A multi-data center training system for endometrial cancer target segmentation models

Abstract

The present invention relates to a multi-data center endometrial cancer target segmentation model training system, belonging to the field of target area delineation technology, and solving the problem of low segmentation model performance in the prior art. The system includes a master control terminal and multiple data centers: each data center stores a local multimodal endometrial cancer sample set; the master control terminal stores a target sample set; a target area segmentation model is constructed on the master control terminal and sent to each data center; the master control terminal controls the multiple data centers to train local target area segmentation models based on the local multimodal endometrial cancer sample set, and updates the parameters of the master control terminal's target area segmentation model based on the parameters of the local target area segmentation models of the multiple data centers to obtain a pre-trained target area segmentation model; the master control terminal fine-tunes the pre-trained target area segmentation model based on the target sample set to obtain a trained endometrial cancer target area segmentation model. Efficient and accurate endometrial cancer target area delineation is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of target area segmentation, and in particular to a training system for an endometrial cancer target area segmentation model in a multi-data center. Background Art

[0002] Endometrial cancer is one of the most common malignant tumors of the female reproductive system, and its incidence is on the rise worldwide. In China, with the aging of the population and changes in lifestyle, the number of new cases of endometrial cancer is also increasing. This makes accurate diagnosis and effective treatment of endometrial cancer an important medical need. For patients with endometrial cancer, post-operative target segmentation is crucial in the formulation of subsequent treatment plans (such as radiotherapy plans). Accurate target segmentation can ensure that treatments such as radiotherapy act precisely on the tumor area, reduce damage to surrounding normal tissues, improve the patient's quality of life and improve prognosis.

[0003] The amount of data collected from endometrial cancer patients at different medical centers is small and variably present. Training a model solely on data collected from a single center can lead to overfitting due to the small sample size, resulting in inaccurate image semantic segmentation, and thus target segmentation, leading to poor model performance. If data from various centers are combined for training, consistent delineation rules are required across centers. When delineation rules differ or the delineation styles differ between centers, training performance is typically poor, resulting in poor model performance. Summary of the Invention

[0004] In view of the above analysis, an embodiment of the present invention aims to provide a multi-data center endometrial cancer target segmentation model training system to solve the problem of low performance of existing segmentation models.

[0005] On the one hand, an embodiment of the present invention provides a multi-data center endometrial cancer target segmentation model training system, the system comprising a master control terminal and multiple data centers:

[0006] Each of the data centers stores a local multimodal endometrial cancer sample set; the main control terminal stores a target sample set;

[0007] Constructing a target area segmentation model on the master control end, and sending the target area segmentation model to each of the data centers;

[0008] The master control end controls the multiple data centers to train the local target segmentation model based on the local multimodal endometrial cancer sample set, and updates the parameters of the target segmentation model of the master control end based on the local target segmentation model parameters of the multiple data centers to obtain a pre-trained target segmentation model;

[0009] The main control end fine-tunes the pre-trained target segmentation model based on the target sample set to obtain a trained endometrial cancer target segmentation model.

[0010] Based on further improvements to the above system, the master control end controls the multiple data centers to train the local target segmentation model based on the local multimodal endometrial cancer sample set in the following manner, and updates the parameters of the target segmentation model of the master control end based on the local target segmentation model parameters of the multiple data centers to obtain a pre-trained target segmentation model:

[0011] S21. The master control terminal randomly selects N data centers and sends training instructions to the selected N data centers;

[0012] S22. After receiving the training instruction, each selected data center trains the local target segmentation model based on the local multimodal endometrial cancer sample set, calculates the training loss of the local target segmentation model, and sends the calculated training loss to the master control end;

[0013] S23. The main control end updates the parameters of the local target segmentation model according to the training losses of the selected N data centers, and sends the updated parameters of the local target segmentation model to each data center; determines whether the current stopping condition is met, and if so, stops training to obtain the pre-trained target segmentation model; otherwise, the main control end executes step S21.

[0014] Based on further improvements to the above system, the master terminal updates the parameters of the local target segmentation model based on the training losses of the selected N data centers, including:

[0015] The master end calculates the pre-updated parameters of each selected data center based on the training losses of the selected N data centers, and sends the pre-updated parameters of each selected data center to the corresponding data center;

[0016] Each selected data center pre-updates the local target segmentation model according to the pre-update parameters, calculates the pre-update loss based on the pre-updated local target segmentation model, and sends the pre-update loss to the master control end;

[0017] The master calculates the weights of N data centers based on their pre-update losses.

[0018] Update the parameters of the local target segmentation model based on the training losses and weights of the N data centers.

[0019] Based on the further improvement of the above system, the master end calculates the pre-update parameters of each selected data center based on the training losses of the selected N data centers, including:

[0020] Obtaining first pre-updated parameters of the N selected data centers based on the training losses of the N selected data centers;

[0021] For the i-th selected data center, a second pre-updated parameter of the i-th selected data center obtained based on the training losses of the other N-1 selected data centers;

[0022] The first pre-updated parameter and the second pre-updated parameter of the i-th selected data center constitute the pre-updated parameter of the i-th selected data center.

[0023] Based on the further improvement of the above system, each selected data center pre-updates the local target segmentation model according to the pre-update parameters, and calculates the pre-update loss based on the pre-updated local target segmentation model, including:

[0024] Each selected data center pre-updates the local target segmentation model based on the first pre-update parameter and the second pre-update parameter to obtain a first pre-update model and a second pre-update model;

[0025] The loss is calculated based on the first pre-updated model on the test set of the local multimodal endometrial cancer sample set to obtain the first pre-updated loss;

[0026] The second pre-updated loss is obtained by calculating the loss on the test set of the local multimodal endometrial cancer sample set based on the second pre-updated model;

[0027] The difference between the second pre-update loss and the first pre-update loss yields the pre-update loss.

[0028] Based on the further improvement of the above system, the weight of each selected data center is calculated based on the pre-updated loss of N data centers using the following formula:

[0029]

[0030] Among them, q i represents the data quality of the i-th selected data center, δ i represents the pre-update loss of the i-th selected data center, and γ represents the adjustment parameter.

[0031] Based on a further improvement of the above system, the multimodal endometrial cancer sample set includes a training set and a test set; the training set is used to train the local target segmentation model, and the test set is used to calculate the training loss of the local target segmentation model;

[0032] The multimodal endometrial cancer sample includes a pair of plain scan CT images and enhanced CT images of endometrial cancer; some multimodal endometrial cancer samples also include mask images corresponding to the pair of plain scan CT images and enhanced CT images of endometrial cancer;

[0033] The training loss is calculated using the following formula:

[0034] Loss = L1 + L2 + L3

[0035] Among them, L1 represents the multimodal image segmentation loss, L2 represents the multimodal cross loss, and L3 represents the contrast loss.

[0036] Based on the further improvement of the above system, the multimodal cross loss is calculated using the following formula:

[0037]

[0038] in, represents the number of samples without corresponding mask images in the test set of the i-th data center, P k1 (s, j) represents the element value at position (s, j) in the first segmented image of the kth sample that has no corresponding mask image, P k2 (s, j) represents the element value at position (s, j) in the second segmented image of the kth sample without a corresponding mask image, and m represents the number of elements in the first segmented image.

[0039] Based on the further improvement of the above system, the contrast loss is calculated in the following way:

[0040] For each sample in the test set without a corresponding mask image, the first segmented image and the second segmented image of the sample are used as a positive sample pair; the first segmented image of the sample and the second segmented image of other samples in the test set without a corresponding mask image are used as negative sample pairs; the positive sample pair and the negative sample pair corresponding to each sample in the test set without a corresponding mask image are obtained;

[0041] The contrast loss is calculated based on the positive and negative pairs corresponding to each sample in the test set that does not have a corresponding mask image.

[0042] Based on the further improvement of the above system, the contrast loss is calculated based on the positive sample pair and negative sample pair corresponding to each sample in the test set without a corresponding mask image in the following way:

[0043]

[0044] in, represents the number of samples without corresponding mask images in the test set of the i-th data center, P k1 represents the first segmented image of the kth sample without a corresponding mask image, P k2 represents the second segmented image of the kth sample without a corresponding mask image, P j2represents the second segmented image of the jth sample without a corresponding mask image, τ represents the adjustment parameter, and dice(·,·) represents the DICE loss.

[0045] Compared with the existing technology, the present invention improves data security by constructing a main control end and multiple data centers, and the main control end and multiple data centers maintain their own sample sets. At the same time, the main control end constructs a target area segmentation model and sends it to each data center; the main control end controls multiple data centers to pre-train local target area segmentation models based on local sample sets, and then updates the parameters of the target area segmentation model of the main control end to realize the training of the target area segmentation model of the main control end and obtain a pre-trained target area segmentation model. The main control end then fine-tunes the pre-trained target area segmentation model on the target data set, thereby quickly training to obtain a target area segmentation model suitable for the target data set, and will not affect the performance of the model due to the small amount of data in the target data set, thereby improving the segmentation performance of the model on the target data set.

[0046] In the present invention, the above-mentioned technical solutions can be combined with each other to achieve more preferred combinations. Other features and advantages of the present invention will be described in the following description, and some advantages will become apparent from the description or be learned through practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are only used for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. Throughout the drawings, the same reference symbols denote the same components.

[0048] Figure 1 This is a block diagram of a training system for a multi-data center endometrial cancer target segmentation model according to an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, and are not used to limit the scope of the present invention.

[0050] A specific embodiment of the present invention discloses a multi-data center endometrial cancer target segmentation model training system, such as Figure 1 As shown, the system includes a master control terminal and multiple data centers;

[0051] Each of the data centers stores a local multimodal endometrial cancer sample set; the main control terminal stores a target sample set;

[0052] Constructing a target area segmentation model on the master control end, and sending the target area segmentation model to each of the data centers;

[0053] The master control end controls the multiple data centers to train the local target segmentation model based on the local multimodal endometrial cancer sample set, and updates the parameters of the target segmentation model of the master control end based on the local target segmentation model parameters of the multiple data centers to obtain a pre-trained target segmentation model;

[0054] The main control end fine-tunes the pre-trained target segmentation model based on the target sample set to obtain a trained endometrial cancer target segmentation model.

[0055] During implementation, each data center collected multimodal image data of postoperative endometrial cancer patients in the center to construct a multimodal endometrial cancer sample set.

[0056] The control server also has a local multimodal endometrial cancer sample set, known as the target dataset. This sample set is a small dataset. Training a model solely based on a small dataset is prone to overfitting, resulting in poor model performance.

[0057] Compared with the existing technology, the training system of the multi-data center endometrial cancer target segmentation model provided in this embodiment improves data security by constructing a main control end and multiple data centers. The main control end and multiple data centers maintain their own sample sets. At the same time, the main control end constructs a target segmentation model and sends it to each data center; the main control end controls multiple data centers to pre-train local target segmentation models based on local sample sets, and then updates the parameters of the target segmentation model of the main control end to train the target segmentation model of the main control end and obtain a pre-trained target segmentation model. The main control end then fine-tunes the pre-trained target segmentation model on the target data set, thereby quickly training to obtain a target segmentation model suitable for the target data set, and will not affect the performance of the model due to the small amount of data in the target data set, thereby improving the segmentation performance of the model on the target data set.

[0058] Specifically, the sample set for each data center includes a training set and a test set. This means that a portion of the sample set is used to form the training set, while another portion is used to form the test set. The training set is used to train the local target segmentation model, while the test set is used to calculate the training loss of the local target segmentation model.

[0059] Each multimodal endometrial cancer sample includes a pair of plain CT images and enhanced CT images of endometrial cancer. Some multimodal endometrial cancer samples also include mask images corresponding to the plain CT images and enhanced CT images.

[0060] During implementation, due to the high workload and low efficiency of manual delineation, some samples in the sample set are allowed to have no labeled data, that is, no corresponding mask images. Therefore, the multimodal endometrial cancer sample set includes paired plain and enhanced CT images of endometrial cancer, as well as the mask images corresponding to some paired plain and enhanced CT images of endometrial cancer.

[0061] The target segmentation model can adopt a deep learning model.

[0062] Specifically, the target segmentation model includes a first segmentation module and a second segmentation module; the first segmentation module and the second segmentation module have the same structure;

[0063] The first segmentation module is used to perform target area segmentation on the plain scan CT image to obtain a first segmented image;

[0064] The second segmentation module is used to perform target segmentation on the enhanced CT image to obtain a first segmented image.

[0065] The master end sends the constructed target segmentation model to each data center, including the model structure and model parameters. Initially, the parameters of the target segmentation model of the master end and the target segmentation model of each data center are the same.

[0066] During implementation, the first segmentation module and the second segmentation module may adopt a 3D-Unet structure.

[0067] Specifically, the master control end controls multiple data centers to train local target segmentation models based on a local multimodal endometrial cancer sample set in the following manner, and updates the parameters of the master control end's target segmentation model based on the parameters of the local target segmentation models of the multiple data centers to obtain a pre-trained target segmentation model:

[0068] S21. The master control terminal randomly selects N data centers and sends training instructions to the selected N data centers;

[0069] S22. After receiving the training instruction, each selected data center trains the local target segmentation model based on the local multimodal endometrial cancer sample set, calculates the training loss of the local target segmentation model, and sends the calculated training loss to the master control end;

[0070] S23. The main control end updates the parameters of the local target segmentation model according to the training losses of the selected N data centers, and sends the updated parameters of the local target segmentation model to each data center; determines whether the current stopping condition is met, and if so, stops training to obtain the pre-trained target segmentation model; otherwise, the main control end executes step S21.

[0071] During implementation, the master control first randomly selects N data centers for the current round of model training, sending training instructions to the selected N data centers. At this point, the parameters of the target segmentation model on the master control and the target segmentation models of each data center are the same, marked as the initial parameters ф.

[0072] For each selected data center, after receiving the training instruction, the local target segmentation model is first trained based on the training set in the local multimodal endometrial cancer sample set, and the parameters of the local target segmentation model are updated. When implemented, multi-step training can be performed, that is, the parameters of the local target segmentation model are updated multiple times, and the parameters of the last update are recorded as θ, θ i Represents the parameters of the local target segmentation model of the i-th data center.

[0073] After updating the parameters of the local target segmentation model, each selected data center calculates the training loss on the test set based on the updated local target segmentation model.

[0074] During implementation, for the i-th data center, the samples on the test set are input into the local target segmentation model (the model parameter is θ i ), and calculate the training loss based on the first segmentation image and the second segmentation image output by the model.

[0075] Specifically, for the i-th data center, the training loss Loss is calculated based on the updated local target segmentation model using the following formula:

[0076] Loss=L1+L2+L3 (1)

[0077] Among them, L1 represents the multimodal image segmentation loss, L2 represents the multimodal cross loss, and L3 represents the contrast loss.

[0078] Specifically, for samples with labels, that is, samples with corresponding mask images, the segmentation loss is calculated based on the first segmentation image and the second segmentation image obtained by the first segmentation module and the second segmentation module.

[0079] Specifically, the multimodal image segmentation loss is calculated using the following formula:

[0080]

[0081] in, Indicates the number of samples with corresponding mask images in the test set of the i-th data center, P k1 represents the first segmented image of the kth sample with a corresponding mask image in the test set, G1 represents the mask image corresponding to the plain scan CT image of the kth sample with a corresponding mask image in the test set, P k2represents the second segmented image of the k-th sample with a corresponding mask image in the test set, G2 represents the mask image corresponding to the enhanced CT image of the k-th sample with a corresponding mask image in the test set, and α1 and α2 represent weight parameters.

[0082] During implementation, α1 and α2 are used to control false negatives and false positives, respectively. Adjusting α1 and α2 can control the trade-off between false negatives and false positives. k1 ∩G1| represents P k1 and the intersection of G1.

[0083] For samples without labels, that is, without corresponding mask images, the loss is calculated through multimodal mutual learning and contrastive learning to obtain multimodal cross loss and contrastive loss.

[0084] Specifically, the multimodal cross loss is calculated using the following formula:

[0085]

[0086] in, represents the number of samples without corresponding mask images in the test set of the i-th data center, P k1 (s, j) represents the element value at position (s, j) in the first segmented image of the kth sample that has no corresponding mask image, P k2 (s, j) represents the element value at position (s, j) in the second segmented image of the kth sample without a corresponding mask image, and m represents the number of elements in the first segmented image.

[0087] Through multimodal cross loss, the images segmented by the first segmentation module and the second segmentation module are trained to be as similar as possible.

[0088] Specifically, the contrast loss is calculated in the following way:

[0089] For each sample in the test set without a corresponding mask image, the first segmented image and the second segmented image of the sample are used as a positive sample pair; the first segmented image of the sample and the second segmented image of other samples in the test set without a corresponding mask image are used as negative sample pairs; the positive sample pair and the negative sample pair corresponding to each sample in the test set without a corresponding mask image are obtained;

[0090] The contrast loss is calculated based on the positive and negative pairs corresponding to each sample in the test set that does not have a corresponding mask image.

[0091] For example, there are samples without corresponding mask images in the test set of the i-th data center. indivual.

[0092] For the kth sample without a corresponding mask image, the first segmentation image P predicted by the model is k1 and the second segmented image P k2 As the positive sample pair corresponding to the sample (P k1 ,P k2 ). Then, the first segmented image P k1 The second segmented images of other samples in the test set constitute the negative sample pair corresponding to the sample, (P k1 ,P j2 ),j≠k.

[0093] Then, based on the positive and negative pairs corresponding to each sample in the test set that does not have a corresponding mask image, the contrast loss is calculated in the following way:

[0094]

[0095] in, represents the number of samples without corresponding mask images in the test set of the i-th data center, P k1 represents the first segmented image of the kth sample without a corresponding mask image, P k2 represents the second segmented image of the kth sample without a corresponding mask image, P j2 represents the second segmented image of the jth sample without a corresponding mask image, τ represents the adjustment parameter, and dice(·,·) represents the DICE loss.

[0096] Through contrast loss, the distance between positive samples is made closer and the distance between negative samples is made farther, thereby deepening the mutual learning between different modalities and improving the segmentation performance of the model.

[0097] After each selected data center calculates the training loss of the local target segmentation model, the calculated training loss is sent to the master control end.

[0098] The master control end updates the parameters of the local target segmentation model according to the training losses of the selected N data centers, and sends the updated parameters of the local target segmentation model to each data center for synchronous update.

[0099] Specifically, the master end updates the parameters of the local target segmentation model based on the training losses of the selected N data centers, including:

[0100] S231. The master control end calculates the pre-update parameters of each selected data center based on the training losses of the selected N data centers, and sends the pre-update parameters of each selected data center to the corresponding data center;

[0101] S232. Each selected data center pre-updates the local target segmentation model according to the pre-update parameters, calculates a pre-update loss based on the pre-updated local target segmentation model, and sends the pre-update loss to the master control end;

[0102] S233. The master control end calculates the weights of the N data centers based on the pre-update losses of the N data centers;

[0103] S234. Update the parameters of the local target segmentation model based on the training losses of the N data centers and the weights of the N data centers.

[0104] Specifically, the master terminal calculates the pre-update parameters of each selected data center based on the training losses of the selected N data centers, including:

[0105] S2311. Obtain first pre-update parameters of the N selected data centers based on the training loss calculation of the N selected data centers;

[0106] S2312: For the i-th selected data center, obtain a second pre-updated parameter of the i-th selected data center based on the training losses of the other N-1 selected data centers;

[0107] The first pre-updated parameter and the second pre-updated parameter of the i-th selected data center constitute the pre-updated parameter of the i-th selected data center.

[0108] Specifically, the first pre-update parameter is calculated based on the losses of the N selected data centers using the following formula:

[0109]

[0110] in, represents the gradient of the total loss L(φ), L(φ) represents the total loss of N selected data centers, Loss(θ i ) represents the loss of the i-th selected data center, β represents the learning rate, φ represents the initial parameter, and φ′1 represents the first pre-update parameter.

[0111] For the i-th selected data center, the second pre-updated parameter of the i-th selected data center is obtained using the following formula based on the losses of the other N-1 selected data centers:

[0112]

[0113] in, Represents the total loss L i2 The gradient of (φ), L i2 (φ) represents the total loss of N-1 selected data centers, Loss(θ j) represents the loss of the i-th selected data center, β represents the learning rate, φ represents the initial parameter, φ′ i2 represents the second pre-updated parameter of the i-th selected data center.

[0114] The first pre-update parameter in each selected data is the same.

[0115] The master control terminal sends the pre-update parameters corresponding to each selected data center to the corresponding selected data center. Each selected data center pre-updates the local target segmentation model based on the pre-update parameters, calculates the pre-update loss based on the pre-updated local target segmentation model, and sends the pre-update loss to the master control terminal.

[0116] Specifically, step S232 includes:

[0117] Each selected data center pre-updates the local target segmentation model based on the first pre-update parameter and the second pre-update parameter to obtain a first pre-update model and a second pre-update model;

[0118] The loss is calculated based on the first pre-updated model on the test set of the local multimodal endometrial cancer sample set to obtain the first pre-updated loss;

[0119] The second pre-updated loss is obtained by calculating the loss on the test set of the local multimodal endometrial cancer sample set based on the second pre-updated model;

[0120] The difference between the second pre-update loss and the first pre-update loss yields the pre-update loss.

[0121] During implementation, for the i-th selected data center, the parameters of the local target segmentation model are updated to the first pre-updated parameters to obtain the first pre-updated model (parameter is φ′1), and the parameters of the local target segmentation model are updated to the second pre-updated parameters to obtain the second pre-updated model (parameter is φ′ i2 ).

[0122] During implementation, a first pre-update loss is calculated on the local test set based on the first pre-update model, and a second pre-update loss is calculated on the local test set based on the second pre-update model. Both the first pre-update loss and the second pre-update loss can be calculated according to formula (1). The pre-update loss of the i-th selected data center is obtained by subtracting the first pre-update loss from the second pre-update loss.

[0123] The i-th selected data center sends the calculated pre-update loss to the master control end, and the master control end calculates the weights of the N data centers based on the pre-update losses of the N data centers.

[0124] During implementation, due to differences in sample size and data quality across data centers, the losses of N data centers have different weights on the initial parameter update. By calculating weights based on the data quality of each data center and dynamically calculating model parameters during each backpropagation, the weights of high-quality data centers are increased, thereby improving model segmentation accuracy and performance.

[0125] Specifically, the master terminal calculates the weight of each selected data center based on the pre-update loss of each selected data center using the following formula:

[0126]

[0127] Among them, q i represents the data quality of the i-th selected data center, δ i represents the pre-update loss of the i-th selected data center, and γ represents the adjustment parameter.

[0128] After obtaining the weights, the master end uses the following formula to update the parameters of the local target segmentation model based on the training losses and weights of the N data centers:

[0129]

[0130] in, represents the gradient of the total training loss L(φ) of the N selected data centers, L(φ) represents the total training loss of the N selected data centers, Loss(θ i ) represents the training loss of the i-th selected data center, ω i represents the weight of the i-th selected data center, β represents the learning rate, and φ represents the parameters of the local target segmentation model.

[0131] It should be noted that Loss(θ i ) represents the training loss of the i-th selected data center, which is the loss calculated in step S22, i.e., in the parameter θ i The training loss is calculated when .

[0132] After the master end updates the parameters of the local target segmentation model, it sends the updated parameters of the local target segmentation model to each data center, updates the parameters of the local target segmentation model in each data center, and thus completes the back propagation of the target segmentation model.

[0133] After completing the back propagation of the target segmentation model, if the current stopping condition is met, the training ends and the pre-trained target segmentation model is obtained. If not, the training returns to step S21 and continues until the stopping condition of the model training is met. In implementation, the stopping condition can be reaching a preset number of training times or the total training loss reaches a preset accuracy.

[0134] After obtaining the pre-trained target segmentation model, the control system fine-tunes it on the target dataset to quickly obtain a trained endometrial cancer target segmentation model. By inputting the multimodal endometrial cancer images to be segmented from the control system into the trained endometrial cancer target segmentation model, the target images are quickly and accurately obtained. This improves the accuracy and efficiency of target segmentation and avoids the risk of overfitting in small sample sizes.

[0135] Those skilled in the art will appreciate that all or part of the process steps of the above-described embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, such as a magnetic disk, an optical disk, a read-only memory, or a random access memory.

[0136] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with this technical field within the technical scope disclosed by the present invention should be covered by the scope of protection of the present invention.

Claims

1. A multi-data center endometrial cancer target segmentation model training system, characterized by: The system includes a main control terminal and multiple data centers: Each of the data centers stores a local multimodal endometrial cancer sample set; the main control terminal stores a target sample set; Constructing a target area segmentation model on the master control end, and sending the target area segmentation model to each of the data centers; The master control end controls the multiple data centers to train the local target segmentation model based on the local multimodal endometrial cancer sample set, and updates the parameters of the target segmentation model of the master control end based on the local target segmentation model parameters of the multiple data centers to obtain a pre-trained target segmentation model; The master control end fine-tunes the pre-trained target segmentation model based on the target sample set to obtain a trained endometrial cancer target segmentation model; The multimodal endometrial cancer sample set includes a training set and a test set; the training set is used to train a local target segmentation model, and the test set is used to calculate the training loss of the local target segmentation model; The multimodal endometrial cancer sample includes a pair of plain scan CT images and enhanced CT images of endometrial cancer; some multimodal endometrial cancer samples also include mask images corresponding to the pair of plain scan CT images and enhanced CT images of endometrial cancer; The training loss is calculated using the following formula: Loss = L1 + L2 + L3 Among them, L1 represents multimodal image segmentation loss, L2 represents multimodal cross loss, and L3 represents contrast loss; The contrast loss is calculated in the following way: For each sample in the test set without a corresponding mask image, the first segmented image and the second segmented image of the sample are used as a positive sample pair; the first segmented image of the sample and the second segmented image of other samples in the test set without a corresponding mask image are used as negative sample pairs; the positive sample pair and the negative sample pair corresponding to each sample in the test set without a corresponding mask image are obtained; The contrast loss is calculated based on the positive and negative pairs corresponding to each sample in the test set that does not have a corresponding mask image in the following way: in, represents the number of samples without corresponding mask images in the test set of the i-th data center, P k1 represents the first segmented image of the kth sample without a corresponding mask image, P k2 represents the second segmented image of the kth sample without a corresponding mask image, P j2 represents the second segmented image of the jth sample without a corresponding mask image, τ represents the adjustment parameter, and dice(·,·) represents the DICE loss.

2. The multi-data center endometrial cancer target segmentation model training system according to claim 1, characterized in that: The master control end controls the multiple data centers to train the local target segmentation model based on the local multimodal endometrial cancer sample set in the following manner, and updates the parameters of the target segmentation model of the master control end based on the local target segmentation model parameters of the multiple data centers to obtain a pre-trained target segmentation model: S21. The master control terminal randomly selects N data centers and sends training instructions to the selected N data centers; S22. After receiving the training instruction, each selected data center trains the local target segmentation model based on the local multimodal endometrial cancer sample set, calculates the training loss of the local target segmentation model, and sends the calculated training loss to the master control end; S23. The main control end updates the parameters of the local target segmentation model according to the training losses of the selected N data centers, and sends the updated parameters of the local target segmentation model to each data center; determines whether the current stopping condition is met, and if so, stops training to obtain the pre-trained target segmentation model; otherwise, the main control end executes step S21.

3. The multi-data center endometrial cancer target segmentation model training system according to claim 2, characterized in that: The master terminal updates the parameters of the local target segmentation model based on the training losses of the selected N data centers, including: The master end calculates the pre-updated parameters of each selected data center based on the training losses of the selected N data centers, and sends the pre-updated parameters of each selected data center to the corresponding data center; Each selected data center pre-updates the local target segmentation model according to the pre-update parameters, calculates the pre-update loss based on the pre-updated local target segmentation model, and sends the pre-update loss to the master control end; The master calculates the weights of N data centers based on their pre-update losses. Update the parameters of the local target segmentation model based on the training losses and weights of the N data centers.

4. The multi-data center endometrial cancer target segmentation model training system according to claim 3, characterized in that: The master end calculates the pre-updated parameters of each selected data center based on the training losses of the selected N data centers, including: Obtaining first pre-updated parameters of the N selected data centers based on the training losses of the N selected data centers; For the i-th selected data center, a second pre-updated parameter of the i-th selected data center obtained based on the training losses of the other N-1 selected data centers; The first pre-updated parameter and the second pre-updated parameter of the i-th selected data center constitute the pre-updated parameter of the i-th selected data center.

5. The multi-data center endometrial cancer target segmentation model training system according to claim 3, characterized in that: Each selected data center pre-updates the local target segmentation model according to the pre-update parameters and calculates the pre-update loss based on the pre-updated local target segmentation model, including: Each selected data center pre-updates the local target segmentation model based on the first pre-update parameter and the second pre-update parameter to obtain a first pre-update model and a second pre-update model; The loss is calculated based on the first pre-updated model on the test set of the local multimodal endometrial cancer sample set to obtain the first pre-updated loss; The second pre-updated loss is obtained by calculating the loss on the test set of the local multimodal endometrial cancer sample set based on the second pre-updated model; The difference between the second pre-update loss and the first pre-update loss yields the pre-update loss.

6. The multi-data center endometrial cancer target segmentation model training system according to claim 3, characterized in that: The weight of each selected data center is calculated based on the pre-update loss of N data centers using the following formula: Among them, q i represents the data quality of the i-th selected data center, δ i represents the pre-update loss of the i-th selected data center, and γ represents the adjustment parameter.

7. The multi-data center endometrial cancer target volume segmentation model training system according to claim 1, characterized in that: The multimodal cross loss is calculated using the following formula: in, represents the number of samples without corresponding mask images in the test set of the i-th data center, P k1 (s, j) represents the element value at position (s, j) in the first segmented image of the kth sample that has no corresponding mask image, P k2 (s, j) represents the element value at position (s, j) in the second segmented image of the kth sample without a corresponding mask image, and m represents the number of elements in the first segmented image.

Images