarized image to obtain a chromosome contour mask image; the chromosome contour mask image is fused with the initial single chromosome image to obtain a target single chromosome image.

[0006] Preferably, the method of reconstructing and optimizing the first chromosome binarized image to obtain the first mask image is as follows: Based on a single chromosome image, the chromosome binarization image is pixel-wise modeled using a Gaussian mixture model to obtain a second chromosome binarization image. Generate a chromosome skeleton map from the second chromosome binarization map; Based on the second chromosome binarized image, a distance weight map corresponding to the binarized image is obtained through a distance transformation algorithm; Perform a logical AND operation on the distance weight map and the chromosome skeleton map to obtain a skeleton distance weight map; Create a new all-zero pixel image of the same size as the chromosome skeleton image, calculate the coordinates and pixel mean of each non-zero point in the chromosome skeleton image, draw a circle image with the coordinates of each non-zero point as the center and the pixel mean as the radius, and fill the circle with pixels to obtain the first mask image of the chromosome.

[0007] Preferably, the method of performing pixel modeling on the chromosome binarization image by using a Gaussian mixture model to obtain the second chromosome binarization image includes: Use the expectation maximization algorithm to iteratively optimize, calculate the posterior probability of the sample point belonging to each Gaussian component and update the mean, covariance and weight of the Gaussian distribution, and then converge the model by maximizing the log-likelihood function; After modeling, each pixel point is correspondingly set with a label. Label values ​​greater than 0.5 are set to 255, and the rest are set to 0 to obtain the second chromosome binarization map.

[0008] Preferably, based on a morphological thinning algorithm, boundary pixels are iteratively deleted until a single-pixel wide skeleton is obtained, a 3×3 neighborhood template is used to detect deletable pixels, and the topological structure and endpoints are kept unchanged, and a skeleton image is generated from the second chromosome binarization image.

[0009] Preferably, based on the second chromosome binarized image, the distance from the pixel to the nearest background point is calculated, the local minimum distance is calculated through a scanning algorithm, and the global distance is optimized, and the distance value is updated based on dynamic programming to obtain the distance weight map corresponding to the second chromosome binarized image.

[0010] Preferably, the chromosome skeleton graph and the distance weight graph have the same coordinate system and corresponding coordinates; In the skeleton distance weighted map, the value V of each non-zero pixel point (x, y) means the width of the chromosome at the coordinate point (x, y) in the second chromosome binarization map.

[0011] Preferably, the single chromosome image with background removed is binarized by a fixed threshold value to obtain a first chromosome binarization image; the threshold value interval is 100-200.

[0012] An electronic device comprises a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or at least one program is loaded and executed by the processor to implement the chromosome profile optimization method.

[0013] A computer-readable storage medium stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by a processor to implement the chromosome profile optimization method.

[0014] Compared with the prior art, the advantages of the present invention are: The present invention binarizes a single chromosome image to remove background impurities in advance, introduces a Gaussian mixture model to optimize the contour of the binarized chromosome image, and more accurately extracts the skeleton of the chromosome. At the same time, based on the extracted skeleton information, a "standard" chromosome is reconstructed using the circle drawing method and merged with the original contour. Compared with the original segmented chromosome area, the target single chromosome image obtained after algorithm optimization almost removes the interference of the overlapping parts of the remaining chromosomes and the background. The lines of the chromosome edges are smoother and more natural, making the final effect more in line with the intuitive feeling of human eye recognition.

[0015] Compared with the existing technology, the edge processing of overlapping and crossing chromosomes in the present invention can almost completely eliminate the interference of the background, accurately segment the area belonging only to the current chromosome, and obtain better image effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 A schematic diagram of the chromosome outline optimization process provided by the present invention; Figure 2 A schematic diagram of a single chromosome image with overlapping regions cut out from the original image provided by the present invention; Figure 3 A schematic diagram of a single chromosome image with overlapping regions after background removal provided by an embodiment of the present invention; Figure 4 A schematic diagram of a first chromosome binarization graph provided by an embodiment of the present invention; Figure 5 A schematic diagram of a chromosome skeleton diagram provided by an embodiment of the present invention; Figure 6 A schematic diagram of a distance weighted graph provided in an embodiment of the present invention; Figure 7 A schematic diagram of a skeleton distance weighted map provided in an embodiment of the present invention; Figure 8 A schematic diagram of a first mask pattern provided by an embodiment of the present invention; Figure 9 A schematic diagram of a chromosome outline mask provided by an embodiment of the present invention; Figure 10 A schematic diagram of a target single chromosome image provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The present invention will be described in further detail below with reference to specific embodiments: The present invention provides a chromosome outline optimization method to solve the problem that the chromosome background impurities after AI segmentation are not removed cleanly, which reduces the reliability of chromosome image application. The flowchart is shown in the attached figure. Figure 1 As shown, the method process is as follows: Step 1: Chromosome pre-processing; Based on the AI ​​segmentation model, an initial contraction is performed on the chromosome outline.

[0018] According to the initial AI segmentation outline, a single chromosome image with overlapping areas is cut out from the original image, such as Figure 2 As shown in the figure, the chromosome region described by the thick line is the target chromatid image. After removing the background, see the attached Figure 3 shown.

[0019] The single chromosome image with the overlapping area is binarized by the threshold Threshold to obtain the first chromosome binarization image (binarization mask image), as shown in the attached figure. Figure 4 The threshold is the pixel mean of the image, and the threshold interval ranges from 0 to 255. The specific interval range depends on the specific image. In this embodiment, the threshold interval is 100 to 200, which is used to separate the foreground and background, retain the required foreground part, and remove the background.

[0020] Step 2: Extract the outline skeleton; a. Based on the single chromosome image with overlapping regions in step 1, use the Gaussian mixture model (GMM) to perform pixel modeling on the first chromosome binary image. Use the EM (Expectation Maximization) algorithm for iterative optimization to calculate the posterior probability that the sample point belongs to each Gaussian component and update the mean, covariance, and weight of the Gaussian distribution. Finally, the model is converged by maximizing the log-likelihood function. Two-dimensional Gaussian distribution formula: ; .

[0021] b. After modeling, each pixel point is assigned a label. Label values ​​greater than 0.5 are set to 255, and the rest are set to 0, resulting in the second chromosome binary map.

[0022] c. Based on morphological thinning algorithms, such as Zhang-Suen or Guo-Hall, iteratively delete boundary pixels until a single-pixel wide skeleton is obtained. Use a 3×3 neighborhood template to detect removable pixels while keeping the topology and endpoints unchanged. Generate a skeleton image from the second chromosome binarization image. See Appendix. Figure 5 shown.

[0023] Step 3: Generate new contour; (1) Based on the second chromosome binary map, calculate the distance from the pixel to the nearest background point. Use a two-pass scanning algorithm, such as the Rosenfeld algorithm. The first pass scans from the upper left to the lower right to calculate the local minimum distance; the second pass scans from the lower right to the upper left to optimize the global distance. Update the distance value based on dynamic programming to obtain the distance weight map corresponding to the second chromosome binary map, such as Figure 6 As shown, the brightness and darkness of the pixels represent the size of the pixel value, and the brighter the area, the larger the pixel value.

[0024] (2) Perform a logical AND operation on the distance weight map and the chromosome skeleton map obtained in step 2 to obtain the skeleton distance weight map, such as Figure 7 shown.

[0025] (3) In the skeleton distance weighted map, the value V of each non-zero pixel point (x, y) means the width of the chromosome at the coordinate point (x, y) in the second chromosome binarization map.

[0026] Count the number N of non-zero pixels and non-zero pixel values ​​Vi in the skeleton distance weighted image, and calculate the pixel mean Vmean. The calculation formula is as follows: .

[0027] (4) Create a new pixel-zero image of the same size as the chromosome skeleton image and calculate the coordinates of each non-zero point in the chromosome skeleton image , we get the point set { };by{ The coordinates of each point in is the center of the circle, and the pixel mean calculated above is As the radius, draw a circle image, and fill the value inside the circle image with 255. Get the first mask image of the chromosome, as shown Figure 8 shown.

[0028] Step 4: Obtain target chromosome image; Perform a logical AND operation on the first mask image and the first chromosome binary image to obtain the chromosome outline mask image, such as Figure 9 shown.

[0029] The chromosome outline mask image is subjected to a logical AND operation with the initially intercepted single chromosome image with overlapping regions to perform chromosome fusion and obtain the target single chromosome image as shown in FIG. Figure 10 shown.

[0030] Through the above Figure 10 and Figure 3 It can be seen from the comparison that the single chromosome obtained by cutting after algorithm optimization ( Figure 10 ) compared to the chromosome region segmented by the original AI ( Figure 3), almost removing the redundant chromosome part (the part visible to the naked eye in the overlapping area) and background interference in the overlapping part, Figure 10 The lines at the edges of the chromosomes are smoother and more natural, making the final effect more in line with the intuitive feeling of human eye recognition.

[0031] This paper introduces the GMM algorithm to more accurately extract the chromosome skeleton. Based on the extracted skeleton information, a "standard" chromosome is reconstructed using a circle-drawing method and merged with the original outline. Compared to existing technologies, this solution can almost completely eliminate background interference when processing the edges of overlapping and crossing chromosomes, accurately segmenting the region belonging to the current chromosome and achieving better image results.

[0032] An embodiment of the present invention also provides an electronic device, which includes a processor and a memory; the memory stores one or more instructions, and the one or more instructions are suitable for the processor to load and execute to implement a chromosome profile optimization method as described in the above method embodiment.

[0033] The memory can be used to store software programs and modules. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory. The memory can mainly include a program storage area and a data storage area. The program storage area can store the operating system, application programs required for functions, etc.; the data storage area can store data created based on the use of the device, etc. In addition, the memory can include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory can also include a memory controller to provide the processor with access to the memory.

[0034] The internal structure of the electronic device provided by the embodiment of the present invention may include but is not limited to: a processor, a memory and a communication interface, wherein the processor, memory and communication interface in the electronic device may be connected via a bus or other means. In the embodiments of this specification, connection via a bus is used as an example.

[0035] Among them, the processor (or CPU, Central Processing Unit) is the computing core and control core of the electronic device. The communication interface is used for communication between the memory and the processor. The memory is used to store programs and data. It can be understood that the memory here can be a high-speed RAM storage device or a non-volatile memory device (non-volatile memory), such as at least one disk storage device; optionally, it can also be at least one storage device located away from the aforementioned processor. The memory provides a storage space, which stores the operating system of the electronic device, which may include but is not limited to: Windows system (an operating system), Linux system (an operating system), etc., and the present invention is not limited to this; and the storage space also stores a computer program (including program code) suitable for being loaded and executed by the processor. In the embodiment of this specification, the processor loads and executes the computer program stored in the memory to implement a chromosome profile optimization method provided by the above method embodiment.

[0036] An embodiment of the present invention also provides a computer-readable storage medium, which can be set in an electronic device to store at least one instruction, at least one program, code set or instruction set related to the chromosome profile optimization method in the method embodiment. The at least one instruction, at least one program, code set or instruction set can be loaded and executed by the processor of the electronic device to implement a chromosome profile optimization method provided by the above method embodiment.

[0037] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store program codes.

[0038] It should be noted that the order of the embodiments of the present invention described above is for illustrative purposes only and does not represent the superiority or inferiority of the embodiments. The above description is of specific embodiments of this specification, and other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0039] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the device embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0040] Those skilled in the art will appreciate that all or part of the steps for implementing the above embodiments may be accomplished by hardware, or by programs instructing related hardware to accomplish the steps. The programs may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk, or an optical disk, etc.

[0041] The above disclosure is only a preferred embodiment of the present invention and certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A chromosome profile optimization method, characterized in that: include: Based on the segmented image, a single chromosome image with overlapping regions is intercepted; The background of a single chromosome image is removed and binarized to obtain the first chromosome binarization image; Reconstruct and optimize the first chromosome binary image, remove the redundant contour areas of the overlapping areas, and obtain the first mask image of the chromosome; Performing image fusion on the first mask image and the first chromosome binarization image to obtain a chromosome outline mask image; The chromosome outline mask image is fused with the initial single chromosome image to obtain the target single chromosome image.

2. A chromosome profile optimization method according to claim 1, characterized in that, The process of reconstructing and optimizing the first chromosome binary image to obtain the first mask image is as follows: Based on a single chromosome image, the chromosome binarization image is pixel-wise modeled using a Gaussian mixture model to obtain a second chromosome binarization image. Generate a chromosome skeleton map from the second chromosome binarization map; Based on the second chromosome binarized image, a distance weight map corresponding to the binarized image is obtained through a distance transformation algorithm; Perform a logical AND operation on the distance weight map and the chromosome skeleton map to obtain a skeleton distance weight map; Create a new all-zero pixel image of the same size as the chromosome skeleton image, calculate the coordinates and pixel mean of each non-zero point in the chromosome skeleton image, draw a circle image with the coordinates of each non-zero point as the center and the pixel mean as the radius, and fill the circle with pixels to obtain the first mask image of the chromosome.

3. A chromosome profile optimization method according to claim 2, characterized in that, A method for performing pixel modeling on the chromosome binarization image through a Gaussian mixture model to obtain a second chromosome binarization image includes: Use the expectation maximization algorithm to iteratively optimize, calculate the posterior probability of the sample point belonging to each Gaussian component and update the mean, covariance and weight of the Gaussian distribution, and then converge the model by maximizing the log-likelihood function; After modeling, each pixel point is correspondingly set with a label. Label values ​​greater than 0.5 are set to 255, and the rest are set to 0 to obtain the second chromosome binarization map.

4. A chromosome profile optimization method according to claim 3, characterized in that, Based on the morphological thinning algorithm, boundary pixels are iteratively deleted until a single-pixel wide skeleton is obtained. A 3×3 neighborhood template is used to detect removable pixels while keeping the topology and endpoints unchanged. A skeleton image is generated from the second chromosome binarization image.

5. A chromosome profile optimization method according to claim 3, characterized in that, Based on the second chromosome binarized image, the distance from the pixel to the nearest background point is calculated. The local minimum distance is calculated through a scanning algorithm, and the global distance is optimized. The distance value is updated based on dynamic programming to obtain the distance weight map corresponding to the second chromosome binarized image.

6. A chromosome profile optimization method according to claim 2, characterized in that, The chromosome skeleton graph and the distance weight graph have the same coordinate system and corresponding coordinates; In the skeleton distance weighted map, the value V of each non-zero pixel point (x, y) means the width of the chromosome at the coordinate point (x, y) in the second chromosome binarization map.

7. A chromosome profile optimization method according to claim 1, characterized in that: The single chromosome image with background removed is binarized by a fixed threshold value to obtain a first chromosome binarization image; the threshold value interval is 100-200.

8. An electronic device, characterized in that: The device includes a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or at least one program is loaded and executed by the processor to implement a chromosome profile optimization method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that The storage medium stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement a chromosome profile optimization method according to any one of claims 1 to 7.

Images