Medical image processing method and system, prediction method, electronic equipment and medium

By preprocessing multiple liver medical images and training a deep learning model, and utilizing segmentation and classification task models, the problem of low classification results for liver medical images was solved, achieving high-precision liver medical image classification.

CN118968116AInactive Publication Date: 2024-11-15HANGZHOU PUJIAN MEDICAL TECH CO LTD
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202410872862.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-01
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The classification results of liver medical images in the current technology have low precision and accuracy.

Method used

By acquiring multiple phases of liver medical images, preprocessing them to obtain training data, extracting features using a segmentation task model of a deep learning model, and performing classification processing using a classification task model, the reliability and accuracy of the classification results are improved by combining multi-task learning and loss function optimization.

Benefits of technology

It achieves high-precision classification of multi-phase liver medical images, improving the reliability and accuracy of liver medical image classification results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118968116A_ABST
    Figure CN118968116A_ABST
Patent Text Reader

Abstract

The invention provides a medical image processing method and system, a prediction method, electronic equipment and a medium. The medical image processing method comprises the following steps: acquiring a plurality of stages of liver medical images; preprocessing the multi-stage liver medical image to obtain training data; training an original deep learning model by using the training data to obtain a deep learning model; performing feature extraction on a to-be-classified liver medical image by using the segmentation task model of the deep learning model to obtain image features; and performing classification processing on the to-be-classified liver medical image according to the image features by using a classification task model of the deep learning model to obtain a classification result. According to the medical image processing method, classification of multiple stages of liver medical images can be realized, and the reliability and precision of a liver medical image classification result are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application belongs to the technical field of medical image processing, and relates to a medical image processing method, and in particular to a medical image processing method, system, prediction method, electronic equipment and medium. Background Art

[0002] Liver medical image processing is an important technical field used in the diagnosis and treatment of liver diseases. With the advancement of medical imaging equipment and the development of computer vision technology, liver medical image processing has made significant progress and has been widely used in recent years.

[0003] Liver medical images are usually derived from various imaging techniques, such as computed tomography, magnetic resonance imaging, and ultrasound imaging. These imaging techniques can provide high-resolution liver structural information, including the liver's morphology, size, tissue density, and possible lesion areas. By analyzing and processing these images, doctors can detect early lesions, classify lesion types, and monitor treatment effects. In liver medical image processing, key technologies include image enhancement, feature extraction, segmentation, and classification. Image enhancement aims to improve image quality and reduce noise and artifacts to more clearly display liver structures and lesion areas. Feature extraction is to extract meaningful features such as shape, texture, and density from images through mathematical and computer vision algorithms to help distinguish normal tissue from abnormal areas. Image segmentation is the process of separating the area of ​​interest in the liver image, which is usually used to locate and quantify liver lesions. This step is crucial for surgical planning and treatment planning. Finally, the classification algorithm classifies the processed image areas into different categories, such as tumors, cysts, inflammation, or normal tissue, to assist doctors in disease diagnosis and treatment decisions.

[0004] Although significant progress has been made in liver medical image processing technology, it still faces some challenges, one of the most notable problems being the low precision and accuracy of classification results. Summary of the invention

[0005] In view of the above-mentioned shortcomings of the prior art, the purpose of the present application is to provide a medical image processing method, system, prediction method, electronic device and medium, which are used to solve the problem of low precision and accuracy of the classification results of liver medical images in the prior art.

[0006] In a first aspect, the present application provides a medical image processing method, which includes: acquiring multi-phase liver medical images; preprocessing the multi-phase liver medical images to obtain training data; using the training data to train an original deep learning model to obtain a deep learning model; using a segmentation task model of the deep learning model to extract features of the liver medical images to be classified to obtain image features; using a classification task model of the deep learning model to classify the liver medical images to be classified according to the image features to obtain classification results.

[0007] In this application, the model is trained using multi-phase liver medical images, the features of the liver medical images to be classified are extracted using the acquired deep learning model, and then the classification results are obtained using the deep learning model based on the acquired image features. This medical image processing method can achieve the classification of multi-phase liver medical images and improve the reliability and accuracy of the liver medical image classification results.

[0008] In an implementation of the first aspect, the multi-phase liver medical image includes a plain scan phase liver medical image, an arterial phase liver medical image, a venous phase liver medical image and / or a delayed phase liver medical image.

[0009] In an implementation of the first aspect, preprocessing the multi-phase liver medical images to obtain training data includes: slicing the multi-phase medical images to obtain a two-dimensional medical image; obtaining a corresponding mask and a tumor position based on the two-dimensional medical image; and obtaining the training data using the corresponding mask and the tumor position of the two-dimensional medical image.

[0010] In an implementation of the first aspect, training the original deep learning model using the training data includes: compensating the segmentation task model of the deep learning model using a first loss function, and compensating the classification task model of the deep learning model using a second loss function.

[0011] In an implementation of the first aspect, using the segmentation task model of the deep learning model to perform feature extraction on multi-phase liver medical images to be classified to obtain image features includes: using an encoder in the segmentation task model of the deep learning model to extract high-level features of the liver medical images to be classified to obtain the image features.

[0012] In an implementation of the first aspect, using the classification task model of the deep learning model to classify the liver medical image to be classified according to the image features to obtain a classification result includes: using the segmentation task model of the deep learning model to segment the liver medical image to be classified to obtain a segmentation result; using the classification task model of the deep learning model to classify the liver medical image to be classified according to the segmentation result and the image features to obtain confidence levels of each category; analyzing and processing the confidence levels of each category to obtain the category with the highest confidence level as the classification result.

[0013] In a second aspect, the present application provides a medical image prediction method for testing the deep learning model described in any one of the first aspects, the medical image prediction method comprising: obtaining a multi-phase liver medical image to be predicted and a corresponding mask; locating the position of the tumor according to the multi-phase liver medical image to be predicted and the corresponding mask; classifying the multi-phase liver medical image to be predicted using a deep learning model to obtain a prediction result; and verifying and analyzing the prediction result according to the position of the tumor to obtain the credibility of the prediction result.

[0014] In a third aspect, the present application provides a medical image processing system, which includes: an image acquisition module for acquiring multi-phase liver medical images; a data acquisition module for preprocessing the multi-phase liver medical images to obtain training data; a model training module for training an original deep learning model using the training data to obtain a deep learning model; a feature acquisition module for extracting features of the liver medical images to be classified using the segmentation task model of the deep learning model to obtain image features; and a result acquisition module for classifying the liver medical images to be classified according to the image features using the classification task model of the deep learning model to obtain classification results.

[0015] In a fourth aspect, the present application provides an electronic device, comprising: a memory for storing a computer program; and a processor for executing the computer program stored in the memory, so that the electronic device performs the medical image processing method as described in any one of the first aspect and / or the medical image prediction method described in the second aspect.

[0016] In a fifth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the medical image processing method described in any one of the first aspects and / or the medical image prediction method described in the second aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1AShown is a schematic diagram of an application scenario of the medical image processing method described in this application.

[0018] Figure 1B Shown are structural diagrams of the client-cloud interaction scenarios in these implementations.

[0019] Figure 2A Shown is a flowchart of the medical image processing method described in an embodiment of the present application.

[0020] Figure 2B Shown is a schematic diagram of an implementation method described in an embodiment of the present application.

[0021] Figure 3A Shown is a flowchart of the medical image processing method described in an embodiment of the present application.

[0022] Figure 3B Shown is a schematic diagram of an implementation method described in an embodiment of the present application.

[0023] Figure 4 Shown is a flowchart of the medical image processing method described in an embodiment of the present application.

[0024] Figure 5 Shown is a flow chart of the medical image prediction method described in an embodiment of the present application.

[0025] Figure 6 Shown is a structural schematic diagram of the medical image processing system described in an embodiment of the present application.

[0026] Figure 7 Shown is a schematic diagram of the structure of an electronic device described in an embodiment of the present application.

[0027] Component number description

[0028] 1 Medical Image Processing Device

[0029] 11 Image acquisition equipment

[0030] 12 Local Processor

[0031] 13 Display Terminal

[0032] 2-end-cloud interactive system

[0033] 20 Terminal

[0034] 21 Cloud Server

[0035] 100 Medical Image Processing System

[0036] 110 Image acquisition module

[0037] 120 Data acquisition module

[0038] 130 Model training module

[0039] 140 Feature acquisition module

[0040] 150 Result Acquisition Module

[0041] 700 Electronic equipment

[0042] 710 Memory

[0043] 720 processor

[0044] 730 Display

[0045] Steps S11 to S15

[0046] Steps S121 to S123

[0047] Steps S151 to S153

[0048] Steps S21 to S24 DETAILED DESCRIPTION

[0049] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0050] It should be noted that the illustrations provided in the following embodiments are only used to illustrate the basic concept of the present application in a schematic manner. Therefore, the drawings only show components related to the present application rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0051] In the field of medical image processing, the segmentation and classification technology of liver tumors has always attracted much attention because it is crucial for the early diagnosis and treatment of liver cancer. With the rapid development of deep learning technology, deep learning models have been widely used in medical image processing. In terms of segmentation, traditional methods mainly rely on pixel-level features and manually designed feature extraction algorithms, such as edge detection and region growing. However, these methods often encounter problems of inaccurate segmentation and long time consumption when processing complex medical images. With the rise of deep learning technology, semantic segmentation methods based on deep learning have gradually become mainstream. Although deep learning models have made significant progress in segmentation tasks, there are still challenges in tumor classification. Traditional classification methods usually rely on manually designed feature extraction and machine learning algorithms, such as support vector machines (SVM) and random forests. However, these methods are often limited by expert experience and manually selected features, and are not applicable to complex medical image data. At the same time, these methods often cannot fully explore the deep features in the image, resulting in limited classification accuracy.

[0052] At least to address the above-mentioned problems, an embodiment of the present application provides a medical image processing method, which includes: acquiring multi-phase liver medical images; preprocessing the multi-phase liver medical images to obtain training data; using the training data to train an original deep learning model to obtain a deep learning model; using a segmentation task model of the deep learning model to perform feature extraction on the liver medical images to be classified to obtain image features; using a classification task model of the deep learning model to classify the liver medical images to be classified according to the image features to obtain classification results.

[0053] In the embodiment of the present application, the model is trained using multi-phase liver medical images, the features of the multi-phase liver medical images are extracted using the acquired deep learning model, and then the classification results are obtained using the deep learning model based on the acquired image features. This medical image processing method can achieve the classification of multi-phase liver medical images and improve the reliability and accuracy of the liver medical image classification results.

[0054] Figure 1A The figure shows an application scenario of the medical image processing method described in the present application. The medical image processing device 1 can be used to implement the medical image processing method provided in the embodiment of the present application, but the application scenario of the medical image processing method provided in the embodiment of the present application is not limited to Figure 1A The medical image processing device 1 shown. Figure 1A As shown, the medical image processing apparatus 1 includes an image acquisition device 11, a local processor 12 and a display terminal 13. The medical image processing method provided in the embodiment of the present application can be applied to the local processor 12.

[0055] in, Figure 1A The local processor 12 in the embodiment may be a local processor or a local processor cluster composed of multiple local processors or a cloud computing center, etc., which is not limited to the specifics herein. Figure 1A Only one image acquisition device 11, one local processor 12 and one display terminal 13 are shown, but it should be understood that Figure 1A The examples are only used to understand the present solution, and the specific numbers of local processors 12 and display terminals 13 should be flexibly determined based on actual conditions.

[0056] In some other implementations, the medical image processing device 1 may not include the display terminal 13, but only include a local processor 12 with a display function and an image acquisition device 11. The medical image processing method provided in the embodiment of the present application can be applied to the local processor 12. The local processor 12 with a display function may include a tablet computer, a laptop computer, a PDA, a mobile phone, a personal computer (PC) and a voice interaction device, or a monitoring device, an ultrasonic detection device, etc., which is not limited here.

[0057] In still other implementations, the medical image processing method described in the present application may be applied to end-cloud interaction scenarios. Figure 1B The following is a schematic diagram of the structure of the client-cloud interaction scenario in these implementations. Figure 1B As shown, the terminal-cloud interaction system 2 includes a terminal 20 and a cloud server 21. The terminal 20 and the cloud server 21 can communicate with each other, and the communication method is not limited to wired or wireless.

[0058] Among them, the terminal 20 can be mobile or fixed, for example, the terminal 20 can be a wireless terminal or a wired terminal. The wireless terminal can refer to a device with wireless transceiver function, which can be deployed indoors, outdoors and in industrial workshops. The terminal 20 can be a mobile phone, a tablet computer, a laptop computer, etc., which is not limited here. The cloud server 21 can include one or more servers, or one or more processing nodes, or one or more virtual machines running on the server. The cloud server 21 can also be called a server cluster, a management platform, a data processing center, etc., which is not limited in the embodiments of the present application.

[0059] The technical solutions in the embodiments of the present application will be described in detail below in conjunction with the drawings in the embodiments of the present application.

[0060] The following embodiments of the present application provide a medical image processing method. The medical image processing method can be, for example, Figure 1A The local processor 12 shown or Figure 1B The cloud server 21 shown is implemented. Fig. 2 is a schematic flow chart of the medical image processing method according to an embodiment of the present application. As shown in Fig. 2, the medical image processing method includes the following steps S11 to S15.

[0061] Step S11, acquiring multi-phase liver medical images, wherein the multi-phase liver medical images are image acquisition methods and features at different time points or different injection stages when acquiring medical images.

[0062] Step S12, preprocessing the multi-phase liver medical images to obtain training data, wherein preprocessing the multi-phase liver medical images includes image addition, image magnification, normalization and other processing methods.

[0063] Step S13, using the training data to train the original deep learning model to obtain a deep learning model. The original deep learning model is an untrained convolutional neural network, corresponding to the deep learning model.

[0064] Step S14, using the segmentation task model of the deep learning model to perform feature extraction on the liver medical image to be classified to obtain image features.

[0065] Step S15, using the classification task model of the deep learning model to classify the liver medical image to be classified according to the image features to obtain a classification result.

[0066] In some possible implementations, multiple liver medical images are acquired, and the multiple liver medical images are preprocessed to acquire training data. The original deep learning model is trained using the training data to acquire a deep learning model. Figure 2B A schematic diagram showing an implementation method described in an embodiment of the present application is shown as follows: Figure 2B As shown, the original deep learning model includes a segmentation task model and a classification task model. The segmentation task model and the classification task model in the original deep learning model are trained by mutual supervision using the training data, so as to achieve better training effect by using the supervision between tasks. The segmentation task model in the deep learning model is used to extract features of the multi-phase liver medical images to obtain image features. The classification task model in the deep learning model is used to classify the multi-phase liver medical images according to the image features to obtain classification results.

[0067] In some other possible implementations, the deep learning model includes a segmentation task model and a classification task model. The segmentation task model includes an encoder and a decoder network, which is based on ResNet50 and uses a dilated convolution technique to expand the receptive field of the convolution kernel by introducing holes between the convolution kernels, thereby increasing the receptive field of the network without increasing the number of parameters. This is conducive to capturing a wider range of information in the segmentation task and improving the segmentation performance of the model.

[0068] In the embodiment of the present application, the model is trained using multi-phase liver medical images, the features of the liver medical images to be classified are extracted using the acquired segmentation task model of the deep learning model, and then the classification results are obtained using the classification task model of the deep learning model based on the acquired image features. This medical image processing method can achieve the classification of multi-phase liver medical images and improve the reliability and accuracy of the liver medical image classification results.

[0069] In one embodiment of the present application, the multi-phase liver medical image includes a plain scan liver medical image, an arterial phase liver medical image, a venous phase liver medical image and / or a delayed phase liver medical image. The plain scan liver medical image refers to a medical image obtained by conventional scanning using X-ray, magnetic resonance or ultrasound technology before the contrast agent is injected, that is, in the venous phase. The arterial phase liver medical image refers to a medical image obtained by imaging examination after the contrast agent is injected into the patient and the contrast agent quickly enters the arterial vascular system of the liver through the blood flow. The venous phase liver medical image refers to a medical image obtained by imaging examination at this time after the contrast agent is injected and the contrast agent reaches the portal vein of the liver through the venous system. The delayed phase liver medical image refers to a medical image obtained by imaging examination performed some time after the contrast agent is injected (usually a few minutes to more than ten minutes).

[0070] Figure 3A The figure shows a flow chart of the medical image processing method described in the embodiment of the present application. Figure 3A As shown, the step S12 includes the following steps S121 to S123.

[0071] Step S121 , performing slicing processing on the multiple medical images to obtain a two-dimensional medical image.

[0072] Step S122: acquiring a corresponding mask and a tumor position according to the two-dimensional medical image.

[0073] Step S123: Acquire the training data using the corresponding mask of the two-dimensional medical image and the position of the tumor.

[0074] In some possible implementations, since the deep learning model is two-dimensional and the acquired multi-phase liver medical images are three-dimensional, the multi-phase liver medical images need to be sliced ​​to obtain the two-dimensional medical images. The positions of each tumor are found in the mask corresponding to the two-dimensional medical images by the connected domain method, ensuring that each mask corresponding to the two-dimensional medical image contains only one liver tumor, and the training data is obtained using the corresponding mask of the two-dimensional medical image and the position of the tumor to ensure the accuracy and consistency of the training data. Figure 3B The diagram is a schematic diagram of an implementation method described in an embodiment of the present application. Figure 3B As shown, the unprocessed multi-phase liver medical image contains two tumors, and each of the two-dimensional medical images after slice processing contains only one tumor.

[0075] In one embodiment of the present application, using the training data to train the original deep learning model includes: using a first loss function to compensate the segmentation task model of the deep learning model, and using a second loss function to compensate the classification task model of the deep learning model.

[0076] In some possible implementations, the number of training data based on different lesions may be unbalanced, and the focal loss function is used as the second loss function to compensate the classification task model of the deep learning model. The dice loss function is used as the first loss function to compensate the segmentation task model of the deep learning model to improve the comprehensive performance of the deep learning model, and a multi-task learning method is used to enable the model to better understand the image and improve the accuracy and robustness of the classification.

[0077] In one embodiment of the present application, the segmentation task model of the deep learning model is used to perform feature extraction on the liver medical image to be classified to obtain image features, including: using the encoder in the segmentation task model of the deep learning model to extract high-level features of the liver medical image to be classified to obtain the image features.

[0078] In some possible implementations, the segmentation task model of the deep learning model includes an encoder and a decoder. The encoder of the deep learning model is used to extract high-level features in the liver medical image to be classified. The high-level features are more abstract features than low-level features such as edges and colors, so as to obtain image features containing rich semantic information.

[0079] Figure 4 The figure shows a flow chart of the medical image processing method described in the embodiment of the present application. Figure 4 As shown, the step S15 includes the following steps S151 to S153.

[0080] Step S151, using the segmentation task model of the deep learning model to perform segmentation processing on the liver medical image to be classified to obtain a segmentation result.

[0081] Step S152: classify the liver medical image to be classified using the classification task model of the deep learning model according to the segmentation result and the image features to obtain the confidence of each category.

[0082] Step S153, analyzing and processing the confidences of the categories to obtain the category with the highest confidence as the classification result.

[0083] In some possible implementations, when considering each liver tumor one by one, a tumor may appear in multiple phases of liver medical images, and the slices of the liver medical image where a tumor is located are usually multi-layered, so the liver medical images of each phase corresponding to the tumor and the corresponding mask are input into the deep learning model for classification processing to obtain the confidence of each category. The sum of the confidences of all slices is counted to obtain the category with the largest confidence as the classification result of the multiple phases of liver medical images.

[0084] Figure 5 The figure shows a flow chart of the medical image prediction method described in the embodiment of the present application, which is used to test the deep learning model described in any embodiment of the present application. Figure 5 As shown, the medical image prediction method includes the following steps S21 to S24.

[0085] Step S21, obtaining multi-phase liver medical images to be predicted and corresponding masks.

[0086] Step S22: locating the position of the tumor according to the multi-phase liver medical image to be predicted and the corresponding mask.

[0087] Step S23: classify the multi-phase liver medical images to be predicted using a deep learning model to obtain a prediction result.

[0088] Step S24, verifying and analyzing the prediction result according to the location of the tumor to obtain the credibility of the prediction result.

[0089] In some possible implementations, for multiple phases of liver medical images and corresponding masks, first locate all liver tumors in the multiple phases of liver medical images to be predicted, and determine the position of each tumor on each phase of liver medical images to ensure the accuracy of data subsequently input into a deep learning model for processing. The multiple phases of liver medical images to be predicted are classified using a deep learning model to obtain a prediction result. The prediction result is verified and analyzed according to the position of the tumor to obtain the credibility of the prediction result.

[0090] In the embodiment of the present application, the position of the tumor is located according to the multi-phase liver medical image to be predicted and the corresponding mask, and then the multi-phase liver medical image to be predicted is classified using a deep learning model to obtain a prediction result. The position of the tumor and the prediction result are used for verification analysis to obtain the credibility of the prediction result. This medical image prediction method can use the information of the multi-phase liver medical image to be predicted to improve the reliability of the prediction result.

[0091] This medical image processing method can realize the classification of multi-phase liver medical images and improve the reliability and accuracy of liver medical image classification results.

[0092] Figure 6 The structure diagram of the medical image processing system described in the embodiment of the present application is shown. Figure 6 As shown, the medical image processing system 100 includes an image acquisition module 110 , a data acquisition module 120 , a model training module 130 , a feature acquisition module 140 and a result acquisition module 150 .

[0093] The image acquisition module 110 is used to acquire multi-phase liver medical images.

[0094] The data acquisition module 120 is used to preprocess the multi-phase liver medical images to obtain training data.

[0095] The model training module 130 is used to train the original deep learning model using the training data to obtain a deep learning model.

[0096] The feature acquisition module 140 is used to extract features of the liver medical image to be classified using the segmentation task model of the deep learning model to obtain image features.

[0097] The result acquisition module 150 is used to classify the liver medical image to be classified according to the image features using the classification task model of the deep learning model to obtain a classification result.

[0098] It should be noted that the above modules 110 to 150 included in the medical image processing system 100 correspond one-to-one to steps S11 to S15 in the medical image processing method shown in FIG. 2 , and are not described in detail here.

[0099] In the embodiment of the present application, the model training module 130 is used to implement multi-task learning of multi-stage liver medical image segmentation tasks and classification tasks using a deep learning model, and the two tasks supervise each other. At the same time, the result acquisition module 150 is used to make full use of the information of multi-stage liver medical images to improve the accuracy of the classification results. The deep learning model can also improve the prediction efficiency of medical images by using automated segmentation and classification.

[0100] In the several embodiments provided in the present application, it should be understood that the disclosed system, device or method can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of modules / units is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules or units, which can be electrical, mechanical or other forms.

[0101] The modules / units described as separate components may or may not be physically separated, and the components displayed as modules / units may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules / units may be selected according to actual needs to achieve the purpose of the embodiments of the present application. For example, the functional modules / units in the various embodiments of the present application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0102] Those of ordinary skill in the art should further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0103] An embodiment of the present application also provides an electronic device. Figure 7 The structure diagram of the electronic device 700 described in the embodiment of the present application is shown. Figure 7 As shown, in this embodiment, the electronic device 700 includes a memory 710 and a processor 720 .

[0104] The memory 710 is used to store computer programs; preferably, the memory 710 includes: ROM, RAM, disk, USB flash drive, memory card or CD-ROM and other media that can store program codes.

[0105] Specifically, the memory 710 may include a computer system readable medium in the form of a volatile memory, such as a random access memory (RAM) and / or a cache memory. The electronic device 700 may further include other removable / non-removable, volatile / non-volatile computer system storage media. The memory 710 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present application.

[0106] The processor 720 is connected to the memory 710 and is used to execute the computer program stored in the memory 710 so that the electronic device 700 executes the medical image processing method described in any embodiment of the present application and / or the medical image prediction method described in the embodiments of the present application.

[0107] Optionally, the processor 720 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0108] Optionally, the electronic device 700 in this embodiment may further include a display 730. The display 730 is communicatively connected to the memory 710 and the processor 720, and is used to display a graphical user interface (GUI) interaction interface related to the medical image processing method described in the embodiment of the present application and / or the medical image prediction method described in other embodiments of the present application.

[0109] The present application also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the medical image processing method and / or the medical image prediction method described in any embodiment of the present application is implemented.

[0110] The descriptions of the processes or structures corresponding to the above-mentioned figures have different emphases. For parts that are not described in detail in a certain process or structure, please refer to the relevant descriptions of other processes or structures.

[0111] The above embodiments are merely illustrative of the principles and effects of the present application and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.

Claims

1. A medical image processing method, It is characterized in that include: Acquire multi-phase medical images of the liver; Preprocessing the multiple liver medical images to obtain training data; Using the training data to train the original deep learning model to obtain a deep learning model; Using the segmentation task model of the deep learning model to perform feature extraction on the liver medical image to be classified to obtain image features; The classification task model of the deep learning model is used to classify the liver medical image to be classified according to the image features to obtain a classification result.

2. The medical image processing method according to claim 1, It is characterized in that The multi-phase liver medical image includes a plain scan phase liver medical image, an arterial phase liver medical image, a venous phase liver medical image and / or a delayed phase liver medical image.

3. The medical image processing method according to claim 1, It is characterized in that Preprocessing the multi-phase liver medical images to obtain training data includes: Performing slicing processing on the multiple medical images to obtain a two-dimensional medical image; Acquire a corresponding mask and a position of a tumor according to the two-dimensional medical image; The training data is acquired using the corresponding mask of the two-dimensional medical image and the position of the tumor.

4. The medical image processing method according to claim 1, It is characterized in that Training the original deep learning model using the training data includes: The segmentation task model of the deep learning model is compensated by using the first loss function, and the classification task model of the deep learning model is compensated by using the second loss function.

5. The medical image processing method according to claim 1, It is characterized in that The segmentation task model of the deep learning model is used to extract features of the liver medical image to be classified to obtain image features including: The encoder in the segmentation task model of the deep learning model is used to extract high-level features of the liver medical image to be classified to obtain the image features.

6. The medical image processing method according to claim 1, It is characterized in that Using the classification task model of the deep learning model to classify the liver medical image to be classified according to the image features to obtain a classification result includes: Using the segmentation task model of the deep learning model to perform segmentation processing on the liver medical image to be classified to obtain a segmentation result; Classify the liver medical image to be classified using the classification task model of the deep learning model according to the segmentation result and the image features to obtain confidence levels of various categories; The confidences of the categories are analyzed and processed to obtain the category with the greatest confidence as the classification result.

7. A medical image prediction method, It is characterized in that For testing the deep learning model according to any one of claims 1 to 6, the method comprises: Acquire multi-phase liver medical images to be predicted and corresponding masks; Locating the position of the tumor according to the multi-phase liver medical image to be predicted and the corresponding mask; The multi-phase liver medical images to be predicted are classified using a deep learning model to obtain a prediction result; and the prediction result is verified and analyzed according to the location of the tumor to obtain the credibility of the prediction result.

8. A medical image processing system, It is characterized in that include: An image acquisition module, used for acquiring multi-phase liver medical images; A data acquisition module, used for preprocessing the multi-phase liver medical images to obtain training data; A model training module, used to train the original deep learning model using the training data to obtain a deep learning model; A feature acquisition module, used to extract features of the liver medical image to be classified using the segmentation task model of the deep learning model to obtain image features; A result acquisition module is used to classify the liver medical image to be classified according to the image features using the classification task model of the deep learning model to obtain a classification result.

9. An electronic device, It is characterized in that The electronic device comprises: Memory for storing computer programs; A processor, wherein the processor is used to execute the computer program stored in the memory so that the electronic device performs the medical image processing method according to any one of claims 1 to 6 and / or the medical image prediction method according to claim 7.

10. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the program is executed by a processor, the medical image processing method according to any one of claims 1 to 6 and / or the medical image prediction method according to claim 7 are implemented.

Citation Information

Patent Citations

  • Pathological image classification method, device and equipment for membranous nephropathy and storage medium

    CN114141339A

  • Liver tumor segmentation method and device based on symbol distance

    CN116309621A

  • KR20230139257A