Method, device, storage medium and electronic equipment for determining a lesion area
By acquiring the registration deformation field and lesion location information of the target image, and using the initial template image to divide the lung lesion area, the problem of lesion area recognition relying on doctors' experience in the existing technology is solved, and automated and efficient lesion area recognition is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHENYANG NEUSOFT INTELLIGENT MEDICAL TECH RES INST
- Filing Date
- 2022-11-25
- Publication Date
- 2026-05-29
AI Technical Summary
In existing technologies, lesion area identification relies on the professional knowledge and experience of doctors, resulting in low identification efficiency and difficulty in quickly and accurately determining the lesion area, especially in the identification of lung lesions, which is time-consuming and involves a large amount of annotation work.
By acquiring the registration deformation field and lesion location information of the target image, the functional regions of the target image are divided using the initial template image, and the lesion region is automatically determined by combining the lesion location information, thus avoiding manual intervention.
It enables automatic identification of lesion areas without human intervention, improving the efficiency and accuracy of lesion area identification, especially shortening the identification time and reducing the workload of annotation when identifying lung lesions.
Smart Images
Figure CN115861215B_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of image processing technology, and more specifically, to a method, apparatus, storage medium, and electronic device for determining lesion regions. Background Technology
[0002] With the rapid development of medical technology, an increasing number of diagnostic tests are being used in medical settings. The images generated during these tests significantly aid in the diagnosis and treatment process, allowing doctors to determine the location, size, and benign or malignant nature of lesions. Typically, lesion regions in medical images need to be predicted based on the doctor's professional knowledge and experience. The accuracy of this prediction depends heavily on the doctor's expertise, and given the current disparity in the ratio of cases to specialists, the efficiency of lesion region identification is relatively low. Therefore, improving the efficiency of lesion region identification has become an urgent problem to be solved. Summary of the Invention
[0003] To address the aforementioned problems, this disclosure provides a method, apparatus, storage medium, and electronic device for determining lesion regions.
[0004] In a first aspect, this disclosure provides a method for determining a lesion region, including:
[0005] Obtain the registration deformation field of the target image of the lesion region to be determined and the lesion location information of the target lesion in the target image;
[0006] Based on the registration deformation field, a registration template image corresponding to the initial template image is determined. The initial template image is an image that does not contain the target lesion and other lesions, and has been divided into template functional regions. The other lesions are lesions other than the target lesion.
[0007] Based on the registration template image, multiple target functional regions of the target image are determined; the target functional regions of the target image are located at the same positions as the template functional regions of the initial template image.
[0008] Based on the lesion location information and multiple target functional regions, the lesion region of the target lesion in the target image is determined.
[0009] Optionally, the target image includes a first image region and a second image region; determining multiple target functional regions of the target image based on the registration template image includes:
[0010] Based on the position of the template functional region in the registration template image, determine multiple undetermined functional regions of the target image;
[0011] When the target lesion is determined to be located in the first image region based on the lesion location information, multiple undetermined functional regions of the second image region are taken as multiple target functional regions of the second image region, and multiple target functional regions of the first image region are determined based on the location of the multiple target functional regions of the second image region.
[0012] The target functional regions contained in the first image region are symmetrical in position to the target functional regions contained in the second image region.
[0013] Optionally, determining the lesion region of the target lesion in the target image based on the lesion location information and multiple target functional regions includes:
[0014] When the organ corresponding to the target image is a first preset organ, the target functional region where the centroid of the target lesion is located is determined from multiple target functional regions according to the lesion location information, and the target functional region where the centroid of the target lesion is located is taken as the lesion region of the target image. The lesion type of the first preset organ does not include hemorrhage.
[0015] Optionally, determining the lesion region of the target lesion in the target image based on the lesion location information and multiple target functional regions includes:
[0016] When the organ corresponding to the target image is a second preset organ, the overlap ratio between the target lesion and each of the target functional regions of the first image region is determined according to the lesion location information, and the lesion type of the second preset organ includes hemorrhage;
[0017] The target functional region corresponding to the largest overlap ratio among the multiple overlap ratios is taken as the region to be determined.
[0018] If the maximum overlap ratio is greater than or equal to a preset ratio threshold, the region to be determined is taken as the lesion region of the target image.
[0019] Optionally, the method further includes:
[0020] If the maximum overlap ratio is less than the preset ratio threshold, determine whether the centroid of the target lesion is located in the undetermined region;
[0021] If the centroid of the target lesion is located in the undetermined region, the undetermined region is taken as the lesion region of the target image.
[0022] Optionally, the method further includes:
[0023] If the centroid of the target lesion is not located in the undetermined region, the base region of the target image is taken as the lesion region of the target image, and the base region is the region in the first image region other than the plurality of target functional regions.
[0024] Optionally, acquiring the registration deformation field of the target image of the lesion region to be determined includes:
[0025] The target image is input into a pre-generated registration deformation field generation model to obtain the registration deformation field output by the registration deformation field generation model.
[0026] Secondly, this disclosure provides an apparatus for determining a lesion region, comprising:
[0027] The acquisition module is used to acquire the registration deformation field of the target image of the lesion area to be determined and the lesion location information of the target lesion in the target image;
[0028] The first determining module is used to determine the registration template image corresponding to the initial template image based on the registration deformation field. The initial template image is an image that does not contain the target lesion and other lesions and has been divided into template functional regions. The other lesions are lesions other than the target lesion.
[0029] The second determining module is used to determine multiple target functional regions of the target image based on the registration template image; the target functional regions of the target image are located at the same positions as the template functional regions of the initial template image;
[0030] The third determining module is used to determine the lesion region of the target lesion in the target image based on the lesion location information and multiple target functional regions.
[0031] Optionally, the target image includes a first image region and a second image region; the second determining module is further configured to:
[0032] Based on the position of the template functional region in the registration template image, determine multiple undetermined functional regions of the target image;
[0033] When the target lesion is determined to be located in the first image region based on the lesion location information, multiple undetermined functional regions of the second image region are taken as multiple target functional regions of the second image region, and multiple target functional regions of the first image region are determined based on the location of the multiple target functional regions of the second image region.
[0034] The target functional regions contained in the first image region are symmetrical in position to the target functional regions contained in the second image region.
[0035] Optionally, the third determining module is further configured to:
[0036] When the organ corresponding to the target image is a first preset organ, the target functional region where the centroid of the target lesion is located is determined from multiple target functional regions according to the lesion location information, and the target functional region where the centroid of the target lesion is located is taken as the lesion region of the target image. The lesion type of the first preset organ does not include hemorrhage.
[0037] Optionally, the third determining module is further configured to:
[0038] When the organ corresponding to the target image is a second preset organ, the overlap ratio between the target lesion and each of the target functional regions of the first image region is determined according to the lesion location information, and the lesion type of the second preset organ includes hemorrhage;
[0039] The target functional region corresponding to the largest overlap ratio among the multiple overlap ratios is taken as the region to be determined.
[0040] If the maximum overlap ratio is greater than or equal to a preset ratio threshold, the region to be determined is taken as the lesion region of the target image.
[0041] Optionally, the device further includes:
[0042] The fourth determining module is used to determine whether the centroid of the target lesion is located in the undetermined region when the maximum overlap ratio is less than the preset ratio threshold.
[0043] The fifth determining module is used to determine the undetermined region as the lesion region of the target image when the centroid of the target lesion is located in the undetermined region.
[0044] Optionally, the device further includes:
[0045] The sixth determining module is used to determine the base region of the target image as the lesion region of the target image when the centroid of the target lesion is not located in the undetermined region. The base region is the region in the first image region other than the plurality of target functional regions.
[0046] Optionally, the acquisition module is further configured to:
[0047] The target image is input into a pre-generated registration deformation field generation model to obtain the registration deformation field output by the registration deformation field generation model.
[0048] Thirdly, this disclosure provides a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps of the method of the first aspect of this disclosure.
[0049] Fourthly, this disclosure provides an electronic device, comprising: a memory having a computer program stored thereon; and a processor for executing the computer program in the memory to implement the steps of the method of the first aspect of this disclosure.
[0050] The technical solutions provided by the embodiments of this disclosure may include the following beneficial effects:
[0051] The method involves acquiring the registration deformation field of a target image containing the lesion region to be determined and the lesion location information of the target lesion in the target image; determining a registration template image corresponding to an initial template image based on the registration deformation field, wherein the initial template image is an image that does not contain the target lesion or other lesions and has been divided into template functional regions, and the other lesions are lesions other than the target lesion; determining multiple target functional regions of the target image based on the registration template image; the target functional regions of the target image are located at the same positions as the template functional regions of the initial template image; and determining the lesion region of the target lesion in the target image based on the lesion location information and the multiple target functional regions. In other words, this disclosure can first determine multiple target functional regions of the target image based on the initial template image, and then determine the lesion region of the target lesion in the target image based on the lesion location information of the target lesion in the target image and the multiple target functional regions. This allows for automatic identification of the lesion region in the target image without manual intervention, improving the efficiency of lesion region identification.
[0052] Other features and advantages of this disclosure will be described in detail in the following detailed description section. Attached Figure Description
[0053] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the following detailed description to explain the present disclosure, but do not constitute a limitation thereof. In the drawings:
[0054] Figure 1 This is a flowchart illustrating a method for determining a lesion region according to an exemplary embodiment of the present disclosure;
[0055] Figure 2 This is a flowchart illustrating a method for generating a registration deformation field generation model according to an exemplary embodiment of the present disclosure;
[0056] Figure 3 This is an initial template image shown according to an exemplary embodiment of the present disclosure;
[0057] Figure 4 This is a flowchart illustrating another method for determining a lesion region according to an exemplary embodiment of the present disclosure;
[0058] Figure 5 This is a schematic diagram of a functional area according to an exemplary embodiment of the present disclosure;
[0059] Figure 6 This is a flowchart illustrating another method for determining a lesion region according to an exemplary embodiment of the present disclosure;
[0060] Figure 7 This is a flowchart illustrating another method for determining a lesion region according to an exemplary embodiment of the present disclosure;
[0061] Figure 8 This is a flowchart illustrating another method for determining a lesion region according to an exemplary embodiment of the present disclosure;
[0062] Figure 9 This is a block diagram illustrating an apparatus for determining a lesion region according to an exemplary embodiment of the present disclosure;
[0063] Figure 10 This is a block diagram illustrating another apparatus for determining a lesion region according to an exemplary embodiment of the present disclosure;
[0064] Figure 11 This is a block diagram illustrating another apparatus for determining a lesion region according to an exemplary embodiment of the present disclosure;
[0065] Figure 12 This is a block diagram illustrating an electronic device according to an exemplary embodiment. Detailed Implementation
[0066] The specific embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit this disclosure.
[0067] In the following description, the words "first" and "second" are used only to distinguish the purpose of the description and should not be interpreted as indicating or implying relative importance or order.
[0068] First, let's explain the application scenarios of this disclosure. Manually predicting lesion areas requires a high level of experience and expertise from doctors. While existing research on image recognition and segmentation algorithms has shown promise in assisting diagnosis, accurately segmenting functional areas with the naked eye based on experience is extremely difficult. For example, the lungs are a major application scenario. The lungs are the largest respiratory organ in the human body, the primary site of gas exchange, and an indispensable vital organ. However, due to factors such as air pollution, the incidence and mortality rates of lung lesions (such as tuberculosis, pulmonary nodules, lung cancer, and pneumonia) are increasing year by year, becoming a major threat to human health. Therefore, the study of lung lesions has become a crucial part of medical image analysis. Structurally, the lungs can be divided into left and right lungs, each further divided into five lobes: three lobes in the right lung and two lobes in the left lung. Anatomical divisions exist between the lobes. Specifically, a horizontal fissure exists between the right upper lobe and the right middle lobe; an oblique fissure separates the right middle lobe and the right lower lobe; and a left oblique fissure separates the left upper lobe and the left lower lobe. Furthermore, each lobe is separated by an independent and intact visceral pleura, resulting in clear anatomical divisions. Clinically, segmentation algorithms can be used to identify each lung lobe to determine the location of lesions and further assess their severity, thus aiding in diagnosis, decision-making, and prognosis. Therefore, accurate identification of the functional regions where lung lesions are located is crucial for understanding lung-related diseases.
[0069] In related technologies, traditional methods or deep learning methods can be used to identify interlobar clefts and determine different regions of the lung lobe. Traditional methods generally use conventional image segmentation algorithms to identify lung regions, blood vessels, and trachea separately, achieving the effect of distinguishing interlobar clefts to assist in lung lobe segmentation. To improve recognition accuracy, traditional image algorithms employ iterative calculations with penalty terms, resulting in very long recognition times and an inability to quickly obtain effective results. For some diseases, this may lead to missing the optimal treatment window. Deep learning, on the other hand, requires manual annotation of lesion regions in sample images during model training. The number of sample images is often extremely large, and the standard for sample annotation is difficult to unify, resulting in a huge workload for annotation and hindering rapid model updates and learning, thus leading to relatively low efficiency in lesion region identification.
[0070] To address the aforementioned problems, this disclosure provides a method, apparatus, storage medium, and electronic device for determining lesion regions. First, multiple target functional regions of a target image are determined based on an initial template image. Then, based on the lesion location information of the target lesion in the target image and the multiple target functional regions, the lesion region of the target lesion in the target image is determined. This allows for automatic identification of lesion regions in the target image without human intervention, thus improving the efficiency of lesion region identification.
[0071] The specific embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.
[0072] Figure 1 This is a flowchart illustrating a method for determining a lesion region according to an exemplary embodiment of the present disclosure, such as... Figure 1 As shown, the method may include:
[0073] S101. Obtain the registration deformation field of the target image of the lesion area to be determined and the lesion location information of the target lesion in the target image.
[0074] The target image can be a medical image captured by medical imaging equipment; for example, the target image can be an image of organs such as the brain, lungs, and heart.
[0075] In this step, after acquiring the target image, the target image can be input into a pre-generated registration deformation field generation model to obtain the registration deformation field output by the registration deformation field generation model. The target image can also be input into a pre-generated location determination model to obtain the lesion location information of the target lesion in the target image output by the location determination model.
[0076] The location determination model can be a mature model from existing technologies, and the registration deformation field generation model can be trained using unsupervised learning methods. For example, Figure 2 This is a flowchart illustrating a method for generating a registration deformation field generation model according to an exemplary embodiment of the present disclosure, such as... Figure 2 As shown, the method may include:
[0077] S21. Obtain motion images of samples and multiple stationary images of samples.
[0078] The moving image of the sample can be an image of a healthy target organ, while the stationary image of the sample can be an image of a target organ containing lesions.
[0079] S22. The target neural network model is trained using the motion image of the sample and multiple fixed images of the samples to obtain the registration deformation field generation model.
[0080] In one possible implementation, after acquiring the sample motion image and multiple sample fixed images, the current sample fixed image can be determined from the multiple sample fixed images. The model training steps are then executed iteratively based on the current sample fixed image until the trained target neural network satisfies a preset stopping iteration condition based on the current sample fixed image and the predicted deformation field. The trained target neural network model is then used as the registration deformation field generation model. The current sample fixed image can be any one of the multiple sample fixed images, and the predicted deformation field is the deformation field output by the target neural network model after the current sample fixed image and the sample motion image are input into it.
[0081] The model training steps may include:
[0082] S1. Input the motion image of the sample and the fixed image of the current sample into the target neural network model, and output the predicted deformation field.
[0083] S2. Based on the predicted deformation field, determine the sample registration image corresponding to the sample motion image.
[0084] S3. Based on the registered image of the sample and the fixed image of the current sample, determine the loss value corresponding to the target neural network model.
[0085] S4. If the target neural network model does not meet the preset stopping iteration condition based on the loss value, update the parameters of the target neural network model based on the loss value to obtain a new target neural network model, and determine a new current sample fixed image from multiple sample fixed images.
[0086] The target neural network model may include an affine transformation sub-model and a deformation field generation sub-model. For example, the sample motion image and the current sample fixed image can be input into the affine transformation sub-model to obtain a transformation matrix. The sample motion image is then distorted using this transformation matrix to obtain a sample transformed image. The sample transformed image and the current sample fixed image are then input into the deformation field generation sub-model to obtain a predicted deformation field. Based on the predicted deformation field, a sample registration image corresponding to the sample motion image is determined. Based on the sample registration image and the current sample fixed image, a loss value corresponding to the target neural network model is determined. If the loss value is greater than a preset loss value threshold, it is determined that the target neural network does not meet a preset stopping iteration condition. The parameters of the affine transformation sub-model and the deformation field generation sub-model are updated based on the loss value to obtain a new target neural network model. Then, one sample fixed image is randomly selected from multiple sample fixed images as the new current sample fixed image, and steps S1 to S4 are continued. If the loss value is less than or equal to the preset loss value threshold, the target neural network model is determined to meet the preset stopping iteration condition, and the target neural network model is used as the registration deformation field generation model.
[0087] It should be noted that the method of determining whether the target neural network model meets the preset stopping iteration condition by using the loss value during the above model training process is only an example. Other conditions in the prior art can also be used to determine whether the target neural network model meets the preset stopping iteration condition, and this disclosure does not limit this. Furthermore, the target neural network model can be a cascaded network. During the iterative execution of the above model training steps, the sample motion image can be registered and deformed recursively. After multiple executions of the model training steps, the sample motion image is continuously distorted, so that the final predicted deformation field can be decomposed into cascaded, small-displacement deformation fields.
[0088] S102. Based on the registration deformation field, determine the registration template image corresponding to the initial template image.
[0089] The initial template image is an image that does not contain the target lesion or other lesions, and whose functional regions have been defined. The other lesions are those other than the target lesion. For example, an image of a healthy organ can be obtained first, template information can be summarized based on expert experience, and the functional regions of the healthy organ image can be defined based on the template information to obtain the initial template image. Figure 3 This is an initial template image shown according to an exemplary embodiment of the present disclosure, such as... Figure 3 As shown, the initial template image is an image obtained by dividing a healthy brain image into functional regions. The initial template image includes 8 template functional regions.
[0090] In this step, after obtaining the registration deformation field of the target image, the initial template image can be resampled using the registration deformation field through existing techniques to obtain the registration template image corresponding to the initial template image.
[0091] S103. Based on the registration template image, determine multiple target functional regions of the target image.
[0092] The target functional region of the target image is located at the same position as the template functional region of the initial template image.
[0093] In this step, after obtaining the registration template image, the functional regions in the target image can be identified by using existing techniques based on the multiple template functional regions divided in the registration template image, thereby obtaining multiple target functional regions of the target image.
[0094] S104. Based on the lesion location information and multiple target functional areas, determine the lesion area of the target lesion in the target image.
[0095] In this step, after identifying multiple target functional regions of the target image, the target functional region where the target lesion is located can be determined based on the lesion location information, and this region is taken as the lesion region. If the target lesion is not located in multiple target functional regions, taking a brain image as an example, the target lesion can be identified as being in the basal ganglia of the target image.
[0096] Using the above method, multiple target functional regions of the target image are first determined based on the initial template image. Then, based on the lesion location information of the target lesion in the target image and the multiple target functional regions, the lesion region of the target lesion in the target image is determined. In this way, the lesion region in the target image can be automatically identified without human intervention, thus improving the efficiency of lesion region identification.
[0097] In one possible implementation, the target image includes a first image region and a second image region. Figure 4 This is a flowchart illustrating another method for determining a lesion region according to an exemplary embodiment of the present disclosure, such as... Figure 4 As shown, the implementation of step S103 may include:
[0098] S1031. Based on the position of the template functional area in the registration template image, determine multiple undetermined functional areas of the target image.
[0099] In this step, existing techniques can be used to identify functional regions in the target image based on multiple template functional regions divided in the registration template image, thereby obtaining multiple undetermined functional regions of the target image.
[0100] S1032. When it is determined that the target lesion is located in the first image region based on the lesion location information, multiple undetermined functional regions of the second image region are taken as multiple target functional regions of the second image region, and multiple target functional regions of the first image region are determined based on the location of the multiple target functional regions of the second image region.
[0101] The target functional area contained in the first image region is symmetrical to the target functional area contained in the second image region. Figure 5 This is a schematic diagram of a functional area according to an exemplary embodiment of the present disclosure, such as... Figure 5 As shown, the black line in the middle divides the target image into two image regions (the first image region and the second image region). The area enclosed by each black line is a target functional region, and the area enclosed by the gray line is the target lesion.
[0102] In this step, the image region where the target lesion is located can be determined based on the lesion location information. Figure 5Taking the target image shown as an example, if the right side of the target image is the first image region and the left side is the second image region, then the target lesion can be determined to be located in the first image region. The second image region of the target image does not contain the lesion. Without the influence of the lesion, the accuracy of determining multiple undetermined functional regions of the second image region based on the position of the template functional region in the registration template image is relatively high. These multiple undetermined functional regions of the second image region can be directly used as multiple target functional regions of the second image region. Since the target functional regions contained in the first image region and the target functional regions contained in the second image region are symmetrically positioned, the functional regions of the first image region can be divided by symmetrical flipping based on the multiple target functional regions of the second image region, thus obtaining multiple target functional regions of the first image region.
[0103] Using the above method, after dividing the functional regions of the target image by registering the template image, the lesion-containing image region is then divided according to the multiple target functional regions of the lesion-free image region, thus obtaining multiple target functional regions of the lesion-containing image region. This can improve the influence of lesions on the functional region division, improve the accuracy of functional region division, and thus improve the accuracy of the determined lesion region.
[0104] In one possible implementation, step S104 may include:
[0105] S1041. If the organ corresponding to the target image is a first preset organ, the target functional region where the centroid of the target lesion is located is determined from multiple target functional regions based on the lesion location information, and the target functional region where the centroid of the target lesion is located is taken as the lesion region of the target image.
[0106] Wherein, the lesion type of the first pre-defined organ does not include hemorrhage. For example, the first pre-defined organ may be the lung, and the target lesion may be a pulmonary nodule.
[0107] In this step, if the organ corresponding to the target image is a first preset organ, the centroid location information of the target lesion can be determined based on the lesion location information. Based on the centroid location information, the target functional area where the centroid of the target lesion is located can be determined, and the target functional area where the centroid of the target lesion is located can be used as the lesion area of the target image.
[0108] Figure 6 This is a flowchart illustrating another method for determining a lesion region according to an exemplary embodiment of the present disclosure, such as... Figure 6 As shown, step S104 can also be implemented by including:
[0109] S1042. If the organ corresponding to the target image is a second preset organ, determine the overlap ratio between the target lesion and each target functional area of the first image region based on the lesion location information.
[0110] The lesion type of the second pre-defined organ includes hemorrhage; for example, the second pre-defined organ may include the brain.
[0111] In this step, if the organ corresponding to the target image is a second preset organ, the overlap ratio between the target lesion and each target functional area in the first image region can be determined based on the lesion location information of the target lesion.
[0112] S1043. Take the target functional area corresponding to the largest overlap ratio among multiple overlap ratios as the undetermined area.
[0113] S1044. If the maximum overlap ratio is greater than or equal to a preset ratio threshold, the region to be determined shall be the lesion region of the target image.
[0114] The preset ratio threshold can be preset based on experience; for example, for brain images, the preset ratio threshold can be 10%.
[0115] In this step, if the maximum overlap ratio is greater than or equal to the preset ratio threshold, the region to be determined can be directly used as the lesion region of the target image.
[0116] Figure 7 This is a flowchart illustrating another method for determining a lesion region according to an exemplary embodiment of the present disclosure, such as... Figure 7 As shown, step S104 can also be implemented by including:
[0117] S1045. If the maximum overlap ratio is less than the preset ratio threshold, determine whether the centroid of the target lesion is located in the undetermined region.
[0118] In this step, if the maximum overlap ratio is less than the preset ratio threshold, it can be further determined whether the centroid of the target lesion is located in the undetermined region based on the centroid location information of the target lesion.
[0119] S1046. If the centroid of the target lesion is located in the undetermined region, the undetermined region shall be taken as the lesion region of the target image.
[0120] S1047. If the centroid of the target lesion is not located in the undetermined region, the basal region of the target image shall be taken as the lesion region of the target image.
[0121] The basal region is the region in the first image region other than the multiple target functional regions. For example, if the target image is a brain image, the basal region may be the basal ganglia of the brain.
[0122] In this step, if the maximum overlap ratio between the target lesion and the target functional area is less than the preset ratio threshold, and the centroid of the target lesion is not located in the undetermined region, it can be determined that the target lesion is not located in multiple target functional areas, and the basal region of the target image can be taken as the lesion region of the target image. For example, if the target image is a brain image, the lesion region can be determined to be the basal ganglia.
[0123] Figure 8 This is a flowchart illustrating another method for determining a lesion region according to an exemplary embodiment of the present disclosure, such as... Figure 8 As shown, after acquiring the target image, three-dimensional correction processing can be performed on the target image. The registration deformation field of the target image is determined by the registration deformation field generation model. Combined with the initial template image, multiple undetermined functional regions of the target image are determined. Referring to the methods of steps S1031 to S1032, multiple target functional regions of the target image are determined. Finally, combined with the location information of the target lesion, the lesion region of the target lesion in the target image is determined.
[0124] Using the above method, the lesion region in the target image can be determined based on the lesion location information. When the division of multiple target functional regions in the target image is more accurate, the determined lesion region in the target image is also more accurate.
[0125] Figure 9 This is a block diagram illustrating an apparatus for determining a lesion region according to an exemplary embodiment of the present disclosure, such as... Figure 9 As shown, the device may include:
[0126] The acquisition module 901 is used to acquire the registration deformation field of the target image of the lesion area to be determined and the lesion location information of the target lesion in the target image;
[0127] The first determining module 902 is used to determine the registration template image corresponding to the initial template image based on the registration deformation field. The initial template image is an image that does not contain the target lesion and other lesions and has been divided into template functional regions. The other lesions are lesions other than the target lesion.
[0128] The second determining module 903 is used to determine multiple target functional regions of the target image based on the registration template image; the target functional regions of the target image are located at the same positions as the template functional regions of the initial template image.
[0129] The third determining module 904 is used to determine the lesion region of the target lesion in the target image based on the lesion location information and multiple target functional regions.
[0130] Optionally, the target image includes a first image region and a second image region; the second determining module 903 is further configured to:
[0131] Based on the location of the template functional region in the registration template image, determine multiple undetermined functional regions of the target image;
[0132] When the location of the lesion is determined to be in the first image region based on the lesion location information, multiple undetermined functional regions of the second image region are taken as multiple target functional regions of the second image region, and multiple target functional regions of the first image region are determined based on the location of the multiple target functional regions of the second image region.
[0133] The target functional area contained in the first image region is symmetrical to the target functional area contained in the second image region.
[0134] Optionally, the third determining module 904 is further configured to:
[0135] When the organ corresponding to the target image is a first preset organ, the target functional region where the centroid of the target lesion is located is determined from multiple target functional regions based on the lesion location information, and the target functional region where the centroid of the target lesion is located is taken as the lesion region of the target image. The lesion type of the first preset organ does not include hemorrhage.
[0136] Optionally, the third determining module 904 is further configured to:
[0137] When the organ corresponding to the target image is a second preset organ, the overlap ratio between the target lesion and each target functional area of the first image region is determined according to the lesion location information, and the lesion type of the second preset organ includes hemorrhage;
[0138] The target functional region corresponding to the largest overlap ratio among multiple overlap ratios is taken as the region to be determined.
[0139] If the maximum overlap ratio is greater than or equal to a preset ratio threshold, the region to be determined is taken as the lesion region of the target image.
[0140] Optionally, Figure 10 This is a block diagram illustrating another apparatus for determining a lesion region according to an exemplary embodiment of the present disclosure, such as... Figure 10 As shown, the device also includes:
[0141] The fourth determining module 905 is used to determine whether the centroid of the target lesion is located in the undetermined region when the maximum overlap ratio is less than the preset ratio threshold.
[0142] The fifth determining module 906 is used to determine the undetermined region as the lesion region of the target image when the centroid of the target lesion is located in the undetermined region.
[0143] Optionally, Figure 11 This is a block diagram illustrating another apparatus for determining a lesion region according to an exemplary embodiment of the present disclosure, such as... Figure 11 As shown, the device also includes:
[0144] The sixth determining module 907 is used to determine the base region of the target image as the lesion region of the target image when the centroid of the target lesion is not located in the undetermined region. The base region is the region in the first image region other than the multiple target functional regions.
[0145] Optionally, the acquisition module 901 is also used for:
[0146] The target image is input into a pre-generated registration deformation field generation model to obtain the registration deformation field output by the registration deformation field generation model.
[0147] The above-mentioned device first determines multiple target functional regions of the target image based on the initial template image, and then determines the lesion region of the target lesion in the target image based on the lesion location information of the target lesion in the target image and the multiple target functional regions. In this way, the lesion region in the target image can be automatically identified without human intervention, thus improving the efficiency of lesion region identification.
[0148] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.
[0149] Figure 12 This is a block diagram illustrating an electronic device 700 according to an exemplary embodiment. Figure 12 As shown, the electronic device 700 may include a processor 701 and a memory 702. The electronic device 700 may also include one or more of a multimedia component 703, an input / output interface 704, and a communication component 705.
[0150] The processor 701 controls the overall operation of the electronic device 700 to complete all or part of the steps in the method for determining the lesion area described above. The memory 702 stores various types of data to support the operation of the electronic device 700. This data may include, for example, instructions for any application or method operating on the electronic device 700, and application-related data such as contact data, sent and received messages, pictures, audio, video, etc. The memory 702 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. Multimedia component 703 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in memory 702 or transmitted via communication component 705. The audio component also includes at least one speaker for outputting audio signals. Input / output interface 704 provides an interface between processor 701 and other interface modules, such as a keyboard, mouse, buttons, etc. These buttons may be virtual or physical buttons. Communication component 705 is used for wired or wireless communication between the electronic device 700 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, 4G, NB-IoT, eMTC, or other 5G technologies, or combinations thereof, is not limited here. Therefore, the corresponding communication component 705 may include: a Wi-Fi module, a Bluetooth module, an NFC module, etc.
[0151] In an exemplary embodiment, the electronic device 700 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the method for determining the lesion region described above.
[0152] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, which, when executed by a processor, implement the steps of the method for determining a lesion region described above. For example, the computer-readable storage medium may be the memory 702 including program instructions described above, which may be executed by the processor 701 of the electronic device 700 to complete the method for determining a lesion region described above.
[0153] In another exemplary embodiment, a computer program product is also provided, the computer program product comprising a computer program executable by a programmable device, the computer program having a code portion for performing the above-described method for determining a lesion region when executed by the programmable device.
[0154] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and these simple modifications all fall within the protection scope of this disclosure. It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, this disclosure will not further describe the various possible combinations.
[0155] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.
Claims
1. A method for determining a lesion area, characterized in that, include: Obtain the registration deformation field of the target image of the lesion region to be determined and the lesion location information of the target lesion in the target image; The target image includes a first image region and a second image region; Based on the registration deformation field, a registration template image corresponding to the initial template image is determined. The initial template image is an image that does not contain the target lesion and other lesions, and has been divided into template functional regions. The other lesions are lesions other than the target lesion. Based on the registration template image, multiple target functional regions of the target image are determined; the target functional regions of the target image are located at the same positions as the template functional regions of the initial template image. Based on the lesion location information and multiple target functional regions, the lesion region of the target lesion in the target image is determined; The step of determining the lesion region of the target lesion in the target image based on the lesion location information and multiple target functional regions includes: When the organ corresponding to the target image is a second preset organ, the overlap ratio between the target lesion and each of the target functional regions of the first image region is determined according to the lesion location information, and the lesion type of the second preset organ includes hemorrhage; The target functional region corresponding to the largest overlap ratio among the multiple overlap ratios is taken as the region to be determined. If the maximum overlap ratio is greater than or equal to a preset ratio threshold, the region to be determined is taken as the lesion region of the target image. If the maximum overlap ratio is less than the preset ratio threshold, determine whether the centroid of the target lesion is located in the undetermined region; If the centroid of the target lesion is located in the undetermined region, the undetermined region is taken as the lesion region of the target image.
2. The method according to claim 1, characterized in that, The step of determining multiple target functional regions of the target image based on the registration template image includes: Based on the position of the template functional region in the registration template image, determine multiple undetermined functional regions of the target image; When the target lesion is determined to be located in the first image region based on the lesion location information, multiple undetermined functional regions of the second image region are taken as multiple target functional regions of the second image region, and multiple target functional regions of the first image region are determined based on the location of the multiple target functional regions of the second image region. The target functional regions contained in the first image region are symmetrical in position to the target functional regions contained in the second image region.
3. The method according to claim 2, characterized in that, The step of determining the lesion region in the target image based on the lesion location information and multiple target functional regions includes: When the organ corresponding to the target image is a first preset organ, the target functional region where the centroid of the target lesion is located is determined from multiple target functional regions according to the lesion location information, and the target functional region where the centroid of the target lesion is located is taken as the lesion region of the target image. The lesion type of the first preset organ does not include hemorrhage.
4. The method according to claim 1, characterized in that, The method further includes: If the centroid of the target lesion is not located in the undetermined region, the base region of the target image is taken as the lesion region of the target image, and the base region is the region in the first image region other than the plurality of target functional regions.
5. The method according to any one of claims 1-4, characterized in that, The registration deformation field for acquiring the target image of the lesion region to be determined includes: The target image is input into a pre-generated registration deformation field generation model to obtain the registration deformation field output by the registration deformation field generation model.
6. A device for determining a lesion area, characterized in that, include: The acquisition module is used to acquire the registration deformation field of the target image of the lesion area to be determined and the lesion location information of the target lesion in the target image; The target image includes a first image region and a second image region; The first determining module is used to determine the registration template image corresponding to the initial template image based on the registration deformation field. The initial template image is an image that does not contain the target lesion and other lesions and has been divided into template functional regions. The other lesions are lesions other than the target lesion. The second determining module is used to determine multiple target functional regions of the target image based on the registration template image; the target functional regions of the target image are located at the same positions as the template functional regions of the initial template image; The third determining module is used to determine the lesion region of the target lesion in the target image based on the lesion location information and multiple target functional regions; The step of determining the lesion region of the target lesion in the target image based on the lesion location information and multiple target functional regions includes: When the organ corresponding to the target image is a second preset organ, the overlap ratio between the target lesion and each of the target functional regions of the first image region is determined according to the lesion location information, and the lesion type of the second preset organ includes hemorrhage; The target functional region corresponding to the largest overlap ratio among the multiple overlap ratios is taken as the region to be determined. If the maximum overlap ratio is greater than or equal to a preset ratio threshold, the region to be determined is taken as the lesion region of the target image. If the maximum overlap ratio is less than the preset ratio threshold, determine whether the centroid of the target lesion is located in the undetermined region; If the centroid of the target lesion is located in the undetermined region, the undetermined region is taken as the lesion region of the target image.
7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method described in any one of claims 1-5.
8. An electronic device, characterized in that, include: A memory on which computer programs are stored; A processor for executing the computer program in the memory to implement the steps of the method according to any one of claims 1-5.