Template image library generation method, target positioning method, device and storage medium

Abstract

The application relates to a template image library generation method, a target positioning method, a device and a storage medium. The template image library generation method comprises the following steps: acquiring a plurality of sample image sets of a target part with different attribute information; for each sample image set, a template image corresponding to each sample image set is constructed; a preset template image library of the target part is established according to the template images corresponding to each sample image set; wherein the preset template image library comprises a plurality of template images of the target part corresponding to different attribute information; in the embodiment of the application, a template image generation method of the target part is provided, and the practicability and operability of the template image are improved; in addition, in the embodiment, one target part corresponds to a plurality of template images, so that the classification granularity of the template image of the target part is finer, the matching degree between the template image and the to-be-detected object is higher, and the accuracy of the analysis result after the target part of the to-be-detected object is analyzed is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, in particular to a template image library generation method, a target positioning method and device, a computer device, a storage medium and a computer program product. BACKGROUND

[0002] In the medical field, with the development of medical imaging technology, shooting medical images has become an auxiliary treatment method for doctors to effectively treat the target part of a patient. With the help of medical images of the target part, doctors can quickly understand the pathological conditions of organs, tissues, blood vessels, etc. of the target part, and can also determine the position of the target area or the anatomical position during surgical treatment.

[0003] In the traditional technology, doctors analyze the medical images of patients based on their professional knowledge and experience. However, in the case of unclear development effect of medical images, severe damage to the target part, large individual structural differences of the target part, or insufficient experience of doctors, it is easy to lead to inaccurate analysis results of medical images by doctors.

[0004] Based on this, the inventors found in theoretical research and practical verification that when analyzing the medical images of the target part, more accurate analysis results can be obtained by combining the standard template images of the target part, i.e. the medical images of the target part in a healthy state. Therefore, how to obtain accurate template images has become a technical problem to be solved. SUMMARY

[0005] Therefore, it is necessary to provide a template image library generation method, a target positioning method and device, a computer device, a computer readable storage medium and a computer program product capable of improving the positioning accuracy of anatomical positions to solve the above technical problems.

[0006] In a first aspect, the present application provides a template image library generation method. The method comprises:

[0007] Obtaining a plurality of sample image sets of a target part; each sample image set corresponds to different attribute information;

[0008] For each sample image set, a template image corresponding to each sample image set is constructed;

[0009] According to the template images corresponding to each sample image set, a preset template image library of the target part is established; the preset template image library includes a plurality of template images of the target part corresponding to different attribute information.

[0010] In one embodiment, for each sample image set, a template image corresponding to each sample image set is constructed, comprising:

[0011] For each sample image set, a multi-level clustering operation is performed on the sample images in the sample image set to obtain at least one sample subset corresponding to each level of clustering operation; the sample subset includes a plurality of sample images satisfying a clustering condition;

[0012] For each sample subset corresponding to each level of clustering operation, a first template image corresponding to each sample subset is determined according to each sample image in each sample subset corresponding to each level of clustering operation;

[0013] According to the first template image corresponding to each sample subset, a second template image corresponding to each level of clustering operation is determined;

[0014] According to the second template image corresponding to each level of clustering operation, a template image corresponding to the sample image set is determined.

[0015] In one embodiment, a multi-level clustering operation is performed on the sample images in the sample image set to obtain at least one sample subset corresponding to each level of clustering operation, including:

[0016] A first level of clustering operation is performed on the sample images in the sample image set to obtain a plurality of candidate sample subsets corresponding to the first level of clustering operation;

[0017] It is judged whether each candidate sample subset satisfies a clustering condition; the clustering condition includes that the number of sample images in the candidate sample subset is greater than or equal to a preset threshold;

[0018] In the case where it is judged that the candidate sample subset satisfies the clustering condition, the candidate sample subset is taken as a sample subset corresponding to the first level of clustering operation;

[0019] In the case where it is judged that the candidate sample subset does not satisfy the clustering condition, a second level of clustering operation is performed on each candidate sample subset that does not satisfy the clustering condition until a clustering termination condition is satisfied, to obtain at least one sample subset corresponding to each level of clustering operation.

[0020] In one embodiment, according to each sample image in each sample subset corresponding to each level of clustering operation, a first template image corresponding to each sample subset is determined, including:

[0021] For each sample subset corresponding to each level of clustering operation, each sample image in the sample subset is input into a preset registration model to obtain an intermediate sample image corresponding to each sample image in the sample subset, respectively;

[0022] The intermediate sample images corresponding to each sample image in the sample subset are fused to obtain a first template image corresponding to the sample subset.

[0023] In one of the embodiments, each sample image in the sample subset is input into a preset registration model to obtain an intermediate sample image corresponding to each sample image in the sample subset, including:

[0024] A reference sample image is determined from the sample images in the sample subset;

[0025] The reference sample image and each floating sample image are input into a preset registration model to obtain a deformation field of each floating sample image mapped to the reference sample image; the floating sample image is a sample image in the sample subset other than the reference sample image;

[0026] For each floating sample image, the floating sample image is registered according to the deformation field corresponding to the floating sample image to obtain an intermediate sample image corresponding to each floating sample image.

[0027] In one of the embodiments, a reference sample image is determined from the sample images in the sample subset, including:

[0028] The mean square error between a first sample image and a second sample image other than the first sample image in the sample subset is calculated respectively; the first sample image is any sample image in the sample subset;

[0029] The mean square errors are summed to obtain a summation result of the mean square errors;

[0030] The smallest summation result is determined from the summation result, and the first sample image corresponding to the smallest summation result is taken as the reference sample image.

[0031] In one of the embodiments, the intermediate sample images corresponding to the sample images in the sample subset are fused to determine a first template image corresponding to the sample subset, including:

[0032] The intermediate sample images corresponding to the sample images in the sample subset are mean-processed to obtain the first template image corresponding to the sample subset.

[0033] In one of the embodiments, a second template image corresponding to the hierarchical clustering operation is determined according to the first template images corresponding to the sample subsets, including:

[0034] The first template images corresponding to the sample subsets are input into a preset registration model to obtain an intermediate template image corresponding to each first template image;

[0035] The intermediate template images corresponding to the first template images are fused to obtain the second template image corresponding to the hierarchical clustering operation.

[0036] In one of the embodiments, the template image corresponding to the sample image set is determined according to the second template image corresponding to each level of clustering operation, comprising:

[0037] The second template image corresponding to each level of clustering operation is respectively input into a preset registration model to obtain an intermediate template image corresponding to each second template image;

[0038] The intermediate template images corresponding to each second template image are fused to obtain the template image corresponding to the sample image set.

[0039] In a second aspect, the application further provides a target positioning method. The method comprises:

[0040] Obtaining a medical image of a target part of a to-be-tested object;

[0041] According to the attribute information of the to-be-tested object, a target template image of the target part matching the attribute information of the to-be-tested object is obtained from a preset template image library;

[0042] According to the medical image of the target part and the target template image of the target part, a to-be-positioned target in the target part is positioned to obtain position information of the to-be-positioned target; wherein the preset template image library is generated by the method in the first aspect.

[0043] In one of the embodiments, according to the medical image of the target part and the target template image of the target part, the to-be-positioned target in the target part is positioned to obtain the position information of the to-be-positioned target, comprising:

[0044] The medical image of the target part and the target template image of the target part are registered to obtain a registered target medical image;

[0045] According to the target medical image, the to-be-positioned target in the target part is positioned to obtain the position information of the to-be-positioned target.

[0046] In a third aspect, the application further provides a template image library generation device. The device comprises:

[0047] An acquisition module is configured to acquire a plurality of sample image sets of a target part; each sample image set corresponds to different attribute information;

[0048] A construction module is configured to construct a template image corresponding to each sample image set for each sample image set;

[0049] An establishment module is configured to establish a preset template image library of the target part according to the template images corresponding to each sample image set; the preset template image library comprises a plurality of template images of the target part corresponding to different attribute information.

[0050] In a fourth aspect, the present application provides a target positioning device. The device comprises:

[0051] a first obtaining module, configured to obtain a medical image of a target part of a to-be-tested object;

[0052] a second obtaining module, configured to obtain a target template image of the target part from a preset template image library according to attribute information of the to-be-tested object, the target template image being matched with the attribute information of the to-be-tested object;

[0053] a determining module, configured to position a to-be-positioned target in the target part according to the medical image of the target part and the target template image of the target part, and obtain position information of the to-be-positioned target; wherein the preset template image library is generated by using the method in the first aspect.

[0054] In a fifth aspect, the present application provides a computer device. The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the method in the first aspect and the second aspect when executing the computer program.

[0055] In a sixth aspect, the present application provides a computer readable storage medium. The computer readable storage medium stores a computer program, and the computer program implements the steps of the method in the first aspect and the second aspect when executed by a processor.

[0056] In a seventh aspect, the present application provides a computer program product. The computer program product comprises a computer program, and the computer program implements the steps of the method in the first aspect and the second aspect when executed by a processor.

[0057] The method for generating the template image library, the target positioning method, the device and the storage device are characterized in that: a plurality of sample image sets with different attribute information of a target part are acquired, and for each sample image set, a template image corresponding to the sample image set is constructed; then, a preset template image library of the target part is established according to the template images corresponding to the sample image sets; and the preset template image library includes a plurality of template images of the target part corresponding to different attribute information. That is, the method for generating the template image of the target part is provided in the embodiment, and the practicability and operability of the template image are improved. Compared with directly analyzing the medical image of the target part by artificial means, the accuracy of the analysis result after analyzing the medical image of the target part can be improved by combining the template image. In addition, in the embodiment, the preset template image library of the target part includes a plurality of template images corresponding to different attribute information, that is, one target part corresponds to a plurality of template images, so that the classification granularity of the template image of the target part is finer, the application range of the template image can be improved, that is, different template images corresponding to the target part can be applied to different objects to be measured, and the matching degree between the template image and different objects to be measured can be improved, and the accuracy of the analysis result after analyzing the target part of different objects to be measured is further improved. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 An application environment diagram of the method for generating the template image library in one embodiment;

[0059] Figure 2 A flowchart of the method for generating the template image library in one embodiment;

[0060] Figure 3 A flowchart of the method for generating the template image library in another embodiment;

[0061] Figure 4 A flowchart of the method for generating the template image library in another embodiment;

[0062] Figure 5 A flowchart of the method for generating the template image library in another embodiment;

[0063] Figure 6 A flowchart of the method for generating the template image library in another embodiment;

[0064] Figure 7 A flowchart of the method for generating the template image library in another embodiment;

[0065] Figure 8 A flowchart of the method for generating the template image library in another embodiment;

[0066] Figure 9 a structural block diagram of a template image library generation device in an embodiment;

[0067] Figure 10 a structural block diagram of a target positioning device in an embodiment. DETAILED DESCRIPTION

[0068] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application is further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0069] The template image library generation method provided by the embodiments of the present application can be applied to a medical image terminal, or a server connected with the medical image terminal, or even a medical image scanning device and a series of computer devices related to medical image processing. The internal structural diagram of the computer device can be as shown in Figure 1 The computer device includes a processor, a memory, and a communication interface connected through a system bus. Optionally, it can also include a display screen and an input device. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The communication interface of the computer device is used to communicate with other computer devices in a wired or wireless manner. The wireless manner can be achieved through WIFI, mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by the processor to implement a template image library generation method. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc. In addition, in the case of the computer device being a server, the server can be implemented by an independent server or a server cluster composed of multiple servers.

[0070] Those skilled in the art can understand that Figure 1 the structure shown in the figure is only a block diagram of part of the structure related to the present application scheme, and does not constitute a limitation on the computer device to which the present application scheme is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0071] In one embodiment, as shown in Figure 2 a template image library generation method is provided. The method is applied to Figure 1The computer device in the embodiment is taken as an example to illustrate the method, including the following steps.

[0072] In step 201, a plurality of sample image sets of a target part are acquired; each sample image set corresponds to different attribute information.

[0073] Optionally, the target part can be any medical tissue part of a to-be-tested object, such as a knee joint, a hip joint, an elbow joint, a chest, a lung, a brain, etc., and the attribute information can be attribute information of the to-be-tested object, including but not limited to gender, age, nationality of the to-be-tested object, and category information of the target part of the to-be-tested object, etc. The category information of the target part of the to-be-tested object can be position information of the target part, for example, in the case of the target part being a knee joint, the position information of the target part can include a left knee joint and a right knee joint. Different attribute information means that at least one attribute information is different in the plurality of attribute information of the to-be-tested object.

[0074] Optionally, the computer device can divide a plurality of empty subsets with different attribute information according to the attribute information, and for each empty subset, a plurality of sample images corresponding to the attribute information of the empty subset are acquired to form a sample image set corresponding to the empty subset. Thus, the computer device can acquire a plurality of sample image sets corresponding to the target part with different attribute information. Optionally, when acquiring sample images corresponding to different attribute information of the target part, historical scanning images for the target part can be acquired from medical institutions in different regions, or scanning images corresponding to different attribute information of the target part can be acquired from a large database as sample images, which is not limited in the present application.

[0075] In step 202, a template image corresponding to each sample image set is constructed for each sample image set.

[0076] Optionally, for each sample image set, image feature extraction can be performed on each sample image in the sample image set, and a template image corresponding to the sample image set is constructed according to the extracted image features; alternatively, clustering processing can be performed on the sample image set in a clustering manner to obtain at least one sample subset corresponding to the sample image set, and then each sample subset is analyzed to obtain a candidate template image corresponding to each sample subset, and then a template image corresponding to the sample image set can be constructed according to the candidate template image corresponding to each sample subset.

[0077] In step 203, a preset template image library of the target part is established according to the template images corresponding to each sample image set.

[0078] The preset template image library includes a plurality of template images of the target part corresponding to different attribute information.

[0079] In the embodiment, a plurality of sample image sets with different attribute information of the target part are obtained, a template image corresponding to each sample image set is constructed for each sample image set, and then a preset template image library of the target part is established according to the template images corresponding to each sample image set. The preset template image library includes a plurality of template images of the target part corresponding to different attribute information. That is, the embodiment provides a method for generating a template image of a target part, which improves the implementability and operability of the template image. Compared with directly analyzing the medical image of the target part by artificial means, the accuracy of the analysis result after analyzing the medical image of the target part can be improved by combining the template image. In addition, in the embodiment, the preset template image library of the target part includes a plurality of template images corresponding to different attribute information, that is, one target part corresponds to a plurality of template images, so that the classification granularity of the template image of the target part is finer. The template image of the target part can not only improve the application range of the template image, that is, different template images corresponding to the target part can be applied to different objects to be measured, but also improve the matching degree between the template image and different objects to be measured, and further improve the accuracy of the analysis result after analyzing the target part of different objects to be measured.

[0080] Figure 3 The flowchart of the method for generating the template image library in another embodiment is shown. The embodiment relates to an optional implementation process of constructing a template image corresponding to each sample image set for each sample image set. Based on the above embodiment, as shown in Figure 3 the step 202 includes:

[0081] In step 301, for each sample image set, a multi-level clustering operation is performed on the sample images in the sample image set to obtain at least one sample subset corresponding to each level of clustering operation.

[0082] Optionally, each sample image in the sample image set can be unsupervised clustered in an unsupervised clustering manner, and a clustering condition and a clustering termination condition can be set. The clustering condition can be a judgment condition for each level of clustering result. The judgment condition for each level of clustering result can be related to the number of images of the clustering subset obtained after clustering, or can be other judgment conditions for the clustering subset obtained after clustering. In addition, the clustering termination condition can be the number of clustering levels, or a termination condition related to the clustering subset obtained after each level of clustering, for example, the clustering subset obtained after clustering includes only one clustering subset, and the clustering subset does not satisfy the judgment condition of each level of clustering result, so that the next level of clustering cannot be performed.

[0083] In an optional implementation manner of the embodiment, as shown in Figure 4As shown, the step 301 can also include:

[0084] The step 401 performs a first-level clustering operation on the sample images in the sample image set to obtain a plurality of candidate sample subsets corresponding to the first-level clustering operation.

[0085] In an optional implementation, the computer device can regard each sample image in the sample image set as a category cluster, and set a similarity threshold; calculate the similarity between each pair of category clusters in the sample image set, and merge two category clusters with the maximum similarity greater than the similarity threshold to obtain a plurality of merged candidate sample subsets; wherein the similarity of the category cluster can be the pixel similarity between the category clusters (i.e. the sample images), that is, the gray value or signal value between each pair of pixel points in the two sample images is measured to determine the similarity between the two sample images. Optionally, the similarity calculation method can include but is not limited to nearest-neighbor, complete-linkage, average-linkage, etc.

[0086] The step 402 judges whether each candidate sample subset meets the clustering condition; the clustering condition here can include that the number of sample images in the candidate sample subset is greater than or equal to a preset threshold.

[0087] The step 403, in the case where the candidate sample subset meets the clustering condition, takes the candidate sample subset as the sample subset corresponding to the first-level clustering operation.

[0088] The step 404, in the case where the candidate sample subset does not meet the clustering condition, performs a second-level clustering operation on each candidate sample subset that does not meet the clustering condition until the clustering termination condition is met, to obtain at least one sample subset corresponding to each level of clustering operation.

[0089] That is, in the case where the number of sample images in the candidate sample subset is less than the preset threshold, the cluster merging process is continued on the candidate sample subset, that is, the next level of clustering operation is performed; and so on until the clustering termination condition is met, that is, the number of clustering levels is reached, or the cluster merging cannot be performed any more.

[0090] The following is illustrated by a specific embodiment: assuming that the sample image set P1, after the first level clustering operation, the obtained candidate sample subset includes a1, a2, a3, a4, a5, a6, a7, a8, a9, a10, assuming that a1 is the candidate sample subset satisfying the clustering condition, a1 can be determined as the sample subset corresponding to the first level clustering operation; and for a2, a3, a4, a5, a6, a7, a8, a9, a10 which do not satisfy the clustering condition, the second level clustering operation is continued, assuming that after the second level clustering operation, the obtained candidate sample subset includes b1 (merging a2 and a4), b2 (merging a3 and a7), b3 (merging a5 and a6), b4 (merging a8, a9 and a10), assuming that b1 and b4 are the candidate sample subsets satisfying the clustering condition, b1 and b4 can be determined as the sample subsets corresponding to the second level clustering operation; then, for b2 and b3 which do not satisfy the clustering condition, the third level clustering operation is continued, the candidate sample subset c1 (merging b2 and b3) after the third level clustering operation is obtained, if c1 satisfies the clustering condition at this time, c1 is determined as the sample subset corresponding to the third level clustering operation; thus, the sample subset a1 corresponding to the first level clustering operation, the sample subsets b1 and b4 corresponding to the second level clustering operation, and the sample subset c1 corresponding to the third level clustering operation can be obtained.

[0091] It should be noted that the hierarchical clustering manner provided above by adopting cluster merging (i.e. agglomerative) is only used for explanation and illustration, in actual application, for the multi-level clustering operation, a divisive hierarchical clustering manner can also be used, the implementation manner of the embodiments of the present application is not limited; in addition, since the divisive hierarchical clustering manner is prior art, the specific implementation process of the divisive hierarchical clustering manner will not be described herein.

[0092] In step 302, for at least one sample subset corresponding to each level clustering operation, a first template image corresponding to each sample subset is determined according to each sample image in the sample subset corresponding to the level clustering operation.

[0093] Optionally, for each sample image in each sample subset of each level, each sample image can be input into a preset template image model to obtain a first template image corresponding to each sample subset. Based on the above example, for the sample subset a1 corresponding to the first level clustering, the first template image T1 corresponding to the sample subset a1 can be obtained; for the sample subsets b1 and b4 corresponding to the second level clustering, the first template image T2 corresponding to the sample subset b1 and the first template image T3 corresponding to the sample subset b4 can be obtained; for the sample subset c1 corresponding to the third level clustering operation, the first template image T4 corresponding to the sample subset c1 can be obtained.

[0094] In an optional implementation of the embodiment, as shown in FIG. 3, the step 302 can also include: Figure 5

[0095] In step 501, for each sample subset corresponding to the level clustering operation, each sample image in the sample subset is input into a preset registration model to obtain an intermediate sample image corresponding to each sample image in the sample subset.

[0096] The preset registration model is obtained by training an initial registration network based on a reference sample image and a plurality of floating sample images. Optionally, the computer device can first obtain a training data set, wherein the training data set includes a plurality of pairs of training samples, each pair of training samples being composed of a reference sample image and a different floating sample image. Optionally, the training samples can be determined from each of the sample subsets, and the reference sample image can be determined from each sample image in a sample subset, and the sample images other than the reference sample image in the sample subset are used as floating sample images. Then, the reference sample image and a floating sample image are used as a pair of training samples. In the same way, a plurality of pairs of training samples can be obtained from a plurality of sample subsets. In addition, it should be noted that each sample subset can be each sample subset in the same sample image set, or each sample subset in different sample image sets.

[0097] Then, the computer device can construct an initial registration network (such as an unsupervised deep learning registration network). Optionally, the unsupervised deep learning registration network can be formed by cascading an encoder and a decoder. The encoder can be composed of a plurality of convolution scales, and the decoder is symmetrical to the encoder structure. In the case that the image modality of the training sample is CT or MR, a three-dimensional convolution kernel can be selected. In the case that the image modality of the training sample is X-ray, a two-dimensional convolution kernel can be selected. Optionally, the encoder can be composed of 5 convolution scales, wherein the convolution scale refers to a series of convolution operations with different convolution kernel sizes.

[0098] During model training, the plurality of pairs of training samples can be sequentially input into the unsupervised deep learning registration network to output a deformation field corresponding to each pair of training samples for image change. Then, the obtained deformation field can be used in a spatial transformation manner to act on the pair of training samples corresponding to the deformation field to obtain a registered image corresponding to the floating sample image in the pair of training samples. Further, for the pair of training samples, the registered image and the reference sample image are input into a preset loss function to calculate the difference between the two. The preset loss function can be composed of a mean square error loss function (such as MSE Loss Function) and a smoothing term, as shown in formula (1).

[0099] ​

[0100] wherein the smoothing term I fixed is a reference sample image, I i is a floating sample image.

[0101] Then, the network hyperparameters are adjusted and set as needed, the preset loss function is continuously converged, the best image segmentation effect is achieved, and the trained network model is saved as the preset registration model. Optionally, a minimum loss value can be preset, and in a case where the loss value of the preset loss function is less than or equal to the minimum loss value, it can be determined that the preset loss function has reached convergence. Optionally, in a case where the network hyperparameters remain unchanged for a long time or the change amplitude is less than or equal to a preset amplitude threshold, it can be determined that the preset loss function has reached convergence, and the trained network model is stable. At this time, the trained network model can be used as the preset registration model. In this embodiment, the unsupervised deep learning registration network is used, which can avoid the tedious process of manually outlining the registration area, improve the training effect of the registration model, and further improve the construction efficiency of the standard template, compared with the supervised registration network.

[0102] Optionally, after obtaining the preset registration model, for each sample subset, the method steps shown in FIG. 6 can be used to obtain the intermediate sample images corresponding to each sample image in the sample subset, including: Figure 6

[0103] Step 601, determining a reference sample image from each sample image in the sample subset.

[0104] Optionally, any sample image in the sample subset can be determined as the reference sample image. A preset screening algorithm can also be used to determine the reference sample image from each sample image. The preset screening algorithm can be flexibly set according to actual use requirements, and the present application does not make specific limitations thereto.

[0105] Step 602, inputting the reference sample image and each floating sample image into the preset registration model to obtain a deformation field of each floating sample image mapped to the reference sample image; wherein the floating sample image is a sample image other than the reference sample image in the sample subset.

[0106] Step 603, registering each floating sample image according to the deformation field corresponding to the floating sample image to obtain an intermediate sample image corresponding to each floating sample image.

[0107] Based on the above example, it is assumed that the sample subset a1 includes sample images [I1, I2, I3,..., In], the sample subset a2 includes sample images [I1, I2, I3,..., In], and the sample subset a3 includes sample images [I1, I2, I3,..., In]. N ​, where the reference sample image I_fixed is I1, and the floating sample images are [I2, I3, …, I N , I1 and I2, I1 and I3, …, I1 and I N are input into the preset registration model respectively, and finally the intermediate sample images corresponding to each floating sample image are obtained, i.e., I2 corresponds to I2 ’ , I3 corresponds to I3 ’ , …, and I N corresponds to I N ’ . It should be noted that the intermediate sample image corresponding to the reference sample image I_fixed is still the reference sample image, i.e., I1 = I1 ’ .

[0108] Step 502: performing fusion processing on the intermediate sample images corresponding to each sample image in the sample subset to obtain a first template image corresponding to the sample subset.

[0109] Optionally, the fusion processing can include, but is not limited to, mean value processing, median value processing, etc.; wherein, when the mean value processing or the median value processing is performed, the mean value processing or the median value processing can also be performed after removing the maximum value / minimum value.

[0110] Preferably, the computer device can perform mean value processing on the intermediate sample images corresponding to each sample image in the sample subset to obtain a first template image corresponding to the sample subset; optionally, the mean value processing includes, but is not limited to, direct averaging, weighted averaging, etc. Based on the above example, for the intermediate sample images corresponding to the sample subset a1, i.e., [I_fixed, I2 ’ , I3 ’ , …, I N ’ , performing mean value processing can obtain a first template image T1 corresponding to the sample subset a1.

[0111] Step 303: determining a second template image corresponding to the hierarchical clustering operation according to the first template images corresponding to each sample subset.

[0112] Optionally, the first template images corresponding to each sample subset can be input into the preset registration model respectively to obtain intermediate template images corresponding to each first template image; then, performing fusion processing on the intermediate template images corresponding to each first template image to obtain a second template image corresponding to the hierarchical clustering operation. For the specific implementation process, reference can be made to the steps of each method shown in the above Figure 5 and Figure 6 , which will not be described here again.

[0113] It should be noted that in the case that the level clustering operation corresponds to only one sample subset, the first template image corresponding to the sample subset can be taken as the second template image corresponding to the level clustering operation.

[0114] Based on the above example, for the first level clustering operation, since the sample subset corresponding to the first level clustering operation only includes a1, the first template image T1 corresponding to the sample subset a1 can be taken as the second template image U1 corresponding to the first level clustering operation. For the second level clustering operation, the first template image T2 corresponding to the sample subset b1 and the first template image T3 corresponding to the sample subset b4 can be input into the preset registration model to obtain an intermediate template image corresponding to the first template image T2 and an intermediate template image corresponding to the first template image T3; then, the intermediate template image corresponding to the first template image T2 and the intermediate template image corresponding to the first template image T3 are subjected to mean value processing to obtain the second template image U2 corresponding to the second level clustering operation. For the third level clustering operation, since the sample subset corresponding to the third level clustering operation only includes c1, the first template image T3 corresponding to the sample subset c1 can be taken as the second template image U3 corresponding to the third level clustering operation.

[0115] In step 304, the template image corresponding to the sample image set is determined according to the second template image corresponding to each level clustering operation.

[0116] Alternatively, the second template image corresponding to each level clustering operation can be input into the preset registration model respectively to obtain an intermediate template image corresponding to each second template image; then, the intermediate template images corresponding to the second template images are subjected to fusion processing to obtain the template image corresponding to the sample image set. For the specific implementation process, reference can be made to the various method steps shown in the above Figure 6 and Figure 7 Here, no further elaboration is given.

[0117] Based on the above example, the second template image U1 corresponding to the first level clustering operation, the second template image U2 corresponding to the second level clustering operation, and the second template image U3 corresponding to the third level clustering operation can be subjected to the above processing to obtain the template image corresponding to the sample image set P1.

[0118] Similarly, for each sample image set corresponding to different attribute information, the template image corresponding to each sample image set of different attribute information can be obtained according to the above related steps, and then the preset template image library can be generated according to the template images corresponding to the sample image sets of different attribute information.

[0119] In this embodiment, for each sample image set, the computer device obtains at least one sample subset corresponding to each level clustering operation by performing multi-level clustering operation on the sample images in the sample image set; and for each sample subset corresponding to each level clustering operation, determines a first template image corresponding to the sample subset according to each sample image in the sample subset corresponding to the level clustering operation; then, determines a second template image corresponding to the level clustering operation according to the first template image corresponding to each sample subset; and determines a template image corresponding to the sample image set according to the second template image corresponding to each level clustering operation; the image set registration method layer by layer can reduce registration errors and noise generated in image fusion, realize high-precision standard template construction, improve the accuracy of the standard template, and further improve the accuracy of target positioning according to the high-precision template image.

[0120] Figure 7 The flowchart of the generation method of the template image library in another embodiment is shown. The embodiment relates to one of the optional implementation processes in which the computer device determines the reference sample image from each sample image in the sample subset, as shown in Figure 7 The step 701 includes the following steps.

[0121] In step 701, the mean square error between the first sample image and the second sample image other than the first sample image in the sample subset is calculated.

[0122] The first sample image is any sample image in the sample subset.

[0123] The existing image mean square error calculation method can be used here. The image mean square error is well known to those skilled in the art, and therefore, the embodiment of the present application does not make too much explanation.

[0124] In step 702, the mean square errors are summed to obtain the summation result of the mean square errors.

[0125] In step 703, the smallest summation result is determined from the summation result, and the first sample image corresponding to the smallest summation result is taken as the reference sample image.

[0126] In this embodiment, the computer device calculates the mean square error between the first sample image and the second sample image other than the first sample image in the sample subset respectively, then sums the mean square errors to obtain a sum result of the mean square errors, further determines the smallest sum result from the sum result, and takes the first sample image corresponding to the smallest sum result as the reference sample image; wherein the first sample image is any sample image in the sample subset; that is, in this embodiment, the sum of the mean square errors between each sample image and other sample images is calculated, and then the sample image corresponding to the smallest mean square error sum is determined from the plurality of mean square error sums as the reference sample image, the difference between the reference sample image and other sample images is the smallest, so that the quality of the image after registration of other sample images in the sample subset according to the reference sample image is higher, and the accuracy of the template image determined according to the registered images is higher, which can improve the accuracy of target positioning according to the template image.

[0127] The template image generation method provided in each of the above embodiments can generate a template image library corresponding to different target parts respectively, and the template image library of the target part can be used for assisting in analyzing the medical image of the target part; for example, the physiological characteristics of the target part can be analyzed in combination with the template image library of the target part and the medical image of the target part, and a certain positioning target in the target part can be positioned and analyzed in combination with the template image library of the target part and the medical image of the target part. The following will be specifically described by taking the accurate positioning analysis of the positioning target in the target part as an example.

[0128] In one embodiment, as shown in Figure 8 , a target positioning method is provided, which will be described by taking the computer device in Figure 1 as an example, comprising the following steps:

[0129] Step 801, obtaining the medical image of the target part of the to-be-measured object.

[0130] Optionally, the computer device can obtain the medical image of the target part of the to-be-measured object from the image scanning device, or from the server storing the medical image of the target part of the to-be-measured object, or from the local storage of the computer device. The medical image stored in the local storage can be the medical image of the target part of the to-be-measured object scanned by the image scanning device and sent to the computer device. The application does not limit the acquisition method of the medical image.

[0131] Step 802, according to the attribute information of the to-be-measured object, obtaining the target template image of the target part matched with the attribute information of the to-be-measured object from the preset template image library.

[0132] The preset template image library includes a plurality of template images of target parts corresponding to different attribute information, and the preset template image library can be generated by the method of any one of the above embodiments. Figures 2 to 7 The attribute information of the to-be-tested object can include, but is not limited to, gender, age, nationality of the to-be-tested object, and category information of the target part of the to-be-tested object, and the like. The category information of the target part of the to-be-tested object can be position information of the target part, for example, left knee joint or right knee joint. Accordingly, the preset template image library can include template images of knee joints corresponding to different attribute information, and at least one attribute information corresponding to different template images is different; for example, a first template image is a template image corresponding to gender (male), age (20), nationality (Han), and position (left knee joint); a second template image is a template image corresponding to gender (female), age (20), nationality (Han), and position (left knee joint); a third template image is a template image corresponding to gender (male), age (20), nationality (Han), and position (right knee joint), and the like.

[0133] Optionally, for the template images of the target parts corresponding to different attribute information, a plurality of sample medical images of the target parts of different objects under the same attribute information can be obtained, and the sample medical images are all medical images of the target parts in a healthy state. Then, the template image of the target part corresponding to the attribute information can be determined by using a plurality of sample medical images. Optionally, when the template image of the target part corresponding to the attribute information is determined by using a plurality of sample medical images, a common feature of the plurality of sample medical images can be learned by using an existing deep learning technology to obtain the template image of the target part corresponding to the attribute information. Of course, the plurality of sample medical images can also be processed by using an existing image processing technology, such as image fusion, to obtain the template image of the target part corresponding to the attribute information. It should be noted that the application embodiments do not make specific limitations on the acquisition method of the template image of the target part. Further, after obtaining the template images of the target parts corresponding to different attribute information, the preset template image library of the target part can be established according to the template images of the target parts corresponding to different attribute information.

[0134] Further, when the target part of the to-be-tested object is anatomically positioned, the computer device can obtain the attribute information of the to-be-tested object input by the user, or obtain the attribute information of the to-be-tested object from the medical record of the to-be-tested object. Then, the computer device can obtain the target template image of the target part corresponding to the attribute information of the to-be-tested object from the above-mentioned preset template image library according to the attribute information of the to-be-tested object.

[0135] In step 803, the to-be-positioned target in the target part is positioned according to the medical image of the target part and the target template image of the target part, and position information of the to-be-positioned target is obtained.

[0136] The to-be-positioned target can be an organ or tissue in the target part, for example, in the anterior cruciate ligament reconstruction of the knee joint, the to-be-positioned target can be the anterior cruciate ligament in the knee joint, that is, the position information of the anterior cruciate ligament is determined through the medical image of the knee joint, and then the anatomical position in the anterior cruciate ligament reconstruction is determined.

[0137] Optionally, the initial position information of the to-be-positioned target can be determined from the target template image of the target part, and then the initial position information of the to-be-positioned target is matched to the medical image of the target part to obtain the position information of the to-be-positioned target in the medical image. Optionally, the medical image of the target part and the target template image of the target part can also be registered to obtain the registered target medical image, and then the to-be-positioned target in the target part is positioned according to the registered target medical image to obtain the position information of the to-be-positioned target. It should be noted that the registration between the medical image and the target template image can use existing medical image registration technology, for example, the image registration technology realized by rigid transformation, affine transformation, projection transformation or nonlinear transformation, etc. The embodiments of the present application do not make specific limitations thereto.

[0138] In the target positioning method, the computer device acquires a medical image of a target part of the to-be-measured object, and acquires a target template image of the target part matching attribute information of the to-be-measured object from a preset template image library according to the attribute information of the to-be-measured object; then, the to-be-positioned target in the target part is positioned according to the medical image of the target part and the target template image of the target part, and position information of the to-be-positioned target is obtained; wherein the preset template image library includes a plurality of template images of target parts corresponding to different attribute information; that is, the target positioning method in the embodiment of the application determines the position of the to-be-positioned target in the target part by combining the standard template image corresponding to the target part of the to-be-measured object; since the positions of various tissues and organs in the standard template image of the target part are obvious and accurate, compared with directly determining the position information of the to-be-positioned target from the medical image of the target part, no matter how the development effect of the position of the to-be-positioned target in the medical image is or how the damage of the position of the to-be-positioned target is, more accurate position information of the to-be-positioned target can be obtained based on the standard template image. In addition, since the preset template image library of the target part in the embodiment of the application includes template images corresponding to different attribute information, the target template image matching the attribute information of the to-be-measured object can be obtained from the preset template image library of the target part according to the attribute information of the to-be-measured object, the matching degree between the target template image and the to-be-measured object is improved, and the positioning accuracy of the to-be-positioned target in the target part of the to-be-measured object can be further improved.

[0139] It should be understood that, although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least part of other steps or steps or stages in other steps.

[0140] Based on the same inventive concept, the embodiment of the application also provides a template image library generation device for implementing the above-mentioned template image library generation method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more template image library generation device embodiments provided below can refer to the limitations of the template image library generation method in the above text, which will not be repeated here.

[0141] In one embodiment, as shown in Figure 9 A template image library generation apparatus is provided, comprising: an acquisition module 901, a construction module 902, and an establishment module 903, wherein:

[0142] The acquisition module 901 is configured to acquire a plurality of sample image sets of the target part; each sample image set corresponds to different attribute information.

[0143] The construction module 902 is configured to construct, for each sample image set, a template image corresponding to the sample image set.

[0144] The establishment module 903 is configured to establish a preset template image library of the target part according to the template images corresponding to each sample image set.

[0145] In one embodiment, the construction module 902 comprises a clustering unit, a first determination unit, a second determination unit, and a third determination unit; wherein the clustering unit is configured to perform a multi-level clustering operation on the sample images in each sample image set to obtain at least one sample subset corresponding to each level of clustering operation; the sample subset comprises a plurality of sample images satisfying a clustering condition; the first determination unit is configured to determine, for each sample subset corresponding to each level of clustering operation, a first template image corresponding to the sample subset according to the sample images in the sample subset corresponding to the level of clustering operation; the second determination unit is configured to determine a second template image corresponding to the level of clustering operation according to the first template images corresponding to each sample subset; and the third determination unit is configured to determine a template image corresponding to the sample image set according to the second template images corresponding to each level of clustering operation.

[0146] In one embodiment, the clustering unit is specifically configured to perform a first level of clustering operation on the sample images in the sample image set to obtain a plurality of candidate sample subsets corresponding to the first level of clustering operation; determine whether each candidate sample subset satisfies a clustering condition; in a case where it is determined that a candidate sample subset satisfies the clustering condition, take the candidate sample subset as a sample subset corresponding to the first level of clustering operation; in a case where it is determined that a candidate sample subset does not satisfy the clustering condition, perform a second level of clustering operation on each candidate sample subset that does not satisfy the clustering condition until a clustering stop condition is satisfied, to obtain at least one sample subset corresponding to each level of clustering operation; wherein the clustering condition comprises that the number of sample images in the candidate sample subset is greater than or equal to a preset threshold.

[0147] In one of the embodiments, the first determining unit is specifically configured to input each sample image in each sample subset into a preset registration model to obtain an intermediate sample image corresponding to each sample image in the sample subset; and perform fusion processing on the intermediate sample images corresponding to each sample image in the sample subset to obtain the first template image corresponding to the sample subset.

[0148] In one of the embodiments, the first determining unit is specifically configured to determine a reference sample image from the sample images in the sample subset; input the reference sample image and each floating sample image into a preset registration model to obtain a deformation field of each floating sample image mapped to the reference sample image; the floating sample image is a sample image in the sample subset other than the reference sample image; and perform registration on each floating sample image according to the deformation field corresponding to the floating sample image to obtain an intermediate sample image corresponding to each floating sample image.

[0149] In one of the embodiments, the first determining unit is specifically configured to calculate a mean square error between a first sample image and a second sample image other than the first sample image in the sample subset; sum the mean square errors to obtain a summation result of the mean square errors; determine a minimum summation result from the summation result; and take the first sample image corresponding to the minimum summation result as the reference sample image; wherein the first sample image is any sample image in the sample subset.

[0150] In one of the embodiments, the first determining unit is specifically configured to perform mean value processing on the intermediate sample images corresponding to each sample image in the sample subset to obtain the first template image corresponding to the sample subset.

[0151] In one of the embodiments, the second determining unit is specifically configured to input the first template images corresponding to each sample subset into a preset registration model to obtain an intermediate template image corresponding to each first template image; and perform fusion processing on the intermediate template images corresponding to each first template image to obtain the second template image corresponding to the level clustering operation.

[0152] In one of the embodiments, the third determining unit is specifically configured to input the second template images corresponding to each level clustering operation into a preset registration model to obtain an intermediate template image corresponding to each second template image; and perform fusion processing on the intermediate template images corresponding to each second template image to obtain the template image corresponding to the sample image set.

[0153] Similarly, based on the same inventive concept, the embodiments of the present application also provide a target positioning device for implementing the target positioning method described above. The implementation scheme of the device for solving the problem is similar to the implementation scheme described in the above method, so the specific limitations in one or more target positioning device embodiments provided below can refer to the limitations of the target positioning method described above, which will not be repeated here.

[0154] In one embodiment, as shown in Figure 10 a target positioning device is provided, comprising: a first acquisition module 1001, a second acquisition module 1002 and a determination module 1003, wherein:

[0155] The first acquisition module 1001 is configured to acquire a medical image of a target part of a to-be-measured object.

[0156] The second acquisition module 1002 is configured to acquire a target template image of the target part corresponding to attribute information of the to-be-measured object from a preset template image library according to the attribute information of the to-be-measured object; the preset template image library comprises a plurality of template images of target parts corresponding to different attribute information.

[0157] The determination module 1003 is configured to position a to-be-positioned target in the target part according to the medical image of the target part and the target template image of the target part, and obtain position information of the to-be-positioned target.

[0158] In one embodiment, the determination module 1003 comprises a registration unit and a determination unit; the registration unit is configured to register the medical image of the target part and the target template image of the target part to obtain a registered target medical image; and the determination unit is configured to position the to-be-positioned target in the target part according to the target medical image to obtain the position information of the to-be-positioned target.

[0159] The modules in the above template image library generation device and target positioning device can be realized by software, hardware and combinations thereof, in whole or in part. The modules can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the above modules.

[0160] In one embodiment, a computer device is provided, which can be a medical image terminal, a server connected with the medical image terminal, a medical image scanning device, or a series of computer devices related to medical image processing, and its internal structure diagram can be as shown in Figure 1 .

[0161] In an embodiment, a computer device is provided, comprising a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the template image library generation method and the target positioning method provided in the above embodiments when executing the computer program.

[0162] In an embodiment, a computer readable storage medium is provided, storing a computer program, and the computer program implementing the steps of the template image library generation method and the target positioning method provided in the above embodiments when executed by a processor.

[0163] In an embodiment, a computer program product is provided, comprising a computer program, and the computer program implementing the steps of the template image library generation method and the target positioning method provided in the above embodiments when executed by a processor.

[0164] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties.

[0165] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0166] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist contradictory, it should be considered as the scope of the present application.

[0167] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for generating a template image library, characterized in that, The method includes: Multiple sample image sets of the target area are acquired; each sample image set corresponds to different attribute information; For each of the aforementioned sample image sets, a multi-level clustering operation is performed on the sample images in the sample image set to obtain at least one sample subset corresponding to each level of clustering operation; the sample subset includes multiple sample images that satisfy the clustering conditions. For at least one sample subset corresponding to each level of clustering operation, a first template image corresponding to each sample subset is determined based on the sample images in each sample subset corresponding to each level of clustering operation. Based on the first template image corresponding to each of the sample subsets, determine the second template image corresponding to each level of clustering operation; Based on the second template image corresponding to each level of clustering operation, determine the template image corresponding to the sample image set; Based on the template images corresponding to each of the sample image sets, a preset template image library for the target region is established; the preset template image library includes multiple template images of the target region corresponding to different attribute information; The step of determining the first template image corresponding to each sample subset based on the sample images in each sample subset corresponding to each level of clustering operation includes: For each sample subset corresponding to each level of clustering operation, the sample images in the sample subset are respectively input into a preset registration model to obtain the intermediate sample images corresponding to each sample image in the sample subset. The intermediate sample images corresponding to each sample image in the sample subset are fused to obtain the first template image corresponding to the sample subset.

2. The method according to claim 1, characterized in that, The step of performing multi-level clustering operations on the sample images in the sample image set to obtain at least one sample subset corresponding to each level of clustering operation includes: Perform a first-level clustering operation on the sample images in the sample image set to obtain multiple candidate sample subsets corresponding to the first-level clustering operation; Determine whether each of the candidate sample subsets satisfies the clustering condition; the clustering condition includes that the number of sample images in the candidate sample subset is greater than or equal to a preset threshold; If so, the candidate sample subset is taken as the sample subset corresponding to the first-level clustering operation; If not, then a second-level clustering operation is performed on each candidate sample subset that does not meet the clustering conditions until the clustering cutoff condition is met, thereby obtaining at least one sample subset corresponding to each level of clustering operation.

3. The method according to claim 1, characterized in that, The step of inputting each sample image in the sample subset into a preset registration model to obtain intermediate sample images corresponding to each sample image in the sample subset includes: From each sample image in the sample subset, a reference sample image is determined; The reference sample image and each floating sample image are respectively input into a preset registration model to obtain the deformation field of each floating sample image mapped to the reference sample image; the floating sample image is the sample image in the sample subset other than the reference sample image; For each of the floating sample images, the floating sample images are registered according to the deformation field corresponding to the floating sample images to obtain intermediate sample images corresponding to each of the floating sample images.

4. The method according to claim 3, characterized in that, Determining the reference sample image from each sample image in the sample subset includes: Calculate the mean square error between the first sample image in the sample subset and the second sample image other than the first sample image; wherein the first sample image is any sample image in the sample subset. The mean square errors are summed to obtain the summation result of the mean square errors; The smallest summation result is determined from the summation results, and the first sample image corresponding to the smallest summation result is used as the reference sample image.

5. The method according to claim 1, characterized in that, The step of fusing the intermediate sample images corresponding to each sample image in the sample subset to determine the first template image corresponding to the sample subset includes: The intermediate sample images corresponding to each sample image in the sample subset are averaged to obtain the first template image corresponding to the sample subset.

6. The method according to claim 1, characterized in that, The step of determining the second template image corresponding to each level of clustering operation based on the first template image corresponding to each of the sample subsets includes: The first template image corresponding to each of the sample subsets is input into the preset registration model to obtain the intermediate template image corresponding to each of the first template images. The intermediate template images corresponding to each of the first template images are fused to obtain the second template image corresponding to each level of clustering operation.

7. The method according to claim 1, characterized in that, The step of determining the template image corresponding to the sample image set based on the second template image corresponding to each level of clustering operation includes: The second template image corresponding to each level of clustering operation is input into a preset registration model to obtain the intermediate template image corresponding to each second template image. The intermediate template images corresponding to each of the second template images are fused to obtain the template images corresponding to the sample image set.

8. A target localization method, characterized in that, The method includes: Acquire medical images of the target area of ​​the subject; Based on the attribute information of the object to be tested, a target template image of the target part that matches the attribute information of the object to be tested is obtained from a preset template image library; Based on the medical images of the target area and the target template image of the target area, the target to be located in the target area is located to obtain the location information of the target to be located. The preset template image library is generated using the method described in any one of claims 1 to 7.

9. A template image library generation device, characterized in that, The device includes: The acquisition module is used to acquire multiple sample image sets of the target area; each sample image set corresponds to different attribute information; A construction module is configured to perform multi-level clustering operations on sample images in each of the sample image sets to obtain at least one sample subset corresponding to each level of clustering operation; each sample subset includes multiple sample images that satisfy clustering conditions; for each sample subset corresponding to each level of clustering operation, a first template image corresponding to each sample subset is determined based on each sample image in each sample subset corresponding to each level of clustering operation; a second template image corresponding to each level of clustering operation is determined based on the first template image corresponding to each sample subset; and a template image corresponding to the sample image set is determined based on the second template image corresponding to each level of clustering operation. A module is established to create a preset template image library for the target region based on the template images corresponding to each of the sample image sets; the preset template image library includes multiple template images of the target region corresponding to different attribute information; The step of determining the first template image corresponding to each sample subset based on the sample images in each sample subset corresponding to each level of clustering operation includes: For each sample subset corresponding to each level of clustering operation, the sample images in the sample subset are respectively input into a preset registration model to obtain the intermediate sample images corresponding to each sample image in the sample subset. The intermediate sample images corresponding to each sample image in the sample subset are fused to obtain the first template image corresponding to the sample subset.

10. A target positioning device, characterized in that, The device includes: The first acquisition module is used to acquire medical images of the target area of ​​the object to be tested; The second acquisition module is used to acquire a target template image of the target part that matches the attribute information of the object to be tested from a preset template image library based on the attribute information of the object to be tested. The determination module is used to locate the target to be located in the target area based on the medical image of the target area and the target template image of the target area, and obtain the location information of the target to be located. The preset template image library is generated using the method described in any one of claims 1 to 7.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 8.

Images