Target organ contouring method and related products

Abstract

The application discloses a target organ contour segmentation method and related products. The method comprises the following steps: acquiring a to-be-processed image, wherein the to-be-processed image comprises a target organ; extracting a first image containing high-frequency information and a second image containing low-frequency information from the to-be-processed image; performing up-sampling processing on the second image to obtain a third image; fusing the first image and the third image to obtain a fourth image; and segmenting the contour of the target organ in the to-be-processed image according to the fourth image to obtain a first segmentation image.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method for segmenting the contour of a target organ and related products. Background Technology

[0002] By performing feature extraction on the image to be processed, a feature image can be obtained. This feature image can then be used for image processing tasks such as image segmentation, object detection, and image classification. Current methods use downsampling of the image to be processed to achieve feature extraction. However, the feature image obtained by this method does not carry enough information, resulting in low accuracy of the segmentation results obtained based on this feature image. Summary of the Invention

[0003] This application provides a method for target organ contour segmentation and related products.

[0004] Firstly, a method for target organ contour segmentation is provided, the method comprising:

[0005] Acquire an image to be processed, the image including the target organ;

[0006] Extract a first image containing high-frequency information and a second image containing low-frequency information from the image to be processed;

[0007] The second image is upsampled to obtain the third image;

[0008] The first image and the third image are fused to obtain a fourth image;

[0009] Based on the fourth image, the contour of the target organ in the image to be processed is segmented to obtain a first segmented image.

[0010] In any embodiment of this application, extracting a first image containing high-frequency information and a second image containing low-frequency information from the image to be processed includes:

[0011] The image to be processed is subjected to n levels of target processing to obtain the second image; the output image of any level of target processing in the n levels of target processing includes a high-frequency image containing high-frequency information and a low-frequency image containing low-frequency information;

[0012] The upsampling process of the second image to obtain the third image includes:

[0013] The second image is subjected to m-level upsampling processing to obtain the third image; the resolution of the output image of the i-th level upsampling processing in the m-level upsampling processing and the resolution of the output image of the j-th level target processing in the n-level target processing are both the target resolution, where i is a positive integer less than m and j is a positive integer less than or equal to n;

[0014] In the high-frequency image output by the n-level target processing, the image with a resolution lower than the target resolution is the target image.

[0015] The input image of the (i+1)th level upsampling process in the m-level upsampling process includes: the output image of the i-th level upsampling process, the low-frequency image output by the j-th level target processing, and the target image.

[0016] In any embodiment of this application, the high-frequency image includes a first high-frequency image and a second high-frequency image, wherein the direction of the edge represented by the first high-frequency image is different from the direction of the edge represented by the second high-frequency image.

[0017] In conjunction with any embodiment of this application, after obtaining the first segmented image, the method further includes:

[0018] Determine the concave curvature of the contour of the target organ in the first segmented image;

[0019] If the concave curvature is less than a first threshold, the first segmented image is determined to be an incorrect segmentation result;

[0020] If the concave curvature is greater than or equal to the first threshold, the first segmented image is determined to be the segmentation result of the image to be processed.

[0021] In conjunction with any embodiment of this application, after obtaining the first segmented image, the method further includes:

[0022] Determine the convex curvature of the contour of the target organ in the first segmented image;

[0023] If the curvature of the convex surface is greater than the second threshold, the first segmented image is determined to be an incorrect segmentation result;

[0024] If the convex curvature is less than or equal to the second threshold, the first segmented image is determined to be the segmentation result of the image to be processed.

[0025] In conjunction with any embodiment of this application, after determining that the first segmented image is the segmentation result of the image to be processed, the method further includes:

[0026] The first pixel in the first segmented image is identified as the target organ and is located outside the outline of the target organ.

[0027] Remove the first pixel from the first segmented image to obtain the second segmented image.

[0028] In conjunction with any embodiment of this application, after determining that the first segmented image is the segmentation result of the image to be processed, the method further includes:

[0029] Determine a second pixel from the first segmented image that is semantically not the target organ and is located within the outline of the target organ;

[0030] The semantics of the second pixel in the first segmented image are updated to the target organ to obtain the third segmented image.

[0031] Secondly, a target organ contour segmentation device is provided, the contour segmentation device comprising:

[0032] An acquisition unit is used to acquire an image to be processed, the image to be processed including a target organ;

[0033] An extraction unit is used to extract a first image containing high-frequency information and a second image containing low-frequency information from the image to be processed.

[0034] An upsampling processing unit is used to perform upsampling processing on the second image to obtain a third image;

[0035] A fusion unit is used to fuse the first image and the third image to obtain a fourth image;

[0036] The segmentation unit is used to segment the contour of the target organ in the image to be processed based on the fourth image to obtain a first segmented image.

[0037] In conjunction with any embodiment of this application, the extraction unit is used for:

[0038] The image to be processed is subjected to n levels of target processing to obtain the second image; the output image of any level of target processing in the n levels of target processing includes a high-frequency image containing high-frequency information and a low-frequency image containing low-frequency information;

[0039] The upsampling process of the second image to obtain the third image includes:

[0040] The second image is subjected to m-level upsampling processing to obtain the third image; the resolution of the output image of the i-th level upsampling processing in the m-level upsampling processing and the resolution of the output image of the j-th level target processing in the n-level target processing are both the target resolution, where i is a positive integer less than m and j is a positive integer less than or equal to n;

[0041] In the high-frequency image output by the n-level target processing, the image with a resolution lower than the target resolution is the target image.

[0042] The input image of the (i+1)th level upsampling process in the m-level upsampling process includes: the output image of the i-th level upsampling process, the low-frequency image output by the j-th level target processing, and the target image.

[0043] In any embodiment of this application, the high-frequency image includes a first high-frequency image and a second high-frequency image, wherein the direction of the edge represented by the first high-frequency image is different from the direction of the edge represented by the second high-frequency image.

[0044] In conjunction with any embodiment of this application, the segmentation unit is further configured to:

[0045] Determine the concave curvature of the contour of the target organ in the first segmented image;

[0046] If the concave curvature is less than a first threshold, the first segmented image is determined to be an incorrect segmentation result;

[0047] If the concave curvature is greater than or equal to the first threshold, the first segmented image is determined to be the segmentation result of the image to be processed.

[0048] In conjunction with any embodiment of this application, the segmentation unit is further configured to:

[0049] Determine the convex curvature of the contour of the target organ in the first segmented image;

[0050] If the curvature of the convex surface is greater than the second threshold, the first segmented image is determined to be an incorrect segmentation result;

[0051] If the convex curvature is less than or equal to the second threshold, the first segmented image is determined to be the segmentation result of the image to be processed.

[0052] In conjunction with any embodiment of this application, the segmentation unit is further configured to:

[0053] The first pixel in the first segmented image is identified as the target organ and is located outside the outline of the target organ.

[0054] Remove the first pixel from the first segmented image to obtain the second segmented image.

[0055] In conjunction with any embodiment of this application, the segmentation unit is further configured to:

[0056] Determine a second pixel from the first segmented image that is semantically not the target organ and is located within the outline of the target organ;

[0057] The semantics of the second pixel in the first segmented image are updated to the target organ to obtain the third segmented image.

[0058] Thirdly, an electronic device is provided, characterized in that it comprises: a processor and a memory, the memory being used to store computer program code, the computer program code including computer instructions, wherein, when the processor executes the computer instructions, the electronic device performs a method as described in the first aspect above and any possible implementation thereof.

[0059] Fourthly, another electronic device is provided, comprising: a processor, a transmitting device, an input device, an output device, and a memory, the memory being used to store computer program code, the computer program code including computer instructions, wherein, when the processor executes the computer instructions, the electronic device performs the method as described in the first aspect above and any possible implementation thereof.

[0060] Fifthly, a computer-readable storage medium is provided, wherein a computer program is stored therein, the computer program including program instructions that, when executed by a processor, cause the processor to perform a method as described in the first aspect above and any possible implementation thereof.

[0061] In a sixth aspect, a computer program product is provided, the computer program product comprising a computer program or instructions, wherein, when the computer program or instructions are executed on a computer, the computer performs the method described in the first aspect and any possible implementation thereof.

[0062] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this application.

[0063] In this embodiment, the third image is obtained by upsampling the second image, and both the second and third images contain low-frequency information of the image to be processed. Since the first image contains high-frequency information of the image to be processed, the contour segmentation device can fuse the high-frequency and low-frequency information of the image to be processed by fusing the first and third images. Thus, the fourth image contains both the low-frequency and high-frequency information of the image to be processed.

[0064] In other words, the contour segmentation device extracts low-frequency information from the image to be processed to obtain a second image, thus achieving feature extraction and obtaining a second image containing the features of the image to be processed; that is, the second image is the feature image of the image to be processed. Then, by upsampling the second image, its resolution is increased to obtain a third image, thereby increasing the resolution of the feature image of the image to be processed. However, because the contour segmentation device discards high-frequency information in the image to be processed during the extraction of low-frequency information to obtain the second image, the third image lacks high-frequency information from the image to be processed. Therefore, the contour segmentation device can enrich the information in the third image by fusing the first and third images to obtain a fourth image.

[0065] Therefore, the contour segmentation device segments the contour of the target organ in the image to be processed based on the fourth image to obtain the first segmented image, which can improve the accuracy of the first segmented image. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments of this application or the background art, the accompanying drawings used in the embodiments of this application or the background art will be described below.

[0067] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application.

[0068] Figure 1 A flowchart illustrating a target organ contour segmentation method provided in an embodiment of this application;

[0069] Figure 2 This is a schematic diagram of a feature extraction structure provided in an embodiment of this application;

[0070] Figure 3 This is a schematic diagram of another post-processing procedure provided in an embodiment of this application;

[0071] Figure 4 A flowchart for determining whether a segmented image to be confirmed is correct using a curvature algorithm is provided in this application embodiment;

[0072] Figure 5 A statistical schematic diagram of the average concave curvature of a kidney outline provided for an embodiment of this application;

[0073] Figure 6 A schematic diagram illustrating the display effect of a final segmentation result provided in an embodiment of this application;

[0074] Figure 7 A schematic diagram showing the display effect of another final segmentation result provided in an embodiment of this application;

[0075] Figure 8 This is a schematic diagram of the structure of a contour segmentation device provided in an embodiment of this application;

[0076] Figure 9 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0077] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0078] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0079] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0080] The execution subject of this application embodiment is a target organ contour segmentation device (hereinafter referred to as contour segmentation device), wherein the image device can be any electronic device capable of executing the technical solution disclosed in the method embodiment of this application. Optionally, the image device can be one of the following: a computer, a server.

[0081] It should be understood that the method embodiments of this application can also be implemented by a processor executing computer program code. The embodiments of this application are described below with reference to the accompanying drawings. Please refer to... Figure 1 , Figure 1 This is a flowchart illustrating an image processing method provided in an embodiment of this application.

[0082] 101. Obtain the image to be processed.

[0083] In this embodiment, the image to be processed may include any content. For example, the image to be processed may include a kidney, a human face, or a vehicle. The image to be processed may be acquired by any imaging device. For example, the image to be processed may be an ultrasound image acquired by an ultrasound probe, or an image acquired by a camera.

[0084] In this embodiment of the application, the image to be processed includes a target organ, which can be any organ. For example, the target organ is a kidney, or for another example, a lung. Optionally, when the image to be processed includes a target organ, the image to be processed is an ultrasound image, that is, an ultrasound image including the target organ, for example, an ultrasound image including the kidney obtained by scanning the kidney with an ultrasound probe.

[0085] In one implementation of acquiring an image to be processed, the contour segmentation device and the imaging device have a communication connection. The contour segmentation device acquires the image acquired by the imaging device through this communication connection as the image to be processed.

[0086] In another implementation of acquiring the image to be processed, the contour segmentation device receives the image to be processed from the user through input components, wherein the input components include: mouse, keyboard, touch screen, touchpad, and audio input device.

[0087] In another implementation of acquiring the image to be processed, the contour segmentation device receives the image to be processed from a receiving terminal, which may include: a mobile phone, a computer, a tablet computer, or a smart wearable device.

[0088] 102. Extract a first image containing high-frequency information and a second image containing low-frequency information from the above images to be processed.

[0089] In this embodiment, high-frequency information refers to high-frequency signals in the image to be processed, and low-frequency information refers to low-frequency signals in the image to be processed. Optionally, high-frequency information includes information with drastic grayscale changes in the image to be processed. For example, high-frequency information in the image to be processed may include edge information, or texture information. Low-frequency information includes information about pixel regions with gradual grayscale changes in the image to be processed. For example, low-frequency information in the image to be processed may include color information, or image content.

[0090] In this embodiment, the contour segmentation device obtains a first image by extracting high-frequency information from the image to be processed, meaning the first image contains the high-frequency information of the image to be processed. The contour segmentation device obtains a second image by extracting low-frequency information from the image to be processed, meaning the second image contains the low-frequency information of the image to be processed.

[0091] In one possible implementation, the contour segmentation device extracts a first image containing high-frequency information and a second image containing low-frequency information from the image to be processed by performing wavelet transform on the image to be processed.

[0092] Optionally, the contour segmentation device performs wavelet transform on the image to be processed to extract a first image containing high-frequency information, and extracts high-frequency information in different directions to obtain different high-frequency images. In this case, the first image includes all high-frequency images. For example, the contour segmentation device extracts high-frequency information from the image to be processed along the horizontal axis of the pixel coordinate system to obtain a horizontal high-frequency image, extracts high-frequency information from the image to be processed along the vertical axis of the pixel coordinate system to obtain a vertical high-frequency image, and extracts high-frequency information from the image to be processed along directions other than the horizontal and vertical axes to obtain a diagonal high-frequency image. In this case, the first image includes a horizontal high-frequency image, a vertical high-frequency image, and a diagonal high-frequency image.

[0093] It should be understood that the contour segmentation device extracts a second image containing low-frequency information from the image to be processed, which is equivalent to performing feature extraction processing on the image to be processed, that is, the second image includes the features of the image to be processed.

[0094] 103. Upsample the second image to obtain the third image.

[0095] In this embodiment, the contour segmentation device increases the image resolution by upsampling the image. Therefore, when the image processing obtains the third image by upsampling the second image, the resolution of the third image is greater than the resolution of the second image.

[0096] In one possible implementation, the contour segmentation device upsamples the second image by performing linear interpolation to obtain the third image. In another possible implementation, the contour segmentation device upsamples the second image by performing nearest-neighbor interpolation to obtain the third image. In yet another possible implementation, the contour segmentation device upsamples the second image by performing bilinear interpolation to obtain the third image. In yet another possible implementation, the contour segmentation device upsamples the second image by performing transposed convolution to obtain the third image.

[0097] Optionally, the resolution of the third image is the same as that of the first image. For example, if the resolution of the first image is 256*256, then the resolution of the third image is also 256*256.

[0098] Optionally, the image processing device fuses the first image and the second image to obtain a fused image. Then, it upsamples the fused image to obtain a third image.

[0099] 104. The first image and the third image are fused together to obtain the fourth image.

[0100] The contour segmentation device fuses the first and third images to obtain a fourth image with the same resolution as the image to be processed. Since the third image is obtained by upsampling the second image, and both the second and third images contain low-frequency information from the image to be processed, and since the first image contains high-frequency information from the image to be processed, the contour segmentation device can fuse the high-frequency and low-frequency information of the image to be processed by fusing the first and third images. Thus, the fourth image contains both the low-frequency and high-frequency information of the image to be processed.

[0101] In other words, the contour segmentation device extracts low-frequency information from the image to be processed to obtain a second image, thus achieving feature extraction and obtaining a second image containing the features of the image to be processed; that is, the second image is the feature image of the image to be processed. Then, by upsampling the second image, its resolution is increased to obtain a third image, thereby increasing the resolution of the feature image of the image to be processed. However, because the contour segmentation device discards high-frequency information in the image to be processed during the extraction of low-frequency information to obtain the second image, the third image lacks high-frequency information from the image to be processed. Therefore, the contour segmentation device can enrich the information in the third image by fusing the first and third images to obtain a fourth image.

[0102] 105. Based on the fourth image above, the contour of the target organ in the image to be processed is segmented to obtain a first segmented image.

[0103] As mentioned earlier, the fourth image is a feature image of the image to be processed, meaning that the fourth image contains the features of the image to be processed. Therefore, the contour segmentation device can determine the semantics of the pixels in the image to be processed based on the fourth image, and then segment the contour of the target organ from the image to be processed based on the semantics of the pixels in the image to be processed.

[0104] For example, when the target organ is the kidney, the contour segmentation device can determine the pixels in the image to be processed that have the semantic meaning of kidney based on the fourth image, and then take the pixel region composed of the pixels with the semantic meaning of kidney as the pixel region covered by the kidney. At this time, the contour of the pixel region covered by the kidney is the kidney contour.

[0105] Therefore, by executing step 301, the contour segmentation device can obtain a first segmented image by segmenting the target organ from the image to be processed, that is, the first segmented image includes the segmentation result of the target organ.

[0106] As mentioned above, the fourth image obtained based on the technical solution provided above can enrich the low-frequency and high-frequency information in the fourth image. Therefore, when the contour segmentation device obtains the fourth image, it can segment the target organ in the image to be processed according to the fourth image to obtain the first segmented image, which can improve the accuracy of the first segmented image.

[0107] As an optional implementation, the contour segmentation device performs the following steps during step 102:

[0108] 201. Perform n-level target processing on the above image to be processed to obtain the above second image.

[0109] In this embodiment of the application, the output image of any level of target processing in the n-level target processing includes a high-frequency image containing high-frequency information and a low-frequency image containing low-frequency information. That is, each level of target processing will extract a high-frequency image containing high-frequency information and a low-frequency image containing low-frequency information, where n is an integer greater than or equal to 1.

[0110] When n is greater than 1, the n-level target processing is sequentially connected in series. Specifically, the output image of the first-level target processing is the input image of the second-level target processing, ..., the output image of the (n-1)th-level target processing is the input image of the nth-level target processing. Optionally, the output image of the previous-level target processing is the input image of the next-level target processing, that is, the low-frequency and high-frequency images output by the first-level target processing are the input images of the second-level target processing, ..., the low-frequency and high-frequency images output by the (n-1)th-level target processing are the input images of the nth-level target processing.

[0111] In one possible implementation, the target processing is wavelet transform. The contour segmentation device performs n-level target processing on the image to be processed to obtain a second image. For example, n is 2. The contour segmentation device performs a first-level wavelet transform on the image to be processed, outputting a first-level high-frequency image and a first-level low-frequency image. Then, a second-level wavelet transform is performed on the first-level low-frequency image and the first-level high-frequency image to obtain the second image.

[0112] The contour segmentation device extracts low-frequency information from the image to be processed by performing target processing on the image to be processed, thus obtaining a low-frequency image. As mentioned above, extracting low-frequency information is equivalent to extracting features from the image to be processed. Therefore, the contour segmentation device can extract features from the image to be processed by performing target processing on the image to be processed. When n is greater than 1, by performing n levels of target processing on the image to be processed, the low-frequency information extracted by each level of target processing is different, that is, the features extracted by each level of target processing are different.

[0113] Contour segmentation devices extract features by performing progressive target processing on the image to be processed. This gradually extracts low-frequency information from the image while discarding non-low-frequency information, resulting in increasingly lower resolution low-frequency images. In other words, the resolution of the low-frequency image output from each level of the n-level target processing is different; specifically, the higher the level of target processing, the lower the resolution of the output low-frequency image. For example, the resolution of the low-frequency image output from the first level of target processing is higher than that from the second level.

[0114] In this embodiment, the contour segmentation device performs the following steps during step 103:

[0115] 202. Perform m-level upsampling on the second image to obtain the third image.

[0116] In this embodiment, the contour segmentation device improves the resolution of the second image by upsampling it. The contour segmentation device gradually improves the resolution of the second image by performing m-level upsampling, ultimately obtaining a third image, where m is a positive integer. Optionally, m and n are equal.

[0117] The m-level upsampling process is implemented through m upsampling layers, with each layer performing one level of upsampling. When m is an integer greater than 1, the m upsampling layers are concatenated, meaning the output image of the first upsampling layer serves as the input image of the second upsampling layer, and so on, with the output image of the (m-1)th upsampling layer serving as the input image of the m-th upsampling layer. The output image of the m-th upsampling layer is the intermediate feature sequence. Optionally, each upsampling layer includes at least one of the following processing methods: bilinear interpolation, nearest-neighbor interpolation, or deconvolution.

[0118] In this embodiment, the resolution of the output image of the i-th level upsampling process in the m-level upsampling process and the resolution of the output image of the j-th level target processing in the n-level target processing are both the target resolution, where i is a positive integer less than m, and j is a positive integer less than or equal to n. For example, m = n = 5, i = 2, j = 3. In this case, the resolution of the image output by the third-level target processing and the resolution of the image output by the second-level upsampling process are both the target resolution.

[0119] It should be understood that the resolution of the output image of the j-th level target processing is the target resolution, which means that the resolution of both the high-frequency image and the low-frequency image output by the j-th level target processing are the target resolution.

[0120] In this embodiment, the high-frequency image output by the n-level target processing has a resolution lower than the target resolution and is considered the target image; that is, the high-frequency image output by the target processing with a level higher than j is considered the target image. For example, if n = 5 and j = 3, the target processing with a level higher than j includes the fourth-level target processing and the fifth-level target processing. The resolution of the high-frequency image output by the fourth-level target processing and the fifth-level target processing is lower than the resolution of the output image of the third-level target processing. Therefore, both the high-frequency image output by the fourth-level target processing and the high-frequency image output by the fifth-level target processing are considered target images.

[0121] In this embodiment, the input image of the (i+1)th level upsampling process in the m-level upsampling process includes: the output image of the i-th level upsampling process, the low-frequency image output by the j-th level target processing, and the target image. For example, when n = m = 5, i = 2, and j = 3, the high-frequency image output by the fourth level target processing and the high-frequency image output by the fifth level target processing are both target images. The input image of the third level upsampling process includes: the output image of the second level upsampling process, the low-frequency image output by the third level target processing, the high-frequency image output by the fourth level target processing, and the high-frequency image output by the fifth level target processing.

[0122] In this embodiment, the contour segmentation device performs n-level target processing on the image to be processed, extracting high-frequency images and low-frequency images of the image to be processed level by level to obtain a second image. Then, the resolution of the second image is increased by performing m-level upsampling processing on the second image. Furthermore, during the m-level upsampling process on the second image, the low-frequency image with the target resolution and the image output from the previous upsampling process are used as inputs, enriching the low-frequency information at different resolutions, i.e., enriching the low-frequency information at different scales. Conversely, during the m-level upsampling process on the second image, the target image output from the target processing is used as the input image for the upsampling process, enriching the high-frequency information in the output image of the upsampling process. Therefore, the third image obtained through this embodiment contains both the low-frequency information and the high-frequency information of the image to be processed at different scales.

[0123] Based on this implementation method, this application also provides a feature extraction structure, through which steps 201, 202, and 104 can be implemented. Please refer to... Figure 2 , Figure 2 The diagram shown is a schematic representation of the feature extraction structure. Figure 2 As shown, the feature pyramid includes an encoding module and a decoding module. The encoding module includes four levels of wavelet transform downsampling, where wavelet transform downsampling is the target processing described above. The decoding module includes six levels of upsampling processing, where... Figure 2 The sixth-level upsampling process is not shown. Both wavelet transform downsampling and upsampling are implemented using a neural network structure.

[0124] exist Figure 2In the feature extraction structure shown, the input image for the first-level upsampling process is the low-frequency image output by the fourth-level wavelet transform downsampling. The input image for the second-level upsampling process includes the high-frequency image output by the fourth-level wavelet transform downsampling and the output image of the first-level upsampling process. The input image for the third-level upsampling process includes the high-frequency image output by the third-level wavelet transform downsampling, the output image of the second-level upsampling process, and the low-frequency image output by the fourth-level wavelet transform downsampling. The high-frequency image and the low-frequency image output by the third-level wavelet transform downsampling and the fourth-level wavelet transform downsampling are input to the third-level upsampling process through a feature pyramid structure. Optionally, the feature pyramid structure is used to perform image stitching.

[0125] The input image for the fourth-level upsampling process includes the high-frequency image output from the second-level wavelet transform downsampling, the output image from the third-level upsampling process, the low-frequency image output from the third-level wavelet transform downsampling, and the low-frequency image output from the fourth-level wavelet transform downsampling. The input image for the fifth-level upsampling process includes the high-frequency image output from the first-level wavelet transform downsampling, the output image from the fourth-level upsampling process, the low-frequency image output from the second-level wavelet transform downsampling, the low-frequency image output from the third-level wavelet transform downsampling, and the low-frequency image output from the fourth-level wavelet transform downsampling. The image output after the fifth-level upsampling process is the aforementioned third image.

[0126] The input image for the sixth-level upsampling process includes a third image, an image to be processed, and a first image. The first image comprises: low-frequency images output from the first-level wavelet transform downsampling, the second-level wavelet transform downsampling, the third-level wavelet transform downsampling, and the fourth-level wavelet transform downsampling. The image output after the sixth-level upsampling process is the aforementioned fourth image.

[0127] As an optional implementation, the high-frequency image output by the target processing includes a first high-frequency image and a second high-frequency image. The direction of the edges represented by the first high-frequency image differs from the direction of the edges represented by the second high-frequency image. That is, the contour segmentation device extracts high-frequency information along different directions by performing target processing, thereby obtaining the first high-frequency image and the second high-frequency image respectively. For example, the contour segmentation device extracts high-frequency information along the horizontal axis of the pixel coordinate system of the image to be processed by performing target processing, thereby obtaining the first high-frequency image. The contour segmentation device extracts high-frequency information along the vertical axis of the pixel coordinate system of the image to be processed by performing target processing, thereby obtaining the second high-frequency image.

[0128] In this embodiment, the contour segmentation device extracts high-frequency information from different directions by performing target processing, and obtains high-frequency images that characterize the edges in different directions, thus enabling better extraction of high-frequency information from the image to be processed.

[0129] As an optional implementation, after obtaining the first segmented image, the contour segmentation device further performs the following steps:

[0130] 301. Determine the concave curvature of the outline of the target organ in the first segmented image.

[0131] 302. If the concave curvature is less than the first threshold, the first segmented image is determined to be an incorrect segmentation result.

[0132] 303. When the concave curvature is greater than or equal to the first threshold, the first segmented image is determined to be the segmentation result of the image to be processed.

[0133] Because human target organs have specific shapes, that is, their outlines have specific shapes, the concave curvature of the target organ's outline should be within a reasonable range. Specifically, if the concave curvature of the target organ's outline segmented from the image to be processed is outside the reasonable range, then the outline segmented from the image to be processed is incorrect. Conversely, if the concave curvature of the target organ's outline segmented from the image to be processed is within the reasonable range, then the outline segmented from the image to be processed is correct.

[0134] In this embodiment, the contour segmentation device uses a first threshold as a basis to determine whether the concave curvature of the contour of the target organ in the first segmented image is within a reasonable range. Specifically, if the concave curvature of the contour of the target organ in the first segmented image is less than the first threshold, it indicates that the contour of the target organ is outside the reasonable range, and thus the contour of the target organ in the first segmented image is determined to be incorrect, i.e., the first segmented image is determined to be an incorrect segmentation result. If the concave curvature of the contour of the target organ in the first segmented image is greater than or equal to the first threshold, it indicates that the contour of the target organ is within the reasonable range, and thus the contour of the target organ in the first segmented image is determined to be correct, i.e., the first segmented image is determined to be the segmentation result of the image to be processed.

[0135] For example, if the target organ is the kidney and the first threshold is -0.06, then if the concave curvature of the kidney contour in the first segmented image is less than -0.06, the contour segmentation device determines that the first segmented image is an incorrect segmentation result; if the concave curvature of the kidney contour in the first segmented image is greater than or equal to -0.06, the first segmented image is determined to be the segmentation result of the image to be processed.

[0136] In this embodiment, the contour segmentation device determines whether the contour of the target organ in the first segmented image is correct based on a first threshold and the concave curvature, thereby improving the accuracy of the segmentation result of the image to be processed.

[0137] As an optional implementation, after obtaining the first segmented image, the contour segmentation device further performs the following steps:

[0138] 401. Determine the convex curvature of the outline of the target organ in the first segmented image.

[0139] 402. If the curvature of the convex surface is greater than the second threshold, the first segmented image is determined to be an incorrect segmentation result.

[0140] 403. When the curvature of the convex surface is less than or equal to the second threshold, the first segmented image is determined to be the segmentation result of the image to be processed.

[0141] Because human target organs have specific shapes, that is, their outlines have specific shapes, the convex curvature of the target organ's outline should be within a reasonable range. Specifically, if the convex curvature of the target organ's outline segmented from the image to be processed is outside the reasonable range, then the outline segmented from the image to be processed is incorrect. Conversely, if the convex curvature of the target organ's outline segmented from the image to be processed is within the reasonable range, then the outline segmented from the image to be processed is correct.

[0142] In this embodiment, the contour segmentation device uses a second threshold as a basis to determine whether the convex curvature of the target organ's contour in the first segmented image is within a reasonable range. Specifically, if the convex curvature of the target organ's contour in the first segmented image is less than the second threshold, it indicates that the target organ's contour is outside the reasonable range, thus determining that the target organ's contour in the first segmented image is incorrect, i.e., the first segmented image is determined to be an incorrect segmentation result. If the convex curvature of the target organ's contour in the first segmented image is greater than or equal to the second threshold, it indicates that the target organ's contour is within the reasonable range, thus determining that the target organ's contour in the first segmented image is correct, i.e., the first segmented image is determined to be the segmentation result of the image to be processed.

[0143] For example, if the target organ is the kidney and the second threshold is 0.04, then if the convex curvature of the kidney contour in the first segmented image is greater than 0.04, the contour segmentation device determines that the first segmented image is an incorrect segmentation result; if the convex curvature of the kidney contour in the first segmented image is less than or equal to 0.04, the first segmented image is determined to be the segmentation result of the image to be processed.

[0144] In this embodiment, the contour segmentation device determines whether the contour of the target organ in the first segmented image is correct based on the second threshold and the convex curvature, thereby improving the accuracy of the segmentation result of the image to be processed.

[0145] As an optional implementation, after determining that the first segmented image is the segmentation result of the image to be processed, the contour segmentation device further performs the following steps:

[0146] 501. Determine the first pixel in the first segmented image that has the semantic meaning of the target organ and is located outside the outline of the target organ.

[0147] In this embodiment, the semantic meaning of a pixel as a target organ indicates that the contour segmentation device determines that the pixel belongs to the target organ by performing segmentation processing on the image to be processed. If the semantic meaning of a pixel is a target organ and it is located outside the contour of the target organ, it means that the pixel is a scatter point of the target organ segmented by the segmentation processing of the image to be processed, and the scatter point is the aforementioned first pixel.

[0148] 502. Remove the first pixel from the first segmented image to obtain the second segmented image.

[0149] The contour segmentation device can improve the accuracy of the target organ segmentation result by removing the first pixel from the first segmented image. That is, by performing step 502 to obtain the second segmented image, the accuracy of the segmentation result of the image to be processed can be improved.

[0150] In one possible implementation, the contour segmentation device filters the first segmented image to remove the first pixel, thus obtaining the second segmented image.

[0151] In another possible implementation, the contour segmentation device dilates the contour of the target organ so that the contour includes a first pixel, obtaining an enlarged contour of the target organ. Then, the enlarged contour is reduced to obtain a second segmented image, where the dilation ratio is the same as the reduction ratio. For example, if the ratio of the area enclosed by the enlarged contour to the area enclosed by the contour before dilation is 1.2, then the ratio of the area enclosed by the enlarged contour to the area enclosed by the contour of the target organ in the second segmented image is also 1.2, where both the dilation and reduction ratios are 1.2.

[0152] As an optional implementation, after determining that the first segmented image is the segmentation result of the image to be processed, the contour segmentation device further performs the following steps:

[0153] 601. Determine a second pixel from the first segmented image that is semantically different from the target organ and is located within the outline of the target organ.

[0154] In this embodiment, a pixel whose semantics are not those of a target organ indicates that the contour segmentation device, through segmentation processing of the image to be processed, determines that the pixel does not belong to the target organ. If a pixel's semantics are not those of a target organ and it is located within the contour of the target organ, it means that the pixel is a void pixel within the contour of the target organ, i.e., the semantics of the pixel are incorrect; in other words, the semantics of the pixel should be those of the target organ. In this application, void pixels within the contour of the target organ are referred to as second pixels.

[0155] 602. Update the semantics of the second pixel in the first segmented image to the target organ to obtain the third segmented image.

[0156] The contour segmentation device improves the accuracy of target organ segmentation by updating the semantics of the second pixel to the target organ and removing empty pixels within the target organ's contour. Specifically, obtaining the third segmented image through step 602 improves the accuracy of the segmentation result of the image to be processed. In one possible implementation, the contour segmentation device uses a flooding algorithm to process the first segmented image, removing the second pixel from the first segmented image to obtain the third segmented image.

[0157] Based on the technical solutions provided above, this application also provides a possible application scenario. Please refer to... Figure 3 , Figure 3 The diagram shown is a flowchart of a kidney contour segmentation method. Specifically, Figure 3 The diagram shows the process of segmenting the kidney contour from an ultrasound image based on the technical solution provided above.

[0158] like Figure 3 As shown, the image to be processed is an ultrasound image including the kidney, with a resolution of 580*950. The contour segmentation device first preprocesses the image to be processed, obtaining an image with a resolution of 576*768. The preprocessing includes cropping and normalization. Cropping removes non-ultrasound scan areas from the image, and normalizing the pixel values ​​to 0-1 reduces the amount of data processed in subsequent steps. The 576*768 image is then input into the segmentation network to obtain the first segmented image of the image to be processed. The first segmented image has a resolution of 576*768. The segmentation network includes a feature extraction module, the structure of which is shown below. Figure 2 As shown, the feature extraction module can extract the feature image of the image to be processed (i.e., the third image mentioned above), and then the segmentation result of the image to be processed (i.e., the first segmentation image mentioned above) can be obtained based on the third image. Optionally, before using the segmentation network to process the image to be processed, the segmentation network can be trained. Specifically, the training data includes the image to be segmented, and the annotation data of the image to be segmented includes the position of the kidney contour.

[0159] The first segmentation result is post-processed to obtain the final segmentation result, where the resolution of the final segmentation result is 580*950. For example... Figure 3 As shown, the post-processing includes removing scattered points, retaining the largest segmented region as the area covered by the kidney to obtain the segmented image to be confirmed, and using a curvature algorithm to determine whether the segmented image to be confirmed is correct. Figure 4 The flowchart shown is a process for determining whether the segmented image to be confirmed is correct using a curvature algorithm. Figure 4 As shown, the contour segmentation device first acquires the image to be segmented, and then calculates the average concave curvature of the kidney contour in the image. It then determines whether the average concave curvature is less than a first threshold. If it is less than the first threshold, the kidney contour in the image is determined to be incorrect, and the image is deleted. If it is greater than or equal to the first threshold, the kidney contour is determined to be correct, and the image is not processed; instead, it is used as the segmentation result of the image to be processed.

[0160] Optionally, the first threshold can be obtained statistically. Figure 5 The figure shows the results of statistical analysis of the average concave curvature of the kidney contour in 37,962 ultrasound images including the kidney. Figure 5 As shown, the average concave curvature of the kidney contour in 29,081 ultrasound images is between -0.01 and 0; the average concave curvature of the kidney contour in 8,204 ultrasound images is between -0.02 and -0.01; the average concave curvature of the kidney contour in 551 ultrasound images is between -0.03 and -0.02; the average concave curvature of the kidney contour in 90 ultrasound images is between -0.04 and -0.03; the average concave curvature of the kidney contour in 13 ultrasound images is between -0.05 and -0.04; the average concave curvature of the kidney contour in 7 ultrasound images is between -0.06 and -0.05; and the average concave curvature of the kidney contour in 16 ultrasound images is less than -0.06. According to the statistical results, the average concave curvature of the kidney contour is mostly between -0.06 and 0. Therefore, -0.06 to 0 can be taken as a reasonable range of concave curvature of the kidney contour. Optionally, -0.06 can be taken as the first threshold.

[0161] After obtaining the final segmentation result of the image to be processed, the final segmentation result, the image to be processed, and the ground truth (GT) of the image to be processed can be displayed to observe the accuracy of the final segmentation result. Figure 6 The image shown is a display of one possible final segmentation result, as follows: Figure 6As shown, the leftmost image is the image to be processed, the third image from the left is the final segmentation result, the rightmost image is the ground truth (GT) of the image to be processed, and the second image from the left is the effect of overlaying the final segmentation result, the image to be processed, and the GT of the image to be processed. Figure 7 The image shows another possible display of the final segmentation result, as follows: Figure 6 As shown, the leftmost image is the image to be processed, the third image from the left is the final segmentation result, the rightmost image is the ground truth (GT) of the image to be processed, and the second image from the left is the effect of overlaying the final segmentation result, the image to be processed, and the GT of the image to be processed.

[0162] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0163] If the technical solution of this application involves personal information, the product using this technical solution has clearly informed the user of the personal information processing rules and obtained the user's voluntary consent before processing the personal information. If the technical solution of this application involves sensitive personal information, the product using this technical solution has obtained the user's separate consent before processing the sensitive personal information, and also meets the requirement of "express consent". For example, at personal information collection devices such as cameras, clear and prominent signs are set up to inform users that they have entered the scope of personal information collection and that personal information will be collected. If an individual voluntarily enters the collection scope, it is deemed that they have agreed to the collection of their personal information; or on the personal information processing device, while using clear signs / information to inform users of the personal information processing rules, authorization is obtained from the individual through pop-up information or by asking the individual to upload their personal information; wherein, personal information processing may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the types of personal information processed.

[0164] The methods of the embodiments of this application have been described in detail above, and the apparatus of the embodiments of this application is provided below.

[0165] Please see Figure 8 , Figure 8 This is a schematic diagram of a contour segmentation device provided in an embodiment of this application. The contour segmentation device 1 includes: an acquisition unit 11, an extraction unit 12, an upsampling processing unit 13, a fusion unit 14, and a segmentation unit 15. Specifically:

[0166] Acquisition unit 11 is used to acquire an image to be processed, the image to be processed including a target organ;

[0167] Extraction unit 12 is used to extract a first image containing high-frequency information and a second image containing low-frequency information from the image to be processed;

[0168] Upsampling processing unit 13 is used to upsample the second image to obtain a third image;

[0169] The fusion unit 14 is used to fuse the first image and the third image to obtain a fourth image;

[0170] The segmentation unit 15 is used to segment the contour of the target organ in the image to be processed based on the fourth image to obtain a first segmented image.

[0171] In any embodiment of this application, the extraction unit 12 is used for:

[0172] The image to be processed is subjected to n levels of target processing to obtain the second image; the output image of any level of target processing in the n levels of target processing includes a high-frequency image containing high-frequency information and a low-frequency image containing low-frequency information;

[0173] The upsampling process of the second image to obtain the third image includes:

[0174] The second image is subjected to m-level upsampling processing to obtain the third image; the resolution of the output image of the i-th level upsampling processing in the m-level upsampling processing and the resolution of the output image of the j-th level target processing in the n-level target processing are both the target resolution, where i is a positive integer less than m and j is a positive integer less than or equal to n;

[0175] In the high-frequency image output by the n-level target processing, the image with a resolution lower than the target resolution is the target image.

[0176] The input image of the (i+1)th level upsampling process in the m-level upsampling process includes: the output image of the i-th level upsampling process, the low-frequency image output by the j-th level target processing, and the target image.

[0177] In any embodiment of this application, the high-frequency image includes a first high-frequency image and a second high-frequency image, wherein the direction of the edge represented by the first high-frequency image is different from the direction of the edge represented by the second high-frequency image.

[0178] In conjunction with any embodiment of this application, the segmentation unit 15 is further configured to:

[0179] Determine the concave curvature of the contour of the target organ in the first segmented image;

[0180] If the concave curvature is less than a first threshold, the first segmented image is determined to be an incorrect segmentation result;

[0181] If the concave curvature is greater than or equal to the first threshold, the first segmented image is determined to be the segmentation result of the image to be processed.

[0182] In conjunction with any embodiment of this application, the segmentation unit 15 is further configured to:

[0183] Determine the convex curvature of the contour of the target organ in the first segmented image;

[0184] If the curvature of the convex surface is greater than the second threshold, the first segmented image is determined to be an incorrect segmentation result;

[0185] If the convex curvature is less than or equal to the second threshold, the first segmented image is determined to be the segmentation result of the image to be processed.

[0186] In conjunction with any embodiment of this application, the segmentation unit 15 is further configured to:

[0187] The first pixel in the first segmented image is identified as the target organ and is located outside the outline of the target organ.

[0188] Remove the first pixel from the first segmented image to obtain the second segmented image.

[0189] In conjunction with any embodiment of this application, the segmentation unit 15 is further configured to:

[0190] Determine a second pixel from the first segmented image that is semantically not the target organ and is located within the outline of the target organ;

[0191] The semantics of the second pixel in the first segmented image are updated to the target organ to obtain the third segmented image.

[0192] In this embodiment, the third image is obtained by upsampling the second image, and both the second and third images contain low-frequency information of the image to be processed. Since the first image contains high-frequency information of the image to be processed, the contour segmentation device can fuse the high-frequency and low-frequency information of the image to be processed by fusing the first and third images. Thus, the fourth image contains both the low-frequency and high-frequency information of the image to be processed.

[0193] In other words, the contour segmentation device extracts low-frequency information from the image to be processed to obtain a second image, thus achieving feature extraction and obtaining a second image containing the features of the image to be processed; that is, the second image is the feature image of the image to be processed. Then, by upsampling the second image, its resolution is increased to obtain a third image, thereby increasing the resolution of the feature image of the image to be processed. However, because the contour segmentation device discards high-frequency information in the image to be processed during the extraction of low-frequency information to obtain the second image, the third image lacks high-frequency information from the image to be processed. Therefore, the contour segmentation device can enrich the information in the third image by fusing the first and third images to obtain a fourth image.

[0194] Therefore, the contour segmentation device segments the contour of the target organ in the image to be processed based on the fourth image to obtain the first segmented image, which can improve the accuracy of the first segmented image.

[0195] In some embodiments, the functions or modules of the apparatus provided in this application can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0196] Figure 9 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. The electronic device 2 includes a processor 21 and a memory 22. Optionally, the electronic device 2 also includes an input device 23 and an output device 24. The processor 21, memory 22, input device 23, and output device 24 are coupled together via connectors, which include various interfaces, transmission lines, or buses, etc., and are not limited in this embodiment. It should be understood that in the various embodiments of this application, coupling refers to mutual connection in a specific way, including direct connection or indirect connection through other devices, such as through various interfaces, transmission lines, buses, etc.

[0197] Processor 21 can be one or more graphics processing units (GPUs). If processor 21 is a GPU, the GPU can be a single-core GPU or a multi-core GPU. Optionally, processor 21 can be a processor group composed of multiple GPUs, with the multiple processors coupled to each other via one or more buses. Optionally, the processor can also be other types of processors, etc., which are not limited in this embodiment.

[0198] The memory 22 can be used to store computer program instructions, as well as various types of computer program code, including program code for executing the scheme of this application. Optionally, the memory includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or compact disc read-only memory (CD-ROM), which is used for related instructions and data.

[0199] Input device 23 is used to input data and / or signals, and output device 24 is used to output data and / or signals. Input device 23 and output device 24 can be independent devices or an integrated device.

[0200] It is understood that in this embodiment of the application, the memory 22 can be used not only to store related instructions, but also to store related data. This embodiment of the application does not limit the specific data stored in the memory.

[0201] Understandable Figure 9 This is merely a simplified design of an electronic device. In practical applications, the electronic device may also include other necessary components, including, but not limited to, any number of input / output devices, processors, memories, etc., and all electronic devices that can implement the embodiments of this application are within the protection scope of this application.

[0202] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0203] Those skilled in the art will readily understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. Those skilled in the art will also readily understand that the various embodiments of this application have different focuses, and for the sake of convenience and brevity, the same or similar parts may not be repeated in different embodiments. Therefore, parts not described or not described in detail in one embodiment can be referred to the descriptions in other embodiments.

[0204] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0205] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0206] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0207] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted through the computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., digital versatile discs (DVDs)), or semiconductor media (e.g., solid-state disks (SSDs)).

[0208] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for segmenting the contour of a target organ, characterized in that, The method includes: Acquire an image to be processed, the image including the target organ; Extracting a first image containing high-frequency information and a second image containing low-frequency information from the image to be processed; the extraction of the first image containing high-frequency information and the second image containing low-frequency information from the image to be processed includes: performing n-level target processing on the image to be processed to obtain the second image; the output image of any level of target processing in the n-level target processing includes a high-frequency image containing high-frequency information and a low-frequency image containing low-frequency information; The second image is upsampled to obtain a third image; the upsampling of the second image to obtain the third image includes: performing m-level upsampling on the second image to obtain the third image; the resolution of the output image of the i-th level upsampling in the m-level upsampling and the resolution of the output image of the j-th level target processing in the n-level target processing are both target resolutions, where i is a positive integer less than m, and j is a positive integer less than or equal to n; the high-frequency image output by the n-level target processing has a resolution lower than the target resolution and is the target image; the input image of the (i+1)-th level upsampling in the m-level upsampling includes: the output image of the i-th level upsampling, the low-frequency image output by the j-th level target processing, and the target image; The first image and the third image are fused to obtain a fourth image; Based on the fourth image, the contour of the target organ in the image to be processed is segmented to obtain a first segmented image; The first pixel in the first segmented image is identified as the target organ and is located outside the outline of the target organ. The contour of the target organ in the first segmented image is dilated so that the contour of the target organ includes a first pixel, thereby obtaining the dilated contour of the target organ. The ratio of the area enclosed by the dilated contour to the area enclosed by the contour of the target organ in the first segmented image is a first ratio. The inflated contour is reduced to obtain a second segmented image, wherein the ratio of the area enclosed by the inflated contour to the area enclosed by the contour of the target organ in the second segmented image is a second ratio, and the first ratio is equal to the second ratio.

2. The method according to claim 1, characterized in that, The high-frequency image includes a first high-frequency image and a second high-frequency image, wherein the direction of the edge represented by the first high-frequency image is different from the direction of the edge represented by the second high-frequency image.

3. The method according to claim 2, characterized in that, After obtaining the first segmented image, the method further includes: Determine the concave curvature of the contour of the target organ in the first segmented image; If the concave curvature is less than a first threshold, the first segmented image is determined to be an incorrect segmentation result; If the concave curvature is greater than or equal to the first threshold, the first segmented image is determined to be the segmentation result of the image to be processed.

4. The method according to claim 1, characterized in that, After obtaining the first segmented image, the method further includes: Determine the convex curvature of the contour of the target organ in the first segmented image; If the curvature of the convex surface is greater than the second threshold, the first segmented image is determined to be an incorrect segmentation result; If the convex curvature is less than or equal to the second threshold, the first segmented image is determined to be the segmentation result of the image to be processed.

5. The method according to claim 3 or 4, characterized in that, After determining that the first segmented image is the segmentation result of the image to be processed, the method further includes: The first pixel in the first segmented image is identified as the target organ and is located outside the outline of the target organ. Remove the first pixel from the first segmented image to obtain the second segmented image.

6. The method according to claim 3 or 4, characterized in that, After determining that the first segmented image is the segmentation result of the image to be processed, the method further includes: Determine a second pixel from the first segmented image that is semantically not the target organ and is located within the outline of the target organ; The semantics of the second pixel in the first segmented image are updated to the target organ to obtain the third segmented image.

7. A target organ contour segmentation device, characterized in that, The device includes: An acquisition unit is used to acquire an image to be processed, the image to be processed including a target organ; An extraction unit is configured to extract a first image containing high-frequency information and a second image containing low-frequency information from the image to be processed; the extraction of the first image containing high-frequency information and the second image containing low-frequency information from the image to be processed includes: performing n-level target processing on the image to be processed to obtain the second image; the output image of any level of target processing in the n-level target processing includes a high-frequency image containing high-frequency information and a low-frequency image containing low-frequency information; An upsampling processing unit is used to upsample the second image to obtain a third image. The upsampling of the second image to obtain the third image includes: performing m-level upsampling processing on the second image to obtain the third image; the resolution of the output image of the i-th level upsampling process in the m-level upsampling process and the resolution of the output image of the j-th level target processing in the n-level target processing are both target resolutions, where i is a positive integer less than m, and j is a positive integer less than or equal to n; the high-frequency image output by the n-level target processing has a resolution lower than the target resolution, and the input image of the (i+1)-th level upsampling process in the m-level upsampling process includes: the output image of the i-th level upsampling process, the low-frequency image output by the j-th level target processing, and the target image; A fusion unit is used to fuse the first image and the third image to obtain a fourth image; The segmentation unit is used to segment the contour of the target organ in the image to be processed based on the fourth image to obtain a first segmented image; The first pixel in the first segmented image is identified as the target organ and is located outside the outline of the target organ. The contour of the target organ in the first segmented image is dilated so that the contour of the target organ includes a first pixel, thereby obtaining the dilated contour of the target organ. The ratio of the area enclosed by the dilated contour to the area enclosed by the contour of the target organ in the first segmented image is a first ratio. The inflated contour is reduced to obtain a second segmented image, wherein the ratio of the area enclosed by the inflated contour to the area enclosed by the contour of the target organ in the second segmented image is a second ratio, and the first ratio is equal to the second ratio.

8. An electronic device, characterized in that, include: A processor and a memory, the memory being used to store computer program code, the computer program code including computer instructions, wherein, when the processor executes the computer instructions, the electronic device performs the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program including program instructions that, when executed by a processor, cause the processor to perform the method according to any one of claims 1 to 6.

Images