Image segmentation result evaluation method and device, storage medium and program product

Abstract

The invention relates to the technical field of medical treatment, in particular to an image segmentation result assessment method and device, a storage medium and a program product, which can improve the comprehensiveness and accuracy of image segmentation result assessment, and the method comprises the following steps: obtaining a spine segmentation image; the spine segmentation image comprises a slice image, and the slice image comprises at least one of a sagittal plane slice image, a coronal plane slice image and a horizontal plane slice image; based on the at least two evaluation items, evaluating the eligibility of the spine segmentation image; the evaluation items comprise a segmentation area evaluation item, a segmentation boundary continuity evaluation item, a segmentation layer continuity evaluation item and a segmentation volume evaluation item.

Description

TECHNICAL FIELD

[0001] The present application relates to the medical technology field, and in particular to an image segmentation result evaluation method and device, a storage medium and a program product. BACKGROUND

[0002] In the medical imaging field, image segmentation technology plays an extremely important role in disease diagnosis, treatment planning, and surgery navigation. With the rapid development of artificial intelligence technology, medical image segmentation technology based on AI algorithms has gradually developed. This image segmentation technology can quickly and automatically outline the spine region, providing a more efficient image analysis means for clinical treatment and significantly improving the efficiency of diagnosis and treatment.

[0003] Currently, research on spine CT image segmentation technology mostly focuses on the continuous optimization of the segmentation algorithm itself in order to improve the accuracy and speed of segmentation. However, in actual clinical applications, the accuracy of the segmentation results is directly related to the rationality and effectiveness of subsequent diagnosis and treatment plans. Once the segmentation results deviate or are incorrect, it may have a serious adverse effect on the treatment effect of the patient. Therefore, verifying the quality of AI segmentation results is the key to ensuring the safety and reliability of clinical applications.

[0004] Traditional verification methods mainly rely on manual annotation or expert evaluation, which requires high experience and consumes a large amount of time and manpower. It is difficult to achieve real-time and automated segmentation result verification, and the existing verification rules are relatively simple and cannot cope with complex spine morphology segmentation verification. There is a problem of poor verification accuracy, which cannot meet the needs of clinical image segmentation processing. SUMMARY

[0005] The purpose of the present application is to provide an image segmentation result evaluation method, device, storage medium and program product, which aims to improve the comprehensiveness and accuracy of image segmentation result evaluation.

[0006] In a first aspect, an image segmentation result evaluation method is provided, which includes: obtaining at least one spine segmentation image; the spine segmentation image includes a slice image, and the slice image includes at least one of the following: a sagittal slice image, a coronal slice image, and a horizontal slice image; based on at least two evaluation items, evaluating the quality of the spine segmentation image; the evaluation items include: a segmentation area evaluation item, a segmentation boundary continuity evaluation item, a segmentation layer continuity evaluation item, and a segmentation volume evaluation item.

[0007] In some embodiments, the spine segmentation image includes a plurality of sagittal slice images; and when the evaluation items include the segmentation area evaluation item, evaluating the quality of the spine segmentation image includes:

[0008] acquire areas of the spinal column segmentation regions in each of the sagittal slice images and sort the areas; and superimpose the sagittal slice images to obtain a superimposed image; the spinal column segmentation region in the superimposed image is an intersection of the spinal column segmentation regions in each of the superimposed sagittal slice images;

[0009] acquire a first slice image; the first slice image is a raw sagittal slice image corresponding to a sagittal slice image with a largest area of the spinal column segmentation region in the spinal column segmentation image in the original spinal column image;

[0010] determine the eligibility of the spinal column segmentation image in the segmentation area evaluation item based on an overlap degree of the spinal column segmentation region in the superimposed image and the spinal column region in the first slice image and an area ratio of the spinal column segmentation region in the superimposed image and the spinal column region in the first slice image.

[0011] In some embodiments, the determining the eligibility of the spinal column segmentation image in the segmentation area evaluation item based on the overlap degree of the spinal column segmentation region in the superimposed image and the spinal column region in the first slice image and the area ratio of the spinal column segmentation region in the superimposed image and the spinal column region in the first slice image comprises: determining that the spinal column segmentation image is eligible in the segmentation area evaluation item when the overlap degree of the spinal column segmentation region in the superimposed image and the spinal column region in the first slice image is greater than or equal to a first preset threshold and the area ratio of the spinal column segmentation region in the superimposed image and the spinal column region in the first slice image is within a first preset range; otherwise, determining that the spinal column segmentation image is ineligible in the segmentation area evaluation item.

[0012] In the above embodiments, the method further comprises: outputting the overlap degree of the spinal column segmentation region in the superimposed image and the spinal column region in the first slice image and the area ratio of the spinal column segmentation region in the superimposed image and the spinal column region in the first slice image.

[0013] In some embodiments, the spinal column segmentation image comprises a plurality of sagittal slice images; and when the evaluation item comprises a segmentation boundary continuity evaluation item, the evaluating the eligibility of the spinal column segmentation image comprises:

[0014] superimpose the sagittal slice images to obtain a superimposed image; the spinal column segmentation region in the superimposed image is an intersection of the spinal column segmentation regions in each of the superimposed sagittal slice images;

[0015] determine a boundary point set of the spinal column segmentation region in the superimposed image;

[0016] The boundary point set comprises at least one of: a left boundary point set, a right boundary point set, and a center point set.

[0017] calculate a distance difference of adjacent points in the boundary point set in a horizontal direction; the adjacent points are boundary points adjacent in a vertical direction.

[0018] The ratio of the number of distance differences within the preset error range to the total number of distance differences is recorded as a first number ratio. If the first number ratio is within a second preset range, it is determined that the spine segmentation image is qualified in the segmentation boundary continuity evaluation item; otherwise, it is determined that the spine segmentation image is unqualified in the segmentation boundary continuity evaluation item.

[0019] In some embodiments, the spine segmentation image includes a plurality of horizontal slice images. In the case where the evaluation item includes a segmentation layer continuity evaluation item, evaluating the qualification of the spine segmentation image includes: selecting a second image from the plurality of horizontal slice images, the second image not containing a spine segmentation region; and calculating a second number ratio of the number of second images in the plurality of horizontal slice images. If the second number ratio is within a third preset range, it is determined that the spine segmentation image is qualified in the segmentation layer continuity evaluation item; otherwise, it is determined that the spine segmentation image is unqualified in the segmentation layer continuity evaluation item.

[0020] In the above embodiments, the method further includes outputting the second number ratio and an index value of each second image.

[0021] In some embodiments, the spine segmentation image includes a sagittal slice image, a coronal slice image, and a horizontal slice image. In the case where the evaluation item includes a segmentation volume evaluation item, evaluating the qualification of the spine segmentation image includes:

[0022] performing voxel analysis on the spine segmentation image, and constructing a spine three-dimensional model based on at least one of the sagittal slice image, the coronal slice image, and the horizontal slice image;

[0023] determining a segmentation volume of the spine based on the spine three-dimensional model;

[0024] In the case where the segmentation volume of the spine is within a fourth preset range, it is determined that the spine segmentation image is qualified in the segmentation volume evaluation item; otherwise, it is determined that the spine segmentation image is unqualified in the segmentation volume evaluation item.

[0025] In some embodiments, the spine segmentation image includes any one of a sagittal slice image, a coronal slice image, and a horizontal slice image. In the case where the evaluation item includes a segmentation volume evaluation item, evaluating the qualification of the spine segmentation image includes: calculating the number of pixels with a pixel value of 1 in the spine segmentation image; in the case where the number of pixels with a pixel value of 1 is within a fifth preset range, it is determined that the spine segmentation image is qualified in the segmentation volume evaluation item; otherwise, it is determined that the spine segmentation image is unqualified in the segmentation volume evaluation item.

[0026] In a second aspect, an electronic device is provided, comprising: a processor and a memory configured to store processor-executable instructions; wherein the processor is configured to execute the instructions to implement any of the image segmentation result evaluation methods of the first aspect.

[0027] In a third aspect, a computer-readable storage medium is provided, the computer-readable storage medium storing instructions thereon, which when executed by a device, cause the device to perform any of the image segmentation result evaluation methods of the first aspect.

[0028] In a fourth aspect, a computer program product is provided, the computer program product comprising computer instructions, which when executed on a processor of a device, cause the device to perform any of the image segmentation result evaluation methods of the first aspect.

[0029] The image segmentation result evaluation method provided by the present application can verify the qualification of the spine segmentation image from multiple evaluation aspects (segmentation area, segmentation boundary continuity, segmentation layer continuity, and segmentation volume), realize comprehensive and objective quantitative evaluation, effectively reduce subjective judgment errors, accurately and quickly find segmentation problems and provide improvement directions, while ensuring the consistency and reliability of segmentation quality, and support optimization of spine segmentation algorithms and improvement of medical image segmentation technology. BRIEF DESCRIPTION OF DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0031] Figure 1 A flowchart of an image segmentation result evaluation method provided by an embodiment of the present application is shown in the figure.

[0032] Figure 2 A flowchart of another image segmentation result evaluation method provided by an embodiment of the present application is shown in the figure.

[0033] Figure 3 A flowchart of another image segmentation result evaluation method provided by an embodiment of the present application is shown in the figure.

[0034] Figure 4 A flowchart of another image segmentation result evaluation method provided by an embodiment of the present application is shown in the figure.

[0035] Figure 5 A flowchart of another image segmentation result evaluation method provided by an embodiment of the present application is shown in the figure.

[0036] Figure 6 A flowchart of another evaluation method of image segmentation results provided by an embodiment of the present application;

[0037] Figure 7 A structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0038] In the embodiments of the present application, the terms "first", "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features.

[0039] In the embodiments of the present application, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or includes elements inherent to such process, method, article or device. Without more limitations, the element defined by the sentence "including a…" does not exclude the presence of another identical element in the process, method, article or device including the element.

[0040] "A and / or B" includes the following three combinations: only A, only B, and the combination of A and B.

[0041] In the field of medical image segmentation, the application of AI algorithms is gradually mature, especially in the segmentation of spine CT images, AI technology can quickly and automatically delineate the spine region. However, the existing technology mainly focuses on the optimization of the segmentation algorithm itself, and pays less attention to the verification and evaluation of the segmentation results. In actual application, the AI algorithm cannot check whether its segmentation result is correct, and needs additional verification means to evaluate the reliability of the segmentation result. In clinical application, the accuracy of the segmentation result directly affects the effect of subsequent diagnosis and treatment, therefore, how to verify the qualification of the AI segmentation result has become a problem to be solved.

[0042] At present, the verification of segmentation results lacks automatic verification means, mainly relying on manual annotation and fixed rules for verification, which needs medical experts to manually evaluate the segmentation results, with high time cost, and most verification methods are based on the verification of geometric features of the segmentation results, which is difficult to adapt to complex morphological spine regions.

[0043] To solve the above technical problems, the present disclosure provides an image segmentation result evaluation method, which is characterized by: obtaining at least one spine segmentation image; the spine segmentation image comprises a slice image, and the slice image comprises at least one of a sagittal slice image, a coronal slice image and a horizontal slice image; and evaluating the eligibility of the spine segmentation image based on at least two evaluation items; the evaluation items include a segmentation area evaluation item, a segmentation boundary continuity evaluation item, a segmentation layer continuity evaluation item and a segmentation volume evaluation item. The method can improve the comprehensiveness and accuracy of the evaluation of the image segmentation result.

[0044] The image segmentation result evaluation method, device, storage medium and program product provided by the present application are described below.

[0045] As shown in Figure 1 The image segmentation result evaluation method comprises the following steps:

[0046] S101, obtaining at least one spine segmentation image.

[0047] The spine segmentation image comprises a slice image, and the slice image comprises at least one of a sagittal slice image, a coronal slice image and a horizontal slice image.

[0048] In some embodiments, the spine segmentation image is a three-dimensional spine image obtained by a spine segmentation result, and slice images in different directions (sagittal, coronal and horizontal) can be extracted from the three-dimensional image.

[0049] In some embodiments, the spine segmentation image is a mask image.

[0050] In some embodiments, the spine segmentation image is a three-dimensional spine image obtained by image segmentation of an original medical image (such as a CT image) based on an AI image segmentation algorithm.

[0051] S102, evaluating the eligibility of the spine segmentation image based on at least two evaluation items.

[0052] The evaluation items include a segmentation area evaluation item, a segmentation boundary continuity evaluation item, a segmentation layer continuity evaluation item and a segmentation volume evaluation item.

[0053] In some embodiments, the selection and combination of the evaluation items can be determined according to actual needs. For example, in the case that at least two evaluation items are qualified, it is determined that the spine segmentation image is qualified, or in the case that there is one evaluation item that is unqualified among all the evaluation items, it is determined that the spine segmentation image is unqualified.

[0054] It can be understood that the image segmentation result evaluation method provided by the application can verify the qualification of the spine segmentation image from multiple evaluation item perspectives (segmentation area, segmentation boundary continuity, segmentation layer continuity, and segmentation volume), realize comprehensive and objective quantitative evaluation, effectively reduce subjective judgment errors, accurately and quickly find segmentation problems and provide improvement directions, while ensuring the consistency and reliability of the segmentation quality, and support optimization of the spine segmentation algorithm and improvement of the medical image segmentation technology.

[0055] In some embodiments, as shown in FIG. 1, Figure 2 The spine segmentation image includes a plurality of sagittal slice images, and in the case where the evaluation item includes the segmentation area evaluation item, evaluating the qualification of the spine segmentation image includes the following steps S201-S204:

[0056] S201, obtaining the area of the spine segmentation region in each sagittal slice image and sorting.

[0057] In some embodiments, before processing each sagittal slice image, an index value is added to each continuous sagittal slice image obtained in advance for subsequent analysis, and the same index value is added to the sagittal slice image of the original spine image corresponding to the spine segmentation image.

[0058] S202, superimposing the sagittal slice images to obtain a superimposed image.

[0059] The spine segmentation region in the superimposed image is the intersection of the spine segmentation regions in each of the superimposed sagittal slice images.

[0060] In some embodiments, since the spine is only a part of human tissue, not all sagittal slice images contain spine tissue. In order to reduce the consumption of computing resources, a part of the sagittal slice images can be selected for superimposition according to the area sorting result in S201. For example, according to the area sorting result, the sagittal slice images with the top 60% of the area of the spine segmentation region are selected for superimposition.

[0061] In some embodiments, in order to make the spine segmentation region in the superimposed image more accurate, the superimposed image can be preprocessed, including binaryzation processing, morphological operation (erosion, dilation), and connected domain processing.

[0062] For example, the superimposed image is subjected to a binarization operation, so that the target region (spine segmentation region) appears bright and other non-target regions become dark regions, thereby highlighting the target region. Then, a morphological erosion operation is performed on the binarized matrix using a cross-shaped (MORPH_CROSS) structuring element with a kernel size of 1x1 to remove the boundary pixels of the target region and eliminate small noise points. Then, a morphological dilation operation is performed on the matrix using the same cross-shaped structuring element to expand the boundary pixels of the target region, thereby restoring the eroded part and enhancing the connectivity of the target region. Then, a connected component analysis is performed on the image subjected to the morphological operation to label different connected regions and generate a label matrix, obtain the total number of connected regions, and count the area of each connected region. The connected regions are sorted in descending order of area. Finally, the two connected regions with the largest areas (excluding the background) are selected to construct a final binary mask image, i.e., the final superimposed image.

[0063] S203, acquire a first slice image.

[0064] The first slice image is a raw sagittal slice image corresponding to the sagittal slice image with the largest area in the spine segmentation region in the spine segmentation image.

[0065] In some embodiments, the sagittal slice with the largest area of the spine segmentation region is determined based on the area sorting result in S201, and then the raw sagittal slice with the same index value in the original spine image is determined as the first slice image according to the index value of the sagittal slice.

[0066] In some embodiments, the spine region in the first slice image is obtained by performing a preprocessing operation on the first slice image, including binarization processing, morphological operation (erosion, dilation), connected component processing, etc.

[0067] S204, determine the eligibility of the spine segmentation image in the segmentation area evaluation item based on the overlap between the spine segmentation region in the superimposed image and the spine region in the first slice image, and the area ratio of the spine segmentation region in the superimposed image to the spine region in the first slice image.

[0068] In some embodiments, the superimposed image and the first slice image are aligned and superimposed, so that the overlap between the spine segmentation region in the superimposed image and the spine region in the first slice image, and the area ratio of the spine segmentation region in the superimposed image to the spine region in the first slice image are obtained based on the superimposed image.

[0069] The overlap between the spine segmentation region in the superimposed image and the spine region in the first slice image is determined based on the ratio of the intersection to the union of the spine segmentation region in the superimposed image and the spine region in the first slice image in the superimposed image.

[0070] In some embodiments, S204 can be specifically implemented as: determining that the spinal column segmentation image is qualified in the segmentation area evaluation item, in a case that the overlap degree of the spinal column segmentation region in the superimposed image and the spinal column region in the first slice image is greater than or equal to a first preset threshold, and the area ratio of the spinal column segmentation region in the superimposed image and the spinal column region in the first slice image is within a first preset range; otherwise, determining that the spinal column segmentation image is unqualified in the segmentation area evaluation item.

[0071] For example, the first preset threshold is 80%, and the first preset range is 80%-120%. The specific values of the first preset threshold and the first preset range are not limited in the present application, and can be determined according to relevant medical standards and clinical experience in actual application.

[0072] In some embodiments, after determining the qualification of the spinal column segmentation image in the segmentation area evaluation item, the segmentation result (qualified or unqualified) of the segmentation area evaluation item, and the overlap degree and the area ratio of the spinal column segmentation region in the superimposed image and the spinal column region in the first slice image can also be output. By analyzing the output segmentation result, overlap degree and area ratio, reverse optimization can be performed, such as optimizing the image segmentation algorithm, adjusting the treatment plan, etc.

[0073] It can be understood that by comparing the superimposed image and the first slice image, the similarity of the shape of the spinal column segmentation region after image segmentation and the original spinal column region can be verified. At the same time, the overlap degree and the area ratio can quantify the similarity of the shape of the spinal column region in the superimposed image and the original image, thereby improving the accuracy and consistency of the evaluation result.

[0074] In some embodiments, as shown in FIG. 3, the spinal column segmentation image includes a plurality of sagittal slice images, and in a case that the evaluation item includes a segmentation boundary continuity evaluation item, evaluating the qualification of the spinal column segmentation image includes the following steps S301-S304: Figure 3

[0075] S301, superimpose the sagittal slice images to obtain a superimposed image.

[0076] The spinal column segmentation region in the superimposed image is the intersection of the spinal column segmentation regions in each of the superimposed sagittal slice images.

[0077] In some embodiments, in order to make the spinal column segmentation region in the superimposed image more accurate, the superimposed image can be preprocessed, including binaryzation processing, morphological operation (erosion, dilation), connected domain processing.

[0078] S302, determine a boundary point set of the spinal column segmentation region in the superimposed image.

[0079] ​The boundary point set includes at least one of the following: a left boundary point set, a right boundary point set, and a center point (centroid point) set.

[0080] S303, calculate a distance difference of adjacent points in the boundary point set in a horizontal direction.

[0081] The adjacent points are boundary points adjacent in a vertical direction.

[0082] S304, calculate a proportion of the number of distance differences within a preset error range in all distance differences, denoted as a first proportion of the number, and if the first proportion of the number is within a second preset range, determine that the segmented spinal column image is qualified in the segmentation boundary continuity evaluation item; otherwise, determine that the segmented spinal column image is unqualified in the segmentation boundary continuity evaluation item.

[0083] It should be noted that, due to the curvature of the physiological structure of the spine, there is a certain distance difference between adjacent boundary points in the horizontal direction. When the distance difference is within a reasonable range, the segmentation boundary of the segmented spinal column image is continuous, and conversely, if the distance difference exceeds the reasonable range, the segmentation boundary of the segmented spinal column image is discontinuous.

[0084] The application does not specifically limit the reasonable range of the distance difference. For example, the reasonable distance difference range can be 0.2-0.4.

[0085] In some embodiments, due to the similarity in spatial position between the left boundary points and the right boundary points of the spinal column in the sagittal plane slice image, in order to reduce the calculation cost, one of the left boundary point set or the right boundary point set can be used together with the center point set as data for determining the qualification of the segmented spinal column image in the segmentation boundary continuity evaluation item. The first proportions of the number of distance differences within a preset error range in all distance differences are respectively calculated in the two sets, and if both of the two first proportions of the number are within a second preset range, it is determined that the segmented spinal column image is qualified in the segmentation boundary continuity evaluation item; otherwise, it is determined that the segmented spinal column image is unqualified in the segmentation boundary continuity evaluation item.

[0086] In some embodiments, the segmented spinal column image includes a plurality of horizontal plane slice images, and the evaluation item further includes a segmentation layer continuity evaluation item. Since the spinal column presents a regular arrangement and morphology in the horizontal plane, the qualification of the segmented spinal column image in the segmentation layer continuity evaluation item can be evaluated based on the continuity of the horizontal plane slice images in the segmented spinal column image.

[0087] Specifically, as shown in FIG. 4, the method includes the following steps S401-S402: Figure 4

[0088] S401, selecting a second image from a plurality of horizontal plane slice images. ​

[0089] The second image is an image without the spinal column segmentation region.

[0090] In some embodiments, according to actual needs, the multiple horizontal slice images can be all the horizontal slice images obtained by a single spinal column segmentation, or can be a part of the horizontal slice images obtained by a single spinal column segmentation.

[0091] In some embodiments, since the spinal column segmentation image is a mask image, in the slice image (sagittal slice image, coronal slice image or horizontal slice image) in each direction of the spinal column segmentation image, the pixel value of the spinal column segmentation region is 1, and the pixel value without the spinal column segmentation region is 0. Based on this, the slice image (sagittal slice image, coronal slice image or horizontal slice image) in each direction of the spinal column segmentation image can contain the case that all pixel values are 0, i.e., the slice image does not contain the spinal column region.

[0092] S402, count the proportion of the number of the second image in the multiple horizontal slice images, denoted as a second number proportion, and when the second number proportion is within a third preset range, determine that the spinal column segmentation image is qualified in the segmentation layer continuity evaluation item; otherwise, determine that the spinal column segmentation image is unqualified in the segmentation layer continuity evaluation item.

[0093] In some embodiments, the third preset range can be determined according to a large amount of known qualified spinal column segmentation image data, and the normal range of the second number proportion is counted, so as to determine a suitable third preset range. For example, the third preset range can be less than 5%, or can be less than 10%.

[0094] In some embodiments, the amount of image data without the spinal column segmentation region in the continuous N layers and above in the horizontal slice image is counted, and the counted number is taken as the number of the second image. Wherein, N is a positive integer.

[0095] For example, taking N as 3, if there are a total of nine continuous spinal column segmentation images with index values of 1, 2, 3, 4, 5, 6, 7, 8 and 9, and the images with index values of 1, 2, 3, 5, 7 and 8 are images without the spinal column segmentation region, then the images with index values of 1, 2 and 3 are the number of the second image, i.e., the number of the second image is 1, and the proportion of the second image in all horizontal slice images is 1 / 3.

[0096] For example, taking N as 1, if there are a total of nine continuous spinal column segmentation images with index values of 1, 2, 3, 4, 5, 6, 7, 8 and 9, and the images with index values of 1, 3, 5 and 8 are images without the spinal column segmentation region, then the number of the second image is 4, and the proportion of the second image in all horizontal slice images is 4 / 9.

[0097] In some embodiments, after determining the eligibility of the spine segmentation image in the segmentation layer continuity evaluation item, the result of the evaluation (pass or fail) is output, as well as the second proportion and the index value of each second image, so that subsequent analysis can be performed according to these data, such as optimizing the image segmentation algorithm, adjusting the treatment plan, etc.

[0098] It should be noted that once the spine segmentation image exists the second image, it indicates that the spine segmentation image has a fault or image missing condition, and when the proportion of the number of slice images with fault or image missing in the total number of slice images exceeds the third preset range, it indicates that the continuity of the spine segmentation image is poor, and the reliability of the segmentation result is low, which may affect the subsequent clinical analysis. At this time, the segmentation algorithm needs to be optimized or adjusted to improve the segmentation quality.

[0099] In some embodiments, the spine segmentation image can also reflect the volume characteristics of the spine, and based on the spine segmentation image, the eligibility of the spine segmentation image can be evaluated from the perspective of the segmentation volume evaluation item.

[0100] A possible implementation can construct a spine three-dimensional model based on the spine segmentation image to evaluate the eligibility of the spine segmentation image in the segmentation volume evaluation item.

[0101] Specifically, as shown in FIG. 5, Figure 5 The method comprises the following steps S501-S503:

[0102] S501, voxel analysis is performed on the spine segmentation image, and a spine three-dimensional model is constructed based on at least one of the sagittal slice image, the coronal slice image and the horizontal slice image.

[0103] In some embodiments, a suitable three-dimensional reconstruction algorithm is selected to construct the spine three-dimensional model. For example, a surface extraction algorithm can be used to construct the spine three-dimensional model.

[0104] S502, based on the spine three-dimensional model, the segmentation volume of the spine is determined.

[0105] In some embodiments, the number of voxels belonging to the spinal tissue is counted by counting the voxels in the constructed spine three-dimensional model. Since each voxel has a fixed volume (the volume of a voxel is equal to the cube of its size in three-dimensional space, i.e. the product of the length, width and height of the voxel in the x, y and z axis directions), the segmentation volume of the spine is determined by multiplying the number of voxels by the volume of a single voxel.

[0106] In some embodiments, the volume of the spine is directly measured by using a measurement tool in professional three-dimensional modeling software or medical image analysis software, selecting a suitable measurement function and measurement parameters, so as to determine the segmentation volume of the spine.

[0107] S503, in the case that the segmentation volume of the spine is in the fourth preset range, it is determined that the spine segmentation image is qualified in the segmentation volume evaluation item; otherwise, it is determined that the spine segmentation image is unqualified in the segmentation volume evaluation item.

[0108] In some embodiments, the fourth preset range is set in advance based on different population types. For example, the population types can include male, female, children, etc. The fourth preset range is not limited in the present application, and in actual application, the first preset range can be set according to relevant medical standards and clinical experience, etc.

[0109] Another possible implementation can evaluate the qualification of the spine segmentation image in the segmentation volume evaluation item based on the number of pixels belonging to the spinal tissue in the spine segmentation image.

[0110] Specifically, as shown in FIG. 6, Figure 6 includes the following steps S601-S602:

[0111] S601, the number of pixels with a pixel value of 1 in the spine segmentation image is calculated.

[0112] The spine segmentation image includes sagittal slice images, coronal slice images and horizontal slice images, and one of them can be selected for segmentation volume evaluation according to specific needs.

[0113] S602, in the case that the number of pixels with a pixel value of 1 is in the fifth preset range, it is determined that the spine segmentation image is qualified in the segmentation volume evaluation item; otherwise, it is determined that the spine segmentation image is unqualified in the segmentation volume evaluation item.

[0114] In some embodiments, the fifth preset range is set in advance based on different population types. For example, the population types can include male, female, children, etc. The fifth preset range is not limited in the present application, and in actual application, the second preset range can be set according to relevant medical standards and clinical experience, etc.

[0115] The image segmentation result evaluation method in the embodiments of the present application can be applied to the processing device 101 in the following image segmentation result evaluation system.

[0116] Figure 7 A structural schematic diagram of an electronic device is provided in the embodiments of the present application. As shown in FIG. 7, Figure 7 The electronic device 700 includes but is not limited to a processor 701 and a memory 702.

[0117] The memory 702 is configured to store executable instructions of the processor 701. It can be understood that the processor 701 is configured to execute the instructions to implement the radiotherapy control method in the above embodiments.

[0118] It should be noted that those skilled in the art can understand that the structure of the electronic device 700 shown in the above embodiments is not a limitation to the electronic device 700. The electronic device 700 can include more or less components than those shown in the above embodiments, or combine some components, or arrange different components. Figure 7 The electronic device 700 shown in the above embodiments is not a limitation to the electronic device 700. The electronic device 700 can include more or less components than those shown in the above embodiments, or combine some components, or arrange different components. Figure 7 The electronic device 700 shown in the above embodiments is not a limitation to the electronic device 700. The electronic device 700 can include more or less components than those shown in the above embodiments, or combine some components, or arrange different components.

[0119] The processor 701 is a control center of the electronic device 700, which connects all parts of the electronic device 700 through various interfaces and lines, executes software programs and / or modules stored in the memory 702 and data stored in the memory 702, processes data, and thus monitors the whole electronic device 700. The processor 701 can include one or more processing units. Optionally, the processor 701 can integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communication. It can be understood that the modem processor can also not be integrated into the processor 701.

[0120] In some embodiments, the processor 701 is configured to obtain at least one spine segmentation image, wherein the spine segmentation image includes a slice image, and the slice image includes at least one of a sagittal slice image, a coronal slice image, and a horizontal slice image.

[0121] In some embodiments, the processor 701 is further configured to evaluate the eligibility of the spine segmentation image based on at least two evaluation items, wherein the evaluation items include a segmentation area evaluation item, a segmentation boundary continuity evaluation item, a segmentation layer continuity evaluation item, and a segmentation volume evaluation item.

[0122] The memory 702 can be configured to store software programs and various data. The memory 702 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs (such as a determination unit, a processing unit, etc.) required by at least one functional module, and the like. In addition, the memory 702 can include a high-speed random access memory, and can also include a non-volatile memory, for example, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state memory device.

[0123] In an example embodiment, a computer-readable storage medium including instructions, for example, the memory 702 including instructions, is also provided, which can be executed by the processor 701 of the electronic device 700 to implement the method in the above-described embodiments.

[0124] Optionally, the computer-readable storage medium can be a non-transitory computer-readable storage medium, for example, the non-transitory computer-readable storage medium can be a Read-Only Memory (ROM), a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0125] In an example embodiment, the embodiments of the present application also provide a computer program product including one or more instructions, which can be executed by the processor 701 of the electronic device 700 to complete the method in the above-described embodiments.

[0126] It should be noted that the instructions in the above computer-readable storage medium or the one or more instructions in the computer program product are executed by the processor of the electronic device to realize each process of the above method embodiments, and can achieve the same technical effects as the above method. To avoid repetition, it will not be repeated here.

[0127] Through the description of the above embodiments, those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-described division of each functional module is taken as an example for illustration. In actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete the above-described full classification or part of the function.

[0128] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the above-described device embodiments are only illustrative. For example, the division of modules or units is only a logical function division. In actual implementation, another division manner can be used, for example, a plurality of units or components can be combined or integrated into another device, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed each other can be through some interface, indirect coupling or communication connection between devices or units, which can be electrical, mechanical or other forms.

[0129] The units described as separate components may or may not be physically separate, and the components displayed as units may be a physical unit or multiple physical units, that is, may be located in one place, or also can be distributed to multiple different places. Part or all of the classified units can be selected according to actual needs to achieve the purpose of the embodiment of the present application.

[0130] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present alone, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0131] If the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present application essentially or the part that contributes to the prior art or the whole classification or part of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions to make a device (which can be a single chip, a chip, etc.) or a processor execute all or part of the steps of the method of the embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, ROM, RAM, magnetic disk or optical disk and various program codes that can store program codes.

[0132] In the description of the embodiments of the present application, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0133] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method of evaluating an image segmentation result, characterized by, The method comprises: obtaining a spine segmentation image; the spine segmentation image comprises slice images, and the slice images comprise at least one of the following: sagittal slice images, coronal slice images and horizontal slice images; based on at least two evaluation items, evaluating the eligibility of the spine segmentation image; the evaluation items include: segmentation area evaluation item, segmentation boundary continuity evaluation item, segmentation layer continuity evaluation item and segmentation volume evaluation item.

2. The method of claim 1, wherein, The spine segmentation image comprises a plurality of sagittal slice images; in the case where the evaluation items include the segmentation area evaluation item, the evaluation of the eligibility of the spine segmentation image comprises: obtaining the area of the spine segmentation region in each sagittal slice image and sorting; superimposing the sagittal slice images to obtain a superimposed image; the spine segmentation region in the superimposed image is the intersection of the spine segmentation region in each of the superimposed sagittal slice images; obtaining a first slice image; the first slice image is the original sagittal slice image corresponding to the sagittal slice image with the largest area of the spine segmentation region in the spine segmentation image in the original unsegmented spine image; based on the overlap of the spine segmentation region in the superimposed image and the spine region in the first slice image, and the area ratio of the spine segmentation region in the superimposed image to the spine region in the first slice image, determine the eligibility of the spine segmentation image in the segmentation area evaluation item.

3. The method of claim 2, wherein, The method further comprises: outputting the overlap of the spine segmentation region in the superimposed image and the spine region in the first slice image, and the area ratio of the spine segmentation region in the superimposed image to the spine region in the first slice image.

4. The method according to claim 2 or 3, characterized in that, The spine segmentation image comprises a plurality of sagittal slice images; in the case where the evaluation items include the segmentation boundary continuity evaluation item, the evaluation of the eligibility of the spine segmentation image comprises: superimposing the sagittal slice images to obtain a superimposed image; the spine segmentation region in the superimposed image is the intersection of the spine segmentation region in each of the superimposed sagittal slice images; 5. The method of claim 1, wherein, determining a set of boundary points of the spine segmentation region in the superimposed image; wherein the set of boundary points comprises at least one of the following: a left set of boundary points, a right set of boundary points, a center point set; ​ ​ Calculate a distance difference of adjacent points in the boundary point set in a horizontal direction, wherein the adjacent points are adjacent boundary points in a vertical direction; Calculate a distance difference of adjacent points in the boundary point set in a horizontal direction, wherein the adjacent points are adjacent boundary points in a vertical direction; 6. The method of claim 1, wherein, Calculate a distance difference of adjacent points in the boundary point set in a horizontal direction, wherein the adjacent points are adjacent boundary points in a vertical direction; The spine segmentation image includes a plurality of horizontal plane slice images; in the case that the evaluation item includes the segmentation layer continuity evaluation item, the method for evaluating the qualification of the spine segmentation image comprises: Filtering a second image from the plurality of horizontal plane slice images, wherein the second image does not contain a spine segmentation region; 7. The method of claim 6, wherein, Calculate a second quantity ratio of the second image in the plurality of horizontal plane slice images, and determine that the spine segmentation image is qualified in the segmentation layer continuity evaluation item when the second quantity ratio is within a third preset range; otherwise, determine that the spine segmentation image is unqualified in the segmentation layer continuity evaluation item. The method further comprises:

8. The method of claim 1, wherein, Output the second quantity ratio and an index value of each second image. The spine segmentation image includes the sagittal plane slice image, the coronal plane slice image, and the horizontal plane slice image; in the case that the evaluation item includes the segmentation volume evaluation item, the method for evaluating the qualification of the spine segmentation image comprises: Perform voxel analysis on the spine segmentation image, and construct a spine three-dimensional model based on at least one of the sagittal plane slice image, the coronal plane slice image, and the horizontal plane slice image; Determine a segmentation volume of the spine based on the spine three-dimensional model; 9. The method of claim 1, wherein, Determine that the spine segmentation image is qualified in the segmentation volume evaluation item when the segmentation volume of the spine is within a fourth preset range; otherwise, determine that the spine segmentation image is unqualified in the segmentation volume evaluation item. The spine segmentation image includes any one of the sagittal plane slice image, the coronal plane slice image, and the horizontal plane slice image; in the case that the evaluation item includes the segmentation volume evaluation item, the method for evaluating the qualification of the spine segmentation image comprises: Calculate the number of pixels with a pixel value of 1 in the spine segmentation image; 10. An electronic device, comprising: Determine that the spine segmentation image is qualified in the segmentation volume evaluation item when the number of pixels with the pixel value of 1 is within a fifth preset range; otherwise, determine that the spine segmentation image is unqualified in the segmentation volume evaluation item. The electronic device comprises: a processor; a memory configured to store instructions executable by the processor; 11. A computer readable storage medium, characterized in that, wherein the processor is configured to execute the instructions to implement the method of any one of claims 1 to 9. The computer readable storage medium stores computer instructions, when the computer instructions run on a computer, make the computer execute the method of any one of claims 1 to 9.

12. A computer program product, characterised in that, The computer program product comprises computer instructions which, when run on an electronic device, cause the electronic device to perform the method of any one of claims 1 to 9.

Images